Blue Apron's Q3 '18 Results: CAC moves higher, challenging retention trends remain the same, and BA gets religion

Blue Apron released their Q3 2018 earnings today. As chance would have it, today was also the day that I taught my Blue Apron case study to the MBA's in my Marketing Research class here at Emory University. Needless to say, I had a perfect excuse to dig deep into their latest results...

Much has been said about the top-line figures (e.g., revenues are down 28% year-on-year, while customers are down 25%), and the fact that they shed 4% of their workforce to rightsize the business, so I won't comment on them here other than to point interested readers to commentary that other people have written (PYMNTS, Reuters, WSJ).

Instead, I thought it would be of value to take a step back and re-examine the big picture from a customer-based corporate valuation perspective, and look at the new results within this context. To do so, I re-ran the model that first broke the story on Blue Apron's customer retention and customer acquisition cost (CAC) issues, updated to include all of the most recent data. I believe this is the backdrop we need to think through the new strategic direction that Blue Apron emphasized this quarter. Hopefully this will bring everyone up to speed with how things look at Blue Apron from a customer-based perspective, and what it implies for where we are going from here.

The Model (and validation)

I re-ran exactly the same model described in the previous analysis of Blue Apron, using an extension of the methodology described in my Journal of Marketing paper with Peter Fader and Bruce Hardie. In a nutshell, I assume that customers are acquired over time, remain with the firm for some time, place orders while they are alive, and spend some amount on each order. I then simply let the data tell me what acquisition, retention, CAC, order rate, and spend per order patterns (allowing for simple trends over calendar time and cohort time) are most consistent with the varied data that Blue Apron has disclosed over the years.

This relatively simple model continues to do a reasonable job of fitting the observed data, including customers, orders, and revenues:

Of course, I also continue to incorporate Blue Apron's previous disclosures on cumulative spend per customer and cumulative customers acquired.

The goodness of these fits continue to leave me feeling confident that the model I am positing is consistent with what we have observed so far at Blue Apron.

So what does the fitted model tell us?

Retention: hasn't changed

Retention remains the same. The inferred retention curve is generally consistent with what I had inferred before:

As before, I infer that about 70% of all acquired customers churn after they have been with the firm for 6 months. It is hard to build a durable business with a retention rate this low.

At the same time, if there continues to be one silver lining, it is that there does appear to be a "hard-core loyal" segment which continues buying a long time into the future. I discussed this in depth in the case study I co-wrote on Blue Apron with Eric Schwartz a few months ago. It explains in detail the "tale of two loyalty segments" that currently exists at Blue Apron. More on this below.

CAC: was high, has moved even higher

The other major trend I had noted in my previous analysis was that Blue Apron was spending a lot more to acquire customers. It appears that this trend has only gotten worse. Below is a chart of Blue Apron's customer acquisition cost over time, presented on a rolling 12-month basis to smooth out seasonal fluctuations:

I estimate that Blue Apron's average customer acquisition cost in Q3 2018 was $148, up from $133 in the year-earlier period and $90 two years prior. This will make all the more difficult to generate a return on their new customer acquisition budget, before even thinking about covering the fixed costs of the business.

These CAC trends are also entirely consistent with a footnote that was buried within Blue Apron's Supplemental Materials:

"The Customer payback in the year after acquisition has extended in recent cohorts as increases in the Net Contribution per Customer have been more than offset by increases in the Cost per Customer.

In other words, the average CLV of recent cohorts has been going down because CAC has been going up, as we can see above.

Blue Apron finally focuses on customer value segments!

To me, by far the most interesting development this quarter was what CEO Brad Dickerson had to say about Blue Apron's new strategic direction. As noted in their press release and further supported in their investor presentation, they have refocused their attention on building their business around their very best customers. They presented two figures in support of this strategy, looking at the one-year performance of customers acquired in Q3 2017, to emphasize the disproportionate amount of value created by their very best customers:

They are basically saying what I have been saying since my very first analysis. When Blue Apron acquires a new cohort of customers, a non-negligible proportion of those customers are actually very valuable. What is holding the company back is the fact that everyone else is highly unprofitable.

Note too: the chart on the right implies that their (undiscounted) payback period is over a year. In one year, they get 2.5x their CAC on 30% of their customers, and 0.2x their CAC on the other 70% of their customers. Weighting it out, that cohort as a whole was at 89% of CAC one year out. That's not good, especially when we consider (1) their WACC is probably 20% right now and a proper break-even analysis should take into account the time value of money, (2) their "net contribution" profitability only deducts COGS ex-D&A, and not any other variable costs, and (3) as noted above, future cohorts are worse than this. This continues to reinforce the concerns I have been raising since the day after they filed their S-1.

In Exhibit 5 of the case study that Eric Schwartz and I wrote, we fit Blue Apron's retention curve (using Second Measure data) to a model that assumes there are two customer segments at the firm -- those with low retention and those with high retention. The resulting fit of this simple model was excellent:

Taking this a step further, we then estimated the average customer lifetime value after CAC for each of these two segments (Exhibit 9 in the case study). Blue Apron's conclusion above is entirely consistent with what we noted in the case study.

While it took a stock price drop of nearly 90% and the removal of their CEO and CMO to reach this point, it appears that Blue Apron may finally have "gotten religion", and intends to make a full-fledged pivot towards a CLV-oriented, customer-centric business model (SEC):

The company will prioritize customer segments within the direct-to-consumer business that exhibit certain attributes, including: 1) proven retention; 2) strong affinity for the brand; and 3) high potential to increase engagement with Blue Apron product offerings.

The top 30% of Blue Apron’s customers on a net revenue basis acquired in recent cohorts account for more than 80% of its net revenue from such cohorts in the year after acquisition and had an average payback on the acquisition cost per customer of less than six months. This presents opportunity to deepen engagement with this “best customer” segment and unlock value from consumers with similar attributes. Blue Apron intends to concentrate its innovation and marketing efforts on serving the needs of its best customers and attracting more of them, leveraging its extensive insights into their behