Peloton – A Marvel at Customer Acquisition, Brand-Building, and Its Untapped Opportunities

A couple years ago, I dug into the Blue Apron S-1 here. People seemed to liked the way I broke down what wasn’t always explicit in their filing. I figured it’s time to do another breakdown. 

The one similarity they both have is that their S-1’s seemed to be written to disclose only the info they wanted to share.  That’s where the similarities end, as we’ll get into. 

I believe in cutting to the chase and then I’ll get into the nitty gritty. So here’s my summary take on Peloton:

Peloton essentially acquires its customers for free – meaning, their gross margin on the sale of their bikes covers their customer acquisition costs. Not just of their media costs. But of their fully-loaded marketing costs, inclusive of people and showroom expenses.  

They have 500K+ ridiculously fanatical customers who seem to be churning out at less than 1% per month.  We’ll dig in to this figure, but even if the figure is 5% per month, which I don’t believe it is, it’s one of the highest retention rates I’ve seen in a subscription business (comparable to DirecTV in its heyday and Netflix). And certainly remarkable for a fitness biz where the average lifetime is 3-6 months. 

Yes, they have $200MM (and growing) of Operating Expenses they have to cover.  They’ll need to increase their subscriber base by 50%-100% to cover those expenses, which will likely go up as they continue to scale.  Or they increase revenue/margin dollars per customer.  THERE ARE AT LEAST 2 UNTAPPED OPPORTUNITIES THAT ARE NO-BRAINERS FOR THEM TO DO SO.  MORE BELOW…

If/when the economy turns south, my guess is that they’ll dial back their acquisition machine, cut staff, and dial back new major higher-investment initiatives.  The question is whether those 500K subscribers continue to pay their $40/month.  Based on my experience at Beachbody during the Recession, I have some thoughts here. 

-Bottom line, Peloton has built a following that is ripe for much further monetization, and the Company likely can navigate a downturn in the economy and/or new customer growth

Now Digging In…Let me get a couple things out of the way

What I did NOT like about the S-1.  

I’m a big fan of Peloton as a consumer and more broadly about the business based on what I found and was able to decipher thru the S-1 (not to mention what I’ve seen over the past few years).  But rather than start with a love-fest, I’d rather address some of the negatives first.

First, the LTV figure they provided – based on subscription revenues only. 

This is just a silly figure they’ve put out there.  It doesn’t have context and frankly is based on an expected time period that isn’t realistic.  In trying to show a massive figure, I think they hurt their credibility.  It just doesn’t seem like they needed to do so.  But here’s what they did:

It’s based on the monthly fee ($39), gross subs added, expected lifetime (more in a second), and an adjusted subscription margin.  Btw, the number of “Adjusted” figures in the S-1 is a bit too much.  You can see my calc’s below – the numbers tie out, but it kinda gives you a headache trying to reconcile what they’re referring to. 

As for the lifetime in the calc below, it’s based on the monthly churn.  1 divided by the churn to be precise.  In a business with 8% monthly churn the lifetime would 1/.08=12+ months.  Based on the churn figures Peloton has provided (which overall do NOT seem to be radically inflated) – less than 1% – the average lifetime is ~160 months, which translates to 13+ years. 

I just don’t see the point in making this assumption.  I’m a massive Peloton fan myself but there’s no way I can conceive of being a subscriber for 13 years.  Now, it’s highly possible that they generate the same amount of money from me and other subscribers over a shorter period of time, but using 160+ months as the expected lifetime is, well, not believable.  Now, they don’t actually say 160 months – they just describe the calc. But that’s what it’s based on. 

And as you can see in the below, customers who were acquired in 2019 are expected to generate over $1 billion in revenues over their lifetime.  Over 13 years. 

Um, ok. 

A couple other points about the above.  I’ve backed into the estimated gross subs added and churn based on figures they provided.  You can see the delta from my calc of LTV to what they provided – it’s ~5%.  I’m good with that margin of error.

Similarly, on a per subscriber basis, my calc of the # of subs is essentially identical. 

That $3,593 says that over the course of 13 years, these customers will generate $3,593 of margin dollars. 

One word on the margin % calculation. It’s one of their “Adjusted” figures, backing out Depreciation & Amortization, as well as stock compensation expense.  Both adjustments are not unusual.  But they also back out the costs they’ve incurred for “Content costs for past use” – as below.  Basically, they used content they didn’t have the rights to. Or as they say, it was “alleged past use.”  I’m not sure why it’s alleged.  Since everything is recorded, it should be part of the record to prove one way or another.  These fees below are $45MM over the past 3 years.  Really, they sound like fines and shouldn’t give anyone the warm-and-fuzzies. 

But these costs have also been backed out of subscription costs and increase subscription margin by 9 percentage points (not 9%, but percentage points) in 2019.  Meaning, subscription margin increased from 51% to 60% as a result of backing these out.  Whether these will continue is anyone’s guess.

What else that I didn’t love:

No real detail on cohorts – counts, LTV, churn.  All are at the aggregated level.  I take a stab at reconciling their figures below.  It would appear the numbers in the S-1 are at least directionally representative. If that’s the case, then why not share more detail? If your cohorts really are that strong, then share the data – most everyone else doesn’t.  They’d stand out even more if they would do so.  

This awful definition of Churn:

Particularly in a fast-growing business (subs are literally 5x from 2017 and 2x the 2018 figures), taking the current period’s cancellations relative to the beginning of period subs figures – well, it doesn’t tell you much. Who is cancelling? Are they recent subs or older ones? This is why cohort figures are so important.

The quarterly figures provide at least a small bit more visibility into what’s happening, but not at the level of detail that most of us would prefer. And as I alluded to above, the typical thinking is – if you aren’t sharing the info, what are you trying to hide? And from the other side, sharing the info is such a strong sign of confidence.

Later in this post, I do my best to make sense of their numbers. There seems to be a recent trend of increasing cancels – both absolute and percentage-wise, but it’s not a crazy figure. At least not yet. And they still have some pretty strong fundamentals we’ll get into.

-Content Costs for Past Use – I mentioned this above, but it deserves an additional call-out. Essentially, Peloton has paid fees (fines) for using content (presumably music) without having the rights to do so. To the tune of $45MM over the past 3 years. Which doesn’t include at least some of the $19MM in litigation expenses.

I don’t like it for a few reasons: a) anyone who has any IP knows what a pain it is to enforce it – IP rights are only as good as the oftentimes-deep pockets you need to pursue those breaching those right; b) there’s a much bigger lawsuit by more content rights owners, seeking $150 million in damages. Who knows what amounts will be paid and/or whether that’ll show up as cash fines or stock ownership for those parties. But given Peloton’s history, it’s reasonable to expect some amounts will be paid out.

Almost 2,000 employees.  Wowza! That’s a lot of people.  Sure they have ~80 showrooms, and this is a billion dollar biz, but that’s a lot of people.

This wasn’t in the S-1, but it’s posts like this which can be indications of operational issues. There were 235 comments, many describing a similar experience. Not good. There will always be exceptions. No one is perfect. And as much as my personal installation experience (of the bike, not the treadmill) was outstanding, this doesn’t sound like the exception. This type of stuff can be the result of growth outpacing operational support. It just better get addressed very quickly.

Okay, now to the positives.

Customer Acquisition:

I mentioned at the top that they essentially acquire customers for free.  Basically, the margin they make on their bikes and treadmill covers their sales and marketing costs.  What’s particularly amazing is that these costs include people and rent.  In most performance marketing businesses, we look at media costs and then exclude the costs of people.  But Peloton is essentially breakeven for each bike/treadmill they sell.

Below are my estimates of revenue, calc’ed gross profit and CPA (based on my new subs figures above) – all on a per-customer basis.  You can see the comparison of my figures to those provided in the S-1.  It’s a pretty good match.

Now, I wouldn’t say this is unheard of in fitness businesses. To be breakeven on day 0.  I’ve done it in businesses I’ve been involved with and seen it with my clients.  But it’s still pretty awesome when it does happen. 

The implication is that everything they generate afterwards is margin.  Or you could say they don’t make any money until the next sale.  But it’s still better than the majority of venture-funded businesses that go in the hole for 6-18 months before being breakeven on a per-customer basis.  Peloton is breakeven on a per customer basis pretty much on Day 0. 

Subscription Revenues and the Rest of the Economic Model

Whether you believe the fees for use of prior content is in the past or they’ll continue to incur these fees, their Adjusted 2019 subscription margin was somewhere between $92MM and $108MM.  That’s pretty good.  Until you look at the nearly $200M of OpEx they incurred – a combination of R&D and G&A. 

In this case below, I don’t think it’s unreasonable to back out stock-based comp expense nor D&A, especially when considering the business’ going-forward model. 

$179 million in non-sales and marketing expense is pretty hefty, at 20% of revenues.  And of course, that’s for the full year. It’s very likely their run rate figures are even higher as they continue to invest in their business. As for the 20% figure, for Peloton and with their model, it doesn’t seem so crazy.  But what are the implications and what is it going to take to get to breakeven?

Looking at the figures below, assuming $179 million in OpEx, that implies they’ll need 753K subscribers to pay $39 per month for the year.  But they’ll likely incur greater OpEx to get there. 

If we look at their current base of 511k, with no net change in subs, and paying $39 per month for the year, they’ll generate $121 million in gross margin.  That leaves $58 million left to cover. 

Somewhere in this model, there has to be greater efficiency and leverage.

Can Peloton get to 753K subscribers and beyond?

What is going to take to get there?

What happens if they don’t grow?

I’ve got a few thoughts. 

The last two lines in the above table are calc’s showing that if Peloton can generate another $100 in margin dollars per customer, how many would they need to cover that $58 million.  It’s simple math. They’d need 576K subs.  A bit more than they have today. 

And if they could generate another $200 of margin dollars per customers? That’ 288K customers. 

Now you’re probably wondering how feasible it is to get the annual revenue from subscribers to increase from $468 ($39*12) to $568?

I do NOT believe it comes from converting Bike buyers into Treadmill buyers.  But rather…

  1. Apparel – the “Other” revenue line reflected ~$15 million in revenues in 2019.  That’s a massive opportunity waiting to be capitalized on.  Frankly, during my time at Beachbody, we didn’t sell much apparel either.  But when customers are fanatical about a health and fitness product – as they were with P90X and Insanity – and certainly as they are with Peloton, they love being walking billboards for the brand.  “Oh, this Peloton sweatshirt? Oh, I forgot I’m wearing it.  But yea, I have a Peloton. I looooove it. Did you have one? Let me tell you all about it…”

It’s a no-brainer in my estimation.  Now, if they put together a referral program a la Morning Brew, and turn those customers into referrers and compensate them with apparel, that might not generate apparel revenues, but it’s a customer acquisition channel that is a ton cheaper than the current $1000 they pay to acquire a customer. 

You don’t think a customer wouldn’t be absolutely thrilled to get a Peloton sweatshirt or cycling shirts or tank for every customer they referred?

Okay, so I combined 2 point in there – apparel and customer referrals. Yes, they get a ton of word-of-mouth benefit already that is driving referrals. I’d think that compensating customers with a piece of apparel would move the needle further.

2. Supplements – this is the MONSTER opportunity that I think Peloton has waiting in the wings. I have zero info on whether they’re spending any time here.  But think about it.  A rabid customer based that is obsessed with the brand, the trainers, and with being healthier. 

Pills, powders, shakes, bars.  Generally high margin, another source of recurring revenues.  And an entirely reasonable way to generate another $100+ of margin dollars per customer. 

Both of these draw on one of the big wins that Peloton has achieved.  I’ll get to that just below, but I first wanted to share why I think Peloton can win.

Sorry, the marketer in me can’t help but mention untapped opportunity #3:

3. Email – they could do it better.  Frankly, they haven’t needed to.  Communicating with their customers via email extremely well, especially after the purchase, hasn’t been a necessity.  But if/when they really emphasize product sales like the above, they’ll need to get it dialed in.  It’ll happen. Just a matter of time.

Ok, why Peloton can win…

The Peloton Family is Real

I posted this a couple days ago on Twitter.  Look at the post below.  It had 5,000+ likes and over 250 comments.  The fact that a customer shared difficulties about their personal life in the Peloton Facebook group and received such love, both from the community and from Peloton itself, speaks volumes about the brand. 

Major props to the Peloton team for the hand-written card and apparel they sent. 

This type of thing doesn’t happen overnight.  And when you think about competitive advantage, this is huge.

They have nailed the Customer Experience

Here’s a list of what they nail:
-The bike
-The instructors
-The tablet
-The quality of the streams
-The workouts themselves
-The music
-The leaderboard
-Sync’ing with Spotify
-The Facebook group
-The shout-outs for anniversary rides
-Hours upon hours of new content every day

The list goes on and on. So many of the core areas that make a difference. Peloton isn’t top-notch at everything. But they do more than almost everybody at the sheer count of things they nail. It’s for this reason that they are able to use the graphic at the very top of this page describing themselves not just as one thing. But as many things.
I’ll give them their due. They create a customer experience unlike almost anyone else.

Increasing workouts per cohort

This graphic below is pretty unreal. 

For starters, it may not sound like much to get customers to workout 10x per month.  Every 3 days doesn’t seem that much.  2+ times per week.  But there are very few fitness companies that get their customers to do so.  At home!

And then to see the count of workouts increasing for a given cohort over time – that means customers are working out more per month each year.  Again, this is remarkable. 

How are they doing it? Well, they now suggest a workout when you finish a ride – upper body, core, stretching.  An extra 5-10 minutes.  Does this mean that the nature of the current workout count is different than those in prior years? Yes.  But it also highlights something that Peloton has been doing for years.

They announce anniversary rides when their folks hit a special number.  50, 100, 500, 1,000.  Could customers cheat? Yes. Can they do shorter rides versus longer rides? Yes.  But Peloton and their customers don’t care. And don’t care to keep track.  No one asks whether someone did 100 20-minute rides or 45 minute rides.  It’s all about engagement and support.   

The Churn Figures Seem Real

I took a stab at trying to back into the churn figures.

Walking thru the chart below, I note which line items were provided in the S-1.  I then calc’ed the incremental subs each quarter (line A) and calc’ed the monthly rev / subscriber just to make sure the numbers made sense. 

I think looked at their Connected Fitness Product Revenues (basically sales of bikes and treadmills), and then based on some average revenue per customer figures I’ve estimated, I back into the Implied New Customers (line B). 

For the most part, line A and B are pretty close.  Q3/Q4 2019 (the March and June quarters) show the highest variance.  To the extent more recent sales have skewed towards the Treadmill and the AOV is higher, that would explain some of the difference.  It’s also possible that the Digital Subscribers (those with a subscription but without a bike/treadmill) is increasing.  Or, as I mention below, churn may be increasing. 

Using Line B, I show CAC, which mirrors what I’d done above – basically a $1K cost to acquire a customer. 

In Line C, I look at the difference between A and B.  This is a proxy for churn – the new customers are higher than the incremental subs.  Presumably most buyers become subscribers, so the difference is likely churn.  The one potentially worrisome metric is if the Q3/Q4 figures on line C are in fact churn.  In count, those are the highest absolute figures on the line, though still directionally in line as a % of the base.  It’s just a question of which cohort those people are connected to.

Finally, I calculate churn based on the methodology they describe in the S-1.  I don’t have monthly subs figures so I just took the average of the beginning and ending of the quarter, multiplied by the churn percent at the top.  You can see the difference b/w this churn (line D) vs. the one above (line C).  Again, not perfect, but pretty close.    

Yes, it’s possible churn is increasing, albeit with a noticeably larger base.  But for the most part, my calc’s seem in line with the figures the company shared.

Btw, I’ve seen some mention that the Company figures don’t take into account those on an installment plan.  There have been comments suggesting that these customers are not able to cancel their subscription just yet. And when they finish their installment billing (up to 39 months for the bike, 24 months for the treadmill), we’ll see higher cancels. From everything I’ve read, the subscription is NOT tied to the installment plan, so a customer isn’t locked in to the subscription should they want to cancel that part.  Nor does Subscription Revs include the installment payments.  As such, it doesn’t appear that there is mixing and matching, nor do these installment payment customers seem to be skewing the churn figures one way or another necessarily.
(Update: In July 2018, Peloton ended an offer whereby the subscription revenue was either prepaid and/or bundled in to the purchase of the bike. Those plans lasted from 12-39 months. The current offer allows for financing of the bike itself and the subscription is offered separately. There is no info as to how many/what percentage of new subscribers took the bundled installment plan. But it should be noted that some of these customers really haven’t had the ability to churn out, but their revenues are being recognized over time and the subscriber counts continue to include these customers.)

As for the final reason I think Peloton can win…

Have you ever spoken to a Peloton customer?

‘nuf said.  They’re beyond passionate.  And loyal. 

The Bigger Win

What is it that Peloton has achieved?

They’ve acquired a massive number of customers. Of fans. Of devotees. 

Of customers who will fly from wherever they live to do their anniversary ride in the NYC studio. Hopefully with their favorite trainer.  And snapping a pic while they’re at it. 

Of customers who love that they can click a button on their screen when a favorite song plays, and that song sync’s with their Spotify account. 

Over the past few years, they established a foundation of half of a millioin loyal paying customers. That’s a very very asset to have created.

What happens when the market turns?

I’m guessing a few things:

– Sales of bikes and treadmills will drop. This seems like an obvious statement – again, based on the growth we experienced at Beachbody, it’s not entirely a given. But for now, let’s assume it to be true. If the company is watching its numbers closely, they dial back customer acquisition spend.  Let’s assume it goes to zero.  It won’t, but even if it did.  New customer acquisition was operating at breakeven.  It wasn’t adding margin dollars.  But obviously adding new customers into the mix.

-How many customers stay on for $39/month?  I was at Beachbody from 2007 thru 2015.  We grew in a meaningful way thru the Recession, presumably because people gave up their gym memberships to buy a $60-$120 program.  Let’s call it $100 they paid.  Peloton customers are on the hook for $500/year.  Are people going to stay on? And what happens to those that are on installment plans? To date, according to the S-1, bad debt has not been meaningful.  Unless I missed it, they didn’t disclose any figures.  But it will go up. 

-It’s not clear just how much leverage they will get with their current OpEx and how much additional overhead is required to add 250K new subscribers.  

-But if the economy goes south, they likely won’t get that additional 250K and will likely cut staff to consolidate around higher-margin opportunities.  

$200 million of non-sales and marketing OpEx is huge.  Their current employees don’t want to hear it, but realistically, cutting back on R&D and tightening up with those 2,000 employees, is something they’ll need to do. 

But with coffers potentially of over $600 million post-IPO, tightening up as needed on acquisition and/or G&A, and expanding its offering into additional high-margin categories (ingestibles in particular), Peloton appears in position to grow.  And certainly to withstand a market downturn. 

Having 500,000+ active customers is a huge advantage. It’s a pretty good bet that if offered more, they’ll take it.

If all else fails, those 500,000 will come to the Company’s rescue. 

That’s what families do. 

And make no mistake, they’ve got a huge Peloton family.   

How Blue Apron Revealed a Lot More about Customer Acquisition and Retention than It Likely Intended To

I’ll cut to the chase first.  I have no inside information, but my analysis has lead me to conclude that Blue Apron is acquiring customers at a rough CPA of $150, its current monthly churn appears to be ~10%, and the Company eventually nets ~$150+ of contribution margin over the lifetime of an average customer.

Those numbers aren’t written in their S-1, but the analysis used to come up with them is based entirely on the information provided in that S-1.  No one else, as far as I have found, has published them.  I frankly haven’t seen even educated guesses at specific numbers.  Other than saying that the current number is probably higher than the $94 figure they disclose in their filing – but that number is based on the average of 2014 through early 2017 results.

For background, I helped to build the analytics team at Beachbody, then ran media and customer acquisition there for the last 5 of my 8 years.  I’ve since spent the last 2 years helping my clients use information to dramatically improve their customer acquisition and retention efforts.  This isn’t an entirely new area for me.  And certainly, if someone or the Company proves me wrong, I’ll be the first to acknowledge it.

My goal here is to show how you can dig through information that may not be explicitly detailed and how I got to my conclusions.

More important than what the numbers they shared reveal is what the implications are for their own business as well as others in the meal delivery space, and the broader subscription business.  More on these points in a bit.

What Blue Apron Shared

The below graphic is the one that has been making the rounds.  It is pulled out of their S-1.

Chart #1

What the graphic shows is the average cost to acquire a customer during the 2014 to early 2017 period.  I’m not sure why it’s called CPC (“Cost per Customer”), but it’s more normally called CPA (“Cost per Acquisition”).  In addition, it shows the lifetime revenues for customer over various periods after their first paid order.  Since the date range considered is 2014 to early 2017, but we are only in mid-2017, the metrics shown for 30 months out and beyond are based on customers in the earlier part of the range.

Let’s also not gloss over the clearly very intentional label, “Months after First Paid Order.”  One frequently-used component of Blue Apron’s customer acquisition strategy has been to give away meals and/or the first box.  That language is specific, such that if a customer doesn’t pay for the first box, they are excluded from the lifetime revenues part of the table.

As for the Marketing Expense associated with these free boxes, there’s this language from page 67 of the S-1 – “Also included in marketing expenses are the costs of orders through our customer referral program, in which certain existing customers may invite others to receive a complimentary meal delivery…The cost of the customer referral program is based on our costs incurred for fulfilling a complimentary meal delivery, including product and fulfillment costs.”

Re: the definition of a customer:  “We determine our number of Customers by counting the total number of individual customers who have paid for at least one Order from Blue Apron across our meal, wine or market products.”
Which means the marketing expense captures the cost, but shouldn’t capture the count of customers if they didn’t convert into a paying customer.
Net net, I actually think the way I’ve done by analysis is okay given how they’ve calculated their metrics. But real churn, if you included free trial customers, would be dramatically higher.  Churn from paying customers is all I could calculdate.  But the fact that the Marketing Expense includes the product/fulfillment cost means they are correctly loading up their marketing numbers.

Next, the below is another table they presented. It’s an eye-sore for everyone, but has some valuable information in it.

Chart #2

Here is a view of the more telling parts of this table.

Chart #3

Based on Charts 1 and 3, we have the information to calculate their acquisition cost and churn.

I’ll start with churn:

The first line is provided in Chart #1.  The second line is just the difference from one period to the next.  Now, the estimated retention is based on the following:

  • From Chart #3, it looks like the Average Revenue per Customer is roughly $247 per quarter. I’ve had to make some simplifying assumptions – this number is a simple average – not perfect but likely close enough.
  • $247 per quarter implies $494 per 6 month period. So if in the period from 6 months to 12 months in, customers generated an incremental $196 of revenue, then that implies that Blue Apron only has 40% of them left ($196/$494).  This is the calculation behind the 40% figure in the 12mo column.  The same math is used for the rest of that line.
  • Dropoff – this is the percentage point difference from the prior period – Note that I haven’t shown Day 0, which was the starting point, to calc the 17% for the 6mo column.
  • A couple things jump out, whether from the Estimated Retention Line or the Dropoff one.
    • Somewhere between month 6 and 12 is where the 50% figure is crossed, which helps to give a sense of the average stick rate of customers
    • Estimated retention seems to go up from month 30 to 36. My guess is 2 main factors are causing this:  a) we don’t have the exact same cohorts in the analysis, only what was available when the S-1 was put together, so it’s not pure apples-to-apples; and b) I’ve made simplifying assumptions based on Average Revenue per Customer, but I would guess that those customers who have stuck around 30 and 36 months have higher-than-average quarterly revenue.

To get a bit closer to when the 50% threshold is crossed to give a sense of the average retention rate (at this volume, I think it’s fair to say that the mean and median are approximately the same), we can do some additional analysis.

Using a straight-line average of dropoff – use the figures at 6 and 12 months as the starting and ending points, respectively, we can generate this table:

From there, it looks like somewhere around month 10.5 is when the 50% figure would be crossed.  If the average customer sticks around 10.5 months, that implies monthly churn of roughly 100/10.5 = 9.5%.

This figure is going to be supported as I get into the CPA calculation.  But one other way to back into that churn rate is to take the Company’s reported $939 of lifetime revenue, and divide it by the simple average revenue per customer, which I’ve calculated at $247 (per quarter).  That implies 3.8 quarters that a customer sticks around, or 11.4 months.

The simple average of these two calc’s would be 10.5 months.  I’m discounting a bit for scale to get to my 10% figured.

Couple additional points here – the average lifetime revenue from those cohorts in the 36mo column will go up.  They’ve simply reported the data they have now.  So arguably LTV will only get better, but at the same time, as the Company attracts a broader group of customers, that’ll be counter-balanced with lower-quality customers.  We can argue about what retention plans they have in place, do people get stickier to the service, etc., but for now, let’s assume this is all a wash.

Now, to calculating the actual CPA Blue Apron is running at

The top part of this table is from Chart #3, the rest is the analysis based on it – my explanation of the table is below:

Change in customers is simple math – the net change in customers from the end of the prior period to the end of the current period.

The next section – Customers acquired at different CPA’s – shows the number of customers acquired based on the Marketing spend reported by the Company and based on varying CPA’s.  Since the average from 2014 through early 2017 is $94, I’ve made the assumption that the Company CPA has been increasing over time – not unreasonable at their scale.  I’ve just used simple $50 increments after trying out a few different ranges.

For example, for Q1-2017, based on $60MM of Marketing spend, and at a $150 CPA, that would imply 404K customers who joined during Q1.

Based on these figures and the net change in customers (to repeat, a calculated figure based on the table), the Churn section is then created.  At $150 CPA, that would imply the Company lost over 247K customers during Q1-2017.

Monthly churn is based on the average of the prior and current period figures, adjusted for the fact that those numbers are quarterly.  Again, sticking with the Q1-2017 and $150 CPA figures, that would imply churn of ~9%.  I’ve simply multiplied the monthly figures by 12 to show annualized numbers.  Anything over 100% describes a scenario where the Company is essentially turning over its entire customer base each year.  We know that in actuality that isn’t the case – there are clearly some customers who have stayed on for a very long time, but it’s a helpful way to make sense of those over 100% figures.

Finally, Lifetime in monthly is calculated based on the Churn percentages – the opposite of what I did in the original churn analysis.

Which assumption on CPA is correct?

If we assume a lower figure, that means that churn is higher.  And vice versa.

But if we layer on the first analysis here, and use the “starting point” of ~10% churn as a possible figure, that leads us to look at the $150 CPA line.  Note that during Q4 of 2015 and 2016, the Company lost a lot of customers, which isn’t unexpected given the holiday season.  Even with the Q1-2016 figure, as much as CPA doesn’t stay constant, the 10% churn figure and $150 CPA seem like defensible metrics for the business based on the information they provided.

Many folks have commented on the fact that Blue Apron didn’t explicitly reveal its churn.  But I wanted to share that it’s possible to back in to that figure.

As for the $150 CPA, it seems reasonably that the current CPA is ~50% higher than the average over the past few years.  At their volume, that isn’t unexpected, not to mention, they are going after more mainstream customers, as opposed to the early adopters.  That usually means spending more to reach those folks. Finally, competition has increased – everyone like pure competitors such as Hello Fresh and Plated, as well as additional options like Munchery, Uber Eats and even Door Dash – have no doubt made customer acquisition more costly for Blue Apron.

What this really means for the business, and how much does the Company make on each customer:

Using the average revenue per customer, alongside the disclosed cost of goods, the Company makes roughly a 33% Gross Margin on its revenues (Revenues less Cost of goods, which include product and fulfillment expenses).  Adjusting the quarterly revenue per customer to a monthly basis, the average monthly gross margin per customer is ~$26.  (Note, at a $150 CPA, that implies 6 months to breakeven on a customer, which is exclusive of all of the Company’s additional expenses such as G&A, capital expenditures, etc.)  If we then multiply that margin figure by the average lifetime, we get a lifetime gross margin, which when reduced by the acquisition cost, gives a lifetime contribution margin (“CM”) per customer.  At a $150 CPA, the Company is roughly making $152 per customer.

Implications for Other Businesses

For starters, Blue Apron has significant funding – almost $200 million according to Crunchbase. That has allowed the Company to operate at a significant loss for pretty much the entirety of its existence.  Many of the other larger food delivery companies are in a similar situation, though the vast majority of companies at large don’t.  In addition, operating in the red on a per customer basis for 6 months+ is more taxing than most companies can afford.

It’s also tough to run a business when 10% of your customer based is cancelling each month.  Those are figures reminiscent of my days at Napster in 2006, when the iPod was the hot thing and subscription music was widely criticized as a non-starter.  (This was also the period when Blockbuster came out with its competitive product to Netflix, and everyone “knew” that was the end of Netflix.  I think we’ve all seen how that worked out.)

10% monthly churn is really high.  As I mentioned earlier, that equates to turning over your customer base more than once a year.  Which puts dramatically more pressure on customer acquisition.  And while the company only has tapped 2-3 million total customers (I’m again making some simplifying assumptions), which is 1% of the US population, with rising competition, it makes its model more difficult.  Not that I’m betting against them necessarily.  More just looking at what their metrics seem to indicate.

Net net, this is a tough business.  The margins aren’t great, customer acquisition isn’t cheap, and retention is unlikely to jump from 10 month to 20 months in the near future.

Something would need to change for their model to dramatically improve.  With these dynamics, it’s not just a “we’ll make it up on volume” play.  But to the extent one or more of those components – core margin, acquisition cost, or retention – could change markedly, the model could be different.

The only one of these metrics that was likely reasonably understood early on was margin.  Until they (or anyone else) began testing, acquisition costs and stick rate were really unknowns.

Which go to the larger point for other businesses, whether in the same or other categories, and that is about how to get as quick an understanding about the business fundamentals and dynamics as possible.

To be clear, there are plenty of subscription businesses that are thriving, both in the food category and beyond.  This is absolutely not a criticism on the category – and frankly it’s not a criticism as much as it is a deep analysis on the Blue Apron business itself.

Every business has to manage through the relationship of Customer LTV and Customer Acquisition Costs.  And then must layer on constraints such as cash, risk it is willing to take, and business objectives – in determining the better path to proceed down and what strategic decision it wants to make as it evolves.

That is the challenge and the fun of forging ground in new categories.

I, for one, am certainly very interested to continue to watch Blue Apron along its own path.

A Proven Framework for Optimizing a Subscription Business

This is not a “how to start a subscription business from scratch” piece.  If you want to know the first 3 things you should do, stop here.  That’s not what you’re going to find.

This piece is intended for marketers with at least $5 million of annualized revenues, and certainly those into 9 figures.  There are plenty of resources on getting started. My expertise is in optimization and scale.  In bringing together analytics to drive actionable marketing insights for subscription businesses, whether a subscription box business, SAAS service, or another type of recurring revenue model.

Feel free to call this a process or a framework, but at its core, it’s about taking apart the business to ensure that the key components and levers are understood.  And then in putting resources in areas that will provide the biggest bang for the buck.

The fact that the list below looks like a step-by-step process is a bit misleading because you don’t wait until one stage is done to move on to the next.  Similarly, if you move on to the next, it doesn’t mean that the prior stage is “done.”

I’ll provide a bit of rationale for the sequencing, and then specific and deeper insights into each.

Here’s the sequence I use when helping those with subscriptions and recurring revenues:

  1. Retention
  2. Build Customer LTV Model
  3. Evaluate media
  4. Funnel reporting and testing
  5. The value of being a member / customer journey mapping
  6. Email
  7. Social / Content
  8. Operational improvements

If you’re in the TL:DR camp, here’s a brief summary of the steps with a much more extensive write-up on each.

The first thing to make sure is well-understood (regardless of how optimized or not it is) is the value and lifetime of a customer.  What’s the average stick rate and cancel rate of a customer and how does that vary by traffic source, partner, plan/SKU, etc?

Armed with that info, building the customer LTV/unit economics model is a natural next step.  The first step helps to understand how long customers stay around, the second step is about rolling in day 0 revenues and the cost structure to get to a gross margin value per customer.  As I mentioned earlier, getting to step 2 hopefully clearly doesn’t mean that the work on retention is done.  Understanding churn, why people cancel, particular drop-off points in the lifecycle, as well as a sensibility around the various touchpoints that can affect retention, means the beginning of a bunch of work in improving the business. More below on this point.

Once the value of a customer is known, whether in aggregate or by traffic source, it makes sense to (re)-evaluate paid media.  Hopefully, there was a reasonable sense of the value of a customer and target CPA prior to this process, but undoubtedly this framework helps to improve model assumptions as well as inspiring a revisit of how paid media has been managed.  Note that I’m grouping together all media – whether Facebook, Google, or TV on the true paid side, as well as any affiliate and influencer campaigns you may have running.

From there it’s about funnel reporting and a testing plan.  What happens when customers hit your site – what percent of them convert, what’s the average order value, what happens to those who don’t convert?  These are only a few of the necessary metrics to understand and then to test to improve.

At this point, I’ve found that it makes sense to build a customer journey map.  Whether in a simple file like this one, or as a beautiful graphic, the point is to make sure you and your team keep in mind the state (mental and emotional) your customer is in, and then how your goals compare to those of your customer.

I’m the first to say that this type of work hasn’t come naturally to me, but it’s not simply a feel-good exercise.  It should absolutely drive action.  If it doesn’t then scrap it.

From there, dig into the email campaigns, which arguably should have a significant impact on your company’s bottom line.  Remember that many customers who come from paid traffic don’t convert right away – they need to learn more about the business, the brand and otherwise.  Email is a crucial part of keeping them warms, sending offers, etc.  It’s also typically where a lot of the conversions that happen on a last click basis end up being driven.

I have a feeling that some folks may think this next section should come first, not #7 on the list. If you’re a startup, then social and content may be one of the first things you focus on.  I put organic social and content here because they are important, but for most companies with some semblance of scale, not to the extent the prior areas are.  They might’ve gotten you to this place, but as the business has grown, those prior steps becoming increasingly important to get dialed in.  That being said, I know folks who drive a ton of traffic because of the content they create.  Or others with massive engagement on their social platforms.  Putting attention on owned and earned media is a big deal.  As is building a community around your brand.  That’s a bit of what these two can do.

Finally, concurrent with this framework is work that happens on the more operational side of the business.  Working to improve the product and making operational improvements (warehouse, customer support, etc.) can and should be happening on a regular basis.

Now, Digging Deeper

  1. Retention, Cancels, Churn, & Lifetime Revenues

Regardless of the category in which you play, if you’re in the subscription business, these are some of the most important metrics to understand.

I like breaking down retention by cohorts – essentially grouping customers – by their join month at the simplest level, and then ideally by traffic source, the offer or plan they took, and any other ways you might group customers.  Within traffic source, you can of course break things down to varying levels of granularity; I would just be careful about the volume of data you have to make sure that any analysis you or someone on your team does is statistically significant.  Meaning, you might break down Facebook into not just Ad Sets but specific ads within those Ad Sets, or you might look at specific affiliate traffic.  But if you only have 20 new customers from those more granular views, then the difference of 2-3 customers can have a meaningful impact on your percentage calcs and as such, shouldn’t be trusted.

Here’s are 2 views I’ve created for a client:

For a business where the vast majority of customers are on the same frequency of renewing their subscription – whether a Netflix subscription where the concept of pausing is very much the exception or Proactiv which similarly has the majority of customers on “regular” shipment frequencies – this view is the one to start with.  I’ll provide the contrast just below.

The core here is to group customers by join month and then to look at the percent of customers who remain at each billing cycle.  The calc can be based on the percent of customers billed and/or the percent of customers who haven’t cancelled.  There’s nuance to each.  That’s why this view is more relevant for businesses where customers are on a regular cadence.  If you’re a Netflix customer and you don’t pay, then you lose access to the service, so the those who have been billed and those who haven’t cancelled (or been cancelled) is pretty much the same thing.

An important note, yours might be a quarterly billing cycle, so cycle 2 may mean 90 days out from their original purchase, as opposed to the cycle for Netflix would be monthly.  I just want to make sure it’s clear that Cycle doesn’t always equal Monthly.

We’ll talk in a moment about interpreting these tables, but I want to make sure the basics of what’s here are clear.  The filters at the top might allow you to segment customers by traffic source (FB vs. Google, or TV vs. direct mail, etc.), by initial offer (a 30-day sample vs. a 90-day sample), or campaign (this might be a promo code or special offer).  I’ve just picked a few here.  You might include sales rep to the extent that’s relevant, the type of lead they are, or the tier of plan a customer chooses.

This next view is a shipment-based one.

Since many subscription box companies allow you to pause your next shipment, looking at the percent of customers who haven’t cancelled may not give you a sense of the real value you’re getting.  If you’re a Blue Apron customer and you’ve paused your account, you aren’t cancelled, but you also aren’t active.  I know this sounds detailed and subtle, but trust me when I say these distinctions can affect how you perceive retention and customer value.

As such, this view is based on the percent of a cohort that has received shipment x, regardless of whether that was in month x or 3 months after the “normal” time.  Clearly, this has an impact on your P&L and on cash, and you have to do some of that reconciliation.  But it also does help to normalize the analysis around cohorts.

I’ve included the same filters here – clearly, those would be customized to your business as they were in the first example.

How You Can Use These Tables

For starters, they give you a clearer sense of how long customers stick around – whether x number of billing cycles, months, or shipments.  Particularly if you’re out raising money, investors want to know how long the average lifetime is.  You and your team need to know these metrics.

One of the biggest reasons that reports are important is that they help to reveal differences and changes.  How do customers from one cohort compare to another at similar periods? How does the value of a customer differ from one traffic source versus another? Does one sale rep deliver much higher value clients than the rest?

Other questions to consider:

Do you lose more customers at certain points in the lifecycle? Most trial businesses lose 30% of their customers during the trial period, for example, but might only lose another 15% in the next cycle after the trial.  But for some reason it might jump to 25% in the following cycle.  This helps to identify where something might be breaking, where customers are no longer engaged, where your team is dropping the ball, etc.  Then it’s a matter of putting time and attention towards these issues.  To state the potentially obvious, the earlier in their lifecycle the better.  It’ll have a trickle effect downstream, but there may be some quick wins you have for slightly more tenured customers that you can affect.  I detail some tactics you can use here.

-In that above piece, I speak to making sure you understand the reasons customers are cancelling.  For most subscription businesses, these typically fall under:

  • Too expensive/don’t see the value
  • Don’t use it
  • Doesn’t work / Don’t like the quality
  • I’m travelling / moving
  • Your customer service sucks and can’t help me fix problems
  • Delivery issues

Depending on the volume within your buckets, that can help determine prioritization.  And help inform ways to address.  Delivery issues and customer service are operational issues.  If someone cancels because they are travelling, you might add a “pause” feature – or at least tag these customers so that you remarket to them after 30 or 60 days.  The top 3 on that list are common and can be issues around marketing, product, follow-up, messaging, etc.  Sometimes your pricing is off, maybe your product just isn’t that good, and then the reality is that customers leave.  Ramit Sethi wrote a great piece speaking to this point.

What does the cancel curve look like? Of those who cancel, when do they do so? Are there a bunch of people cancelling on day 0 – so maybe they’re gaming a trial or promo? Or on the day you send notifications about their next shipment (which is pretty common)? Or on the day or two after the next cycle (which means they may not have wanted it and may even be asking for a refund)?

One of the reasons the cancel curve is important not just to understand but to monitor on an ongoing basis is that it can serve as an early indicator of something that has changed.  The retention analyses above are crucial, but they can be lagging indicators and analyses around critical business issues.

Look at the chart below:

Note how the cancels in the first 7 days jump from ~4% from Feb-May to 11% in the month of June? That type of behavior will show up in the retention analysis when you do it at the beginning of July, but a) it’s not uncommon to overlook the most recent month, if only because you just have less time and data; and b) those extra days between when an issue pops up, you identify it, and then figure how to fix them are costly.  If you knew you weren’t billing 10% of your customers each month, you’d want to know that as quickly as possible, right? Well, of course cancels mean fewer billings.

In this case, something had changed both with the buyer email sequence, suggesting to customers that they needed to take action when in fact they didn’t (an oversight), and with an in-box insert suggesting the same thing.  Both decisions had been intentional, but obviously the impact wasn’t.  The email was easy to fix; the insert meant getting more printed and then sent to swap out at the fulfillment centers.

Lifetime Revenues

In a nutshell, the above KPI’s have not included any discussion of revenues explicitly.  Sure, revenues are components of retention, cancels, etc. But it’s important to look at these things both in the context of counts as well as revenues.

Certain segments of customers may have similar retention rates but different lifetime revenues.  Depending on the offer, plan, tier, etc., those would be more obvious.  At the same time, some customers may tack on more to their order, whether with additional recurring offers, features, or even 1-time offers.

Regardless, as obvious as it may sound, sometimes we can get so caught up in the above analysis that is based on counts that you forget to layer in actual revenues.  To start, you might find it’s necessary to assume an average order value for all customers, but over time, just as the different filters were put in above, being able to pull in revenues for all customers is just as important.

As we wrap this section, I’ll state the obvious – this was a long and meaty section.  But in the world of subscriptions, these are some of the crucial aspects to put attention to.  Having someone dedicated to owning this part, and then working with others on the team to execute tests and changes, gives you a greater chance of knowing what’s happening in your business and making meaningful changes to it.

As I mentioned early, moving to the next step doesn’t mean step 1 is done.  None of these are ever “done.”  Rather, step 2, building the unit economics model, is a natural extension to step 1.

  1. The Unit Economics Model

I’ve written about the importance of knowing the value of your customer and having a robust unit economics model numerous times. I walk thru the actual model here and go over some additional basics here .

I’ll hit only the key points now:

  1. There are 2 primary reasons you need to understand customer LTV and your unit economics model:
    1. To manage your media. If you’re running paid media of any form and don’t know the value of a customer, you’re headed for trouble. You need to know how much you generate from a customer, what your goals are (% margin, breakeven by a certain day, etc.), and then that helps you back into your target CPA.
    2. To identify the key levers in your model and where to deploy resources for improvement, testing, etc. Knowing your baseline metrics and goals is one thing. Great companies believe that they are never fully-optimized and so have a continual testing program. But everyone is resource-constrained, so knowing where to deploy resources is crucial.
  2. Someone on the team must be accountable for maximizing customer LTV, and as a result, the target CPA. That doesn’t mean they do so at the exclusion of the brand, nor does it mean they do everything on their own. But someone must “own” customer LTV. You’re not going to get better at the pace you want if you don’t do so.
  3. Customer LTV is a combination of revenues AND costs. It’s more fun and sexier to focus on the former. But if you reduce the latter, that goes straight to more dollars you can put to customer acquisition. Here’s are a couple of my posts about optimizing revenues and costs.
  4. G&A, investments, capex, etc. should NOT be a part of your unit economics model. The point of the model is to capture marginal revenues and costs associated with those revenues. You don’t need to hire a new person for each incremental order. Sure, at some point, you do need to make those types of commitments, but that doesn’t mean they should be included in that model.

I cannot overstate how important the unit economics model is to running a performance marketing business.  It should be a living, breathing model that is updated once per month at minimum, and it should help to inform key testing and resource allocation decisions.

  1. Evaluating your media

I felt a little unsure about including this step here, as it should be an ongoing point.  But given the emphasis I place on the Unit Economics Model, it warrants a moment of discussion.

Once you have a target CPA in place – whether overall or by channel – make sure to communicate that to whoever is managing those channels, internally or externally. And to the extent the models and target CPA change, make sure your partners stay up to date.

At the same time, whether you’re managing paid traffic internally or externally, and feel like you have decent scale or not, I believe there should always be a sense of ongoing evaluation of how you’re doing and who’s doing it.  Performance marketing is one of those areas where there should never be a sense of complacency, or really of being content.  Those traits can be a harbinger of reduced performance.

If that sounds stressful, it is meant to be.

Sure, your team and partners should have room to breathe, the time to execute, the resources and support to do their best – whatever that means for your business given the constraints you have.  But as with our own lives, it can be easier to accept the status quo for too long.


Challenge your team.  Challenge your partners.  Have conversations with other marketers.  Keep a decent network of partners or possible full-time people.  The reality is that great people are hard to find, so having this sense of ongoing recruitment, while seemingly time-intensive, can keep you better connected to what’s going on *outside* of your company and keep you connected to good people and partners when you need them.  Because you will need them at some point.

  1.  Funnel reporting and testing

This can sometimes be referred to as Conversion Rate Optimization (“CRO”), but I don’t like that term because it seems to exclude revenue.  For simplicity purposes in this section, I’m going to assume that traffic is being directed to a website, as opposed to being directed to another response mechanism like the phone. And so to that end, the goal of the site isn’t simply to maximize conversion rate.  It’s to maximize revenue per visitor (again, assuming that your model is transactional on your site – if your site is a lead gen tool whereby sales agents call leads, then you just have to adapt the premise here to your model).

But even maximizing revenue per visitor isn’t correct.

Your goal should be to maximize margin dollars per visitor.

I’ve got a ton in this prior post about 6 revenue-enhancing areas to put attention to.

Funnel optimization is implicit across most of those 6 areas.

But once you have a sense of the retention rate and value of a customer, again, without ignoring those areas, it’s time to put attention to what happens when someone hits your site.  Reporting is critical to get a baseline.  Then start to focus in on the highest trafficked pages on the site, the key leverage points in the funnel, and work to identify where the gaps and opportunities lie.

I’m embarrassed to say it took me so long to get going with heat-mapping tools like Hotjar or CrazyEgg, but they’ve revealed a ton about consumer behavior and have just exposed areas that might’ve been missed.

For example, with a recent client, it was only by using Hotjar that we caught a blind spot in the checkout flow.  A ton of potential customers were clicking on the top nav on the final step of the checkout process.  As much as I’m all about giving customers what they want and allowing them to navigate, we were hurting ourselves by essentially “suggesting” they click away.  Now, there’s also a question about why people were clicking those buttons when they should’ve been ready to close.  We addressed that separately, but in the interim, we changed the top nav to a progression bar – and so instead of giving people their “check out the squirrel” opportunity, we kept them focused on the checkout process and saw a 10% lift in conversion rate.

One disconnect I often see with funnel optimization is a lack of communication between customer support and the retention team with whoever is running front-end tests.  Hearing the reasons that people are contacting customer support and why they are cancelling can be really helpful to inform front-end tests.  Which is another reason I believe in starting this process from the retention analytics side first.  Take what’s already happening, and then beginning your way back upstream.

Another area where I see marketers slip is in thinking the funnel work is done once the customer hits the Buy Button.  But what happens on the Thank You Page, the order confirm email and beyond are just as important.  As is making sure that there is clear accountability within the organization for these parts of the consumer experience.

Whereas some people think the funnel ends once the customer buys, others think onboarding begins after they buy, when in fact the onboarding process begins once the customer engages with you.  To that end, a great site with a ton of free content is UserOnboard.  They do an amazing job not just of showing what others are doing (they are primarily focused on SAAS businesses), they mark-up their screenshots and provide commentary to identify what they think is being done well and where there are areas of opportunity.

The point being, yes, put time and attention towards optimizing the “traditional” funnel from ad through to purchase.  But also be intentional about the path you want to take people down once they have shifted from lead to customer.

  1. The Additional Value of a Customer / Member as well as Customer Journey Mapping

After digging in on analytics, retention, and some core components of customer acquisition, I like to take a step back to revisit the customer journey and how customers think of the business.  We’ll get more tactical after this with email and social/content.  As much as I could argue starting with this, oftentimes getting grounded and into these prior areas is a more natural way to understand and to start optimizing the business.  Then taking some time to look at the business in a more holistic approach can be a good thing.

I’m somewhat stealing the term of “Member” from Robbie Kellman Baxter, who wrote a phenomenal book called “The Membership Economy” – I highly recommend checking it out.

Now of course she didn’t invent the word or concept of a “Member”.  But she has done a great job of helping people shift thinking about customers to members.  Think about it for a moment – isn’t a customer different than a member?

Do your customers think about your business as one they subscribe to, or do they think of themselves as being a member of your business or service?

“I get a subscription box from this company” versus “I’m a member of their program” or “I’m one of their members.”

American Express hasn’t used the “Membership has its privileges” campaign in years, as Robbie points out, but you still have the word “Member” on your card, regardless of what color that card is.  They’ve done a phenomenal job not just of building a loyalty/point program, but in giving their members real value.

Most recently, they partnered with Shoprunner to give all their (AmEx’s) members free 2-day shipping from a variety of online retailers.  AmEx knew it couldn’t do this on their own, and this is one of the big reasons people order from Amazon, but this is a huge tangible benefit of paying for that annual fee, which is still called a “membership fee.”

Ultimately, a member expects a certain type of treatment, expects that you’re taking care of them, giving them value above and beyond what they might’ve signed up for.  Whether you’re a subscription box company or a software provider, what are the ways you can add value and make life easier and better for your customers?  Some of this might take the form of enhancements to the actual product, or it certainly could extend beyond.

Before jumping in to these additional areas, I’d be remiss to not state what I think might be obvious – there are a bunch of things that should be “check-the-box” items, which you just have to nail.

  • Are customers clear about what they are getting?
  • Do they get the necessary information about what they’ve ordered quickly (order and ship confirm emails, for example)
  • Does the product or service match what they were promised?
  • If it’s a physical item, does it show up when they expect to and without delivery problems?
  • Can they easily find information about their account?
  • When they reach out to customer support, do they hear back in a timely manner?

These are the basics, but anyone who has had issues with the above can attest how hard it is to run your business when the basics aren’t being nailed.  So make sure those are being nailed.

I’d also say that the bar continues to be raised about what people expect.  As much as I’ve detailed some additional areas below in which to add value to your customers, increasingly, these are being included in your customers’ “check-the-box” category, even though you might not think of it that way.

Here are some areas and examples which you might utilize to build out your membership program.


  • Value-added services (0% margin offers for example)
  • Member versions of promos – perhaps in the form of a bonus, an opportunity to get the bonus for a referral, or otherwise
  • Special offers
  • Buy credits at a discount

Outside Offers

  • Gifts/benefits that aren’t explicitly tied to the business’ products


  • Loyalty / rewards program – including rewards for longevity and continuous longevity
  • Acknowledgment / member recognition
  • Physical card
  • What are the benefits of a higher-tier status (silver, gold, platinum) – pricing, shipping, concierge service
  • Unexpected surprises


  • Connect offline / in-person
  • Community
  • Give them opportunities to participate
  • Access – to FB lives, groups, offers, etc

Something else I’ve found helpful – and I’ll admit that this felt too “traditional marketing” when I first started doing it – has been to map the customer journey.

You can see an example of what I mean here – this isn’t a beautiful graphic but it does serve the purpose of illustrating the more traditional steps customer go through (unaware, aware, consideration, purchase), but I’ve also added in a few sections:

  • For a subscription box business, what are the sub-steps for an Active customer?
  • Where does the interaction take place?
  • What are the company goals at each step?
  • What are the customer’s goals at each step?
  • What is the customer’s mental and emotional state at each step?

For example, once an order is placed, it’s normal practice to send an order confirm email. As the marketer, in addition to making sure the customer gets the necessary confirmation, we might want to give them account details, info about our content, give them additional offers, etc.  But generally speaking, the customer wants to know the order went thru and to move on.  That doesn’t mean that some of them aren’t open to buying more, reading your content, etc., but I think it’s helpful to keep in mind that the customer is very likely busy, might be excited or might even be experiencing buyer’s remorse.  And so, understanding that can help to bridge where they are with our goals.

Everything for me is about driving action.  It might seem that this is an exercise that is just “nice to do,” but it should in fact lead to action, lead to changes in how, what and where you communicate with your customers.

And especially if you want to start moving from having customer to members, it’s critical to be mindful about where they are, not just in their customer journey but in their life overall (hint: they are super busy, want stuff done as easily as possible, and want you to be more than 1 step ahead on servicing them even better than you might expect they’d want.)

  1. Email

For some, email feels outdated and unfortunately is ignored.  For others, email represents the source of 30% of their revenues.

Email is a big deal and one of the primary means that marketers use to communicate with their customers, separate from their blog, social or in the product itself (whether as a physical insert or through an app or software).

There are a couple components of analyzing what’s happening with email:

  1. What’s being done to capture emails?
  2. How effectively are they being marketed to?

In terms of email capture, I’m not a big fan of the pop-up 10% discount offer than shows up on so many sites within seconds of hitting a site. I think it’s a crappy customer experience.  That being said, you have to test what works for you.

What I can say is that it’s worth looking at the most trafficked pages on the site to make sure there are at least 2-3 different ways people can opt in – those might include a newsletter sign-up, an optin for a free e-book, and and exit intent pop.

In terms of monetizing the list, most people can get so concerned about unsubscribes that they become too cautious on hitting their list.  That doesn’t mean that each email has to be an offer, but certainly even if it’s primarily a content-based email, there should be some link or subtle offer driving them to your site.

The early-engagement series, both for buyers and prospects, is particularly important.  Most people are the hottest they’ll be for your brand in the first 7-10 days.  So getting in front of them, whether to make sure they get and stay engaged, or to show them the value of your product to get them to purchase, is critical.

I do want to make a small note about the long tail.  I work with a client where more than 60% of their revenues come from customers who’ve opted in more than 30 days ago.  They have a higher-priced service, so it makes sense that people need time.  But this just emphasizes the need to keep (engaged) people on the list, as well as the importance of content and brand.  When customers take longer to convert, it’s even more likely that your brand is going to have an impact on their chance of success.  So, too, will content – more in a moment on content, but keeping using it to keep your email list engaged can have a big difference on your bottom line.

Finally, remember that solid open rates on emails can be 30%.  Just because someone hasn’t opened or clicked on what you thought was a great offer doesn’t mean they aren’t interested. There’s always a component here of playing the numbers.

  1. Social / Content

Social and content are great ways to increase engagement and connection to your brand.  They are also natural ways to provide value-added content to your audience, whether as an acquisition or retention tool.

Content marketing is no longer a buzzword, it’s what people are expecting. And it can be delivered on your site via a blog, thru email, and on social.  And with sites like Problogger, Upwork and Craigslist, it is getting cheaper to source great writers and editors to help craft quality content for your business.

Plus, don’t forget about soliciting user-generated content (“UGC”).  Sometimes you just need to ask for feedback or a video – whether that helps to inform the product or service, or it can be used as content to market with (make sure you get someone’s okay before doing so), taking advantage of content that is educational, inspirational, and/or entertaining is a big plus.

  1. Operational Improvements

I’ve intentionally focused on the marketing side of the business in this piece.

A few words on Product and then a couple more with an additional link on Cost Optimizations.

Product – I cannot emphasize how important product is.  All the marketing, analytics and optimizations are a ton easier and more effective with a solid product.  And just that much harder with a bad one.  SAAS businesses seem to have a culture of continual improvement, in the form of releases, but many in the physical products world rarely have that culture as a core part of the business.  If you’re in this world, I’d suggest taking a cue from SAAS businesses.

Customers expect that their software or app will constantly be updated.  How many updates have you gotten for your smartphone apps in the past couple of days? Likely a bunch. So why don’t physical products have anything close to that cadence?  Well, for starters, it’s easier to push an update live for software than to do the analogue in the physical world.  And yet that doesn’t mean they aren’t possible.  Please put some attention to do so regardless of your business.

Finally, cost optimizations.

If you’ve made it this far, then I’ll simply suggest you check out this link where I detail some of the core cost-related areas to optimize.

In Summary

Phew, there’s a ton here.  There’s also a ton to your subscription business.  There are a lot of areas to target.  But my final suggestion is to make sure to key in on the main leverage points in your business.  Everyone is resource-constrained so making sure you and your team are focused on the biggest opportunity is crucial.

Hopefully, working your business is a lot of fun.  There are plenty of rough patches and frustrations in any business, but the sheer number of ways you can affect your customer’s experience means there’s never a dull moment and plenty of places to make a difference.

I hope these thoughts help you to improve your business.

And if you have any questions or thoughts, I’d love to hear them.

10 Ways to Increase Your Subscription Stick Rate


Whether you think of yourself as a strong or a weak marketer, it’s likely you are not spending enough time working to keep your customers around longer.  Most folks prefer climbing the next peak.

And yet, you’ve heard it a million times.  It’s a ton easier to sell more to an existing customer than to a new one.

In previous posts, I’ve talked about the fact that every business is one way or another in a subscription model.  You either have customers on recurring billing or are “just” trying to sell them again after their initial transaction.

In this post, I want focus on the traditional subscription (continuity) model, where a customer has signed up for a product or service that has a recurring billing component.  That might mean monthly, quarterly or annually.  The point is that every so often, a customer re-ups with you.  That also means that every so often, they have an explicit opportunity to cancel or ask for a refund.

When I work with marketers about increasing their customer LTV and subscription stick rates, the below list details the top places I dig in.  To make sure we understand what is happening at each level, to set a baseline (if necessary) of where they are, and to help inform prioritization of testing to improve that component of the business.

  1. Do you actually know your subscription / continuity take rate and your stick rate on an overall basis?

This is a simple (note, not always “easy”) one.

Sure you can go crazy by splicing and dicing your numbers countless ways.  By cohort. By channel or front-end campaign.  Gross or net of returns.  By customer count and by revs.

But start with the basics. Of the people who start each month, how many are around for the next few billing cycles. Where are the biggest points where customers drop off? Just starting there will reveal a good bit of info and insight to help inform what you should be testing.

For those of you who are already doing so, hopefully I just gave you a few more ways to think about it.  You can always get more sophisticated about how you segment your customers.  And likely you’ll want to do different things for different segments.

But without a base level of understanding of where you’re at, it’s hard to know what to do next.

  1. Do you know the top 4 reasons customers are cancelling or asking for a refund?

Most likely the list includes “too expensive”, “don’t use the product”, etc.  But you might have something totally different on your list, for example that your product is too complicated to use. While no one wants to hear that type of feedback, I’d always rather know that than not.  Because depending on the order of the most prominent reasons, you can decide what to test next and how to tackle each reason.

  1. Do you have a cancel / save offer, script, and training when customers want to cancel?

Next, what do your agents say when a customer calls to cancel? Do they have specific guidelines or an outline to follow? Have you tested (and measured) different strategies? Ultimately, it’s a business decision on what kind of offers you might want to give to a customer.  I’ll say that it drives me crazy when the mobile phone carriers offer me a big discount when I’ve called to cancel.  Why did it take my wanting to cancel to get such a good offer?  For your business, the scripting might include something as simple as probing why a customer is cancelling.  Sometimes there’s a lack of information or training the customer needs.  Whether or not a rep can keep a customer, at the least, that information needs to be documented and passed along to the appropriate person (See #1 above and #10 below).

  1. Do you have customer care agents or sales agents handling cancel / saves?

This is again another business decision, but if you want a true save strategy, I’d suggest that you test your sales agents, not your customer care agents, for this task.  Sales vs. care agents are totally different people.  Both should come from a place of providing value and service to a customer.  But care agents generally have a softer demeanor, while sales agents are typically more skilled at overcoming objections, speaking to features, benefits, and value of an offer.

  1. Does the product actually meet the promise you made when the customer signed up in the first place?

This one sounds obvious, but beyond making sure your product is a good one, this point goes to real expectation matching.  What does your customer actually think your product or service is going to do? Does their experience match that expectation?

Note, beware if you find yourself saying, “well, if only the customer watched the 20 minute tutorial or did x, y, or z, they would know what to do.”  To those folks, I’d ask what guidance are you giving them to watch the tutorial or to know to take the additional action.  At the same time, if your product is touted as “simple to use” and “you can get going in only 5 minutes”, then arguably it shouldn’t take a 20 minute tutorial to do so.   Or if your site is so confusing that customers who are in fact looking for information can’t find it, I’d posit that your “simple” brand promise is not being delivered upon.  Remember it’s not solely about a customer’s explicit use of your product or service.  But rather their entire experience, from order entry to FAQ’s and customer support.

  1. What info do you put on the credit card descriptor?

You may or may not be as anal as your customers, many of whom look through every line of their credit card statement.  Even though your customers may love you, if what shows up on their credit card statement seemingly has no relevance to the brand they interact with, their chances of challenging the charge and/or calling you to cancel will increase.

This kind of thing can happen for different reasons.  Sometimes, a marketer has a business name that differs from the various products she sells.  And so it’s simply easier to put the company name on the descriptor.  This isn’t exclusive to folks intentionally trying to hide their brands from the company name – my guess is that anyone reading this is a purchaser of something from P&G (Proctor and Gamble) but doesn’t explicitly realize that P&G is the parent company.

Take a revisit of what appears on the credit card statement from a brand new customer’s perspective and think about whether or not it would confuse you.

  1. What happens when a customer’s credit card declines?

Do you have a process to retry, use Account Updater, or to possibly call a customer to let them know their credit card has declined? At the least, testing the frequency and overall quantity of retries is something all marketers should do. Whether that means every day for a week or two, or once a week for 4 weeks, the reality is that the cost per attempt is something you or your team should know (let’s say it’s roughly $0.10). Then it’s simply a matter of comparing the necessary success rate vs. your price point and margin.

Account Updater is something I’ve discussed previously, but the short version is that some banks provide a paid service to update the expiration date of a valid card but one that has simply gone past its expiration date.  It’s an extremely rare customer that will call to update their expiration date.  The beauty of the service is that you only pay per successfully-changed card.

Finally, depending on your price point, margin and call center fees, it might be worth exploring calling customers for whom their card has declined.  By the way, one measure of knowing whether you are proud of your product/service or just trying to sneak under the radar on their bill is how you respond when you consider calling a customer.  If your product is actually of true value, then while yes, some customers will take the call as an opportunity to cancel their subscription (how is this worse than the declined credit card by the way?), you should feel a responsibility to make sure your customers have uninterrupted access to your product.

  1. Do you send an email prior to a new billing period and offer additional products?

This might not seem obvious to extending stick rate.  How would notifying people of an upcoming charge and/or offering them more to buy improve retention?  Like many others out there, I’ve signed up for Dollar Shave Club and must say that their pre-shipment emails are some of the best I’ve seen.  Not only am I informed about my next shipment and given the option to delay it, but they have a very slick one-click tool to add additional items to my order, shaving cream for example.  I may not have needed new blades, but if I needed shaving cream, that keeps me on as a customer.  At the least, they’ve added to the customer LTV.  Best case, because I continue to think of myself as a Dollar Shave Club buyer, I’m more likely to keep my subscription active vs. cancelling altogether.  Which either means new acquisition for them or reactivation, neither of which is easier than maintaining an existing customer relationship.

  1. Do you allow customization?

I alluded to this, but it deserves its own call-out.  Increasingly, customers expect to have some level of control on when and what they receive.  We no longer live in the world of one-size-fits-all.  Customers may really like your product but aren’t using it at the pace you suggest.  Or perhaps they have a busy travel schedule for the next month.  Or maybe they use some of your products dramatically more than others (especially relevant for those in the beauty category).  It’s preferable to give people the ability to customize their order online.  But even if you market it online and only your reps can make the change, I would encourage you to test how much flexibility to give your customers.  This is also where using customer feedback and cancelation reasons can help to inform where you start.  If the number one reason people cancel is that they don’t use it (assuming it’s a consumable) as frequently as suggested – beyond revisiting the amount you’re shipping, how you help them consume the product, etc. – you might want to consider varying the quantity or frequency of their orders.

  1. Do you have someone accountable for retention marketing?

Whether you are a solopreneur or running a 9-figure business, if you have a subscription component of your business, someone should be working to improve it.  It might be the difference between getting to the point where you can stop being a solopreneur and can hire someone.  Or it might mean adding even more margin to your business.  Regardless, do the math.  How many people do you add onto a subscription each month?  What’s the stick rate.  How much do you have to move the needle to justify more attention or a hire to improve the value of those customers?  Again, it might be just you for a while.  Or it might just be a no-brainer to bring someone one.  But make sure it’s an informed choice you’re making.

As one of my friends says, there’s never been a more difficult time to acquire a new customer.  And yet, too often we keep focusing on getting new ones at the expense of maintaining our existing ones.  With a bit of process and working at it, there are ways to keeping your customers around and happy a lot longer.  But it requires attention and some testing.

Now it’s a matter of taking this playbook and applying it your business.  The rewards are worth it.

Serving the Underserved Creates THEIR Experience

In a literal way, inclusivity is the opposite of exclusivity.  And so, while being exclusive means something is for a select group, inclusivity, particularly for marketers, doesn’t necessarily have to about trying to include everyone. At times, companies can simply target people who have been underserved.  Especially when we think about one of the big no-no’s in marketing – that of trying to be all things to all people – it’s important to understand this distinction and not make the presumption that inclusivity is going against that adage.  The below examples demonstrate how engaging and enrolling audiences that have been ignored, in a sense crafting a great experience for the underserved, can yield great results.

Sephora: Clueless Shoppers are Welcome

Plenty of industries are filled with employees who like to make sure (potential) customers know just how “knowledgeable” the employee is, often at the expense of the customer.  For my part, I grew up a cycling fan and wouldn’t have enough time to list off the number of times I felt so belittled because an employee in a cycling store was seemingly offended by what he considered an offensively-simple question.

The beauty industry has a similar dynamic.

For its part, the beauty brand Sephora has done an exceptionally good job of making the beauty-shopping experience actually feel inclusive, not snobbish. Blogger Alicia Jessop has described Sephora as the equivalent to women of what a hardware store is for many men[1]. Just as men often go into hardware stores and browse without having any specific product in mind, women often shop at Sephora with that similar sense of discovery, exploration, and accessibility.

It’s fair to say that most everyone has gone into a store where there was a certain product domain they knew nothing about, then either tried to portray they knew something they didn’t, or they didn’t want to acknowledge that they knew nothing.  Unless we are just plain lucky, it’s a rare occasion that we leave with the right product and certainly none the wiser.

One of the primary reasons we as consumers behave this way is an expectation of how the “expert”, in this case a store employee, is going to respond.  Prior experience has led many of us to be intimidated in environments where there is a lot of domain expertise or technical skill.

But Sephora, while obviously not perfect, mitigates this effect in several ways.

The most obvious way is the store layout.   Unlike department stores where all the merchandise is behind a counter and only accessible by employees, Sephora has put product front and center, so that customers can touch and try products without having to ask anyone.  There is no gatekeeper to get past or to ask what might be an ignorant question.  Instead, customers have free reign to products at Sephora – to touch, hold, and try at their leisure.  This may seem like a minor detail, but given the department store dominance in beauty and just the fact that a customer doesn’t have to speak to someone to try something that grabs their attention – these have been big changes in the beauty industry.

Next, when customers do need help, Sephora staff are there to serve them.  Training is an important aspect of Sephora’s employee onboarding, whether on the customer service side or the technical side (how to help a customer select makeup that suits them as well as applying it).  Training is done at the store level but also through Sephora University, the center for the company’s training.  While creating its own training center, aka the University, is an advantage that a massive company like Sephora can afford to do, putting attention to how employees treat customers is something that any business can do, whether big or small.

It’s important to acknowledge that no business is remotely perfect. In researching Sephora, there were plenty of examples I heard about sub-par, off-brand experiences.  Particularly in retail, where the inherently imperfect human interaction drives a good part of the experience, perfection isn’t the goal.  Being a ton better than everyone else is.  On a consistent basis.  At the same time, setting up aspects of the business, such as the store’s design, that are not dependent on how an employee chooses to behave, can shift that burden and reinforce the bigger experience that is trying to be delivered.

As we all know, a company’s brand isn’t simply created from one component of the business.  It is built everywhere.  And in Sephora’s case, creating a welcoming, interactive and fun environment, whether through the store design or its employees’ attitudes, has been a key to their success.  It affects customers in the store, their likelihood of making a purchase, and certainly the stories they tell (primarily positive in Sephora’s case), which drives the chances they return and/or how they influence others to do so.

Gwynnie Bee: Serving Plus-Sized Women

When the apparel and accessories business Stitch Fix launched, they initially didn’t target plus-size women (anywhere from size 10-14 and up, depending on whose arbitrary definition you want to choose from).  It was a business decision that Stitch Fix (a former client of mine) made, amongst others that also included not serving men or kids, all of which they do now.

This isn’t about right or wrong for Stitch Fix, but their success and focus on its core group of customers, meant that there were opportunities for others to target those that weren’t being served.  Stitch Fix was not the only subscription business focused on delivering these same products to women, but in being one of the earliest and certainly the largest (most recently topping $1 billion in revenues and 2+ million active clients), who they were and were not serving was much more visible.

So as Stitch Fix was building its business, and to a certain extent educating the broader market on what they offered, Gwynnie Bee launched at a similar time, exclusively serving plus-sized women. Not that Stitch Fix played into the stereotypes that are all-too-often found in beauty magazines, but the fact that they didn’t stock apparel for plus-size women gave Gwynnie Bee the opportunity to highlight how it was differentiated from the category leader in Stitch Fix.  Gwynnie Bee was specifically offering clothes for a group that are not always served by apparel retailers.  To this day, beauty magazines constantly reinforce the idea that skinny is better.  To its credit, Gwynnie Bee embraced its target customer, and didn’t try to show size 2 women on its site.

Body image is admittedly a sensitive topic; and yet from a practical side, the reality is that many women in the US fall under the plus-sized definition.  So while Gwynnie Bee was targeting an underserved and perhaps niche category, theirs was by no means a group with small numbers.

(It’s interesting to note that Stitch Fix eventually did serve plus-size women, in addition to men and children.  And Gwynnie Bee announced in early 2018 that to be considered truly inclusive, they were not remaining exclusive to plus-sized women.)

There are plenty of other examples on servicing what isn’t considered mainstream; certain shops target “big and tall” men, some restaurants are “family friendly,” while others are clearly designed for couples.

In all of these cases, whether Sephora, Gwynnie Bee, or others, inclusivity isn’t about trying to be everything to everyone. It’s about targeting a particular group of people, especially those who may be underserved.  And then making them feel warm, understood, and served in the manner which we’d all want to be treated.


[1] Jessop, Alicia. “What’s Good Wednesday: Why Women Love Sephora.” August 22, 2012.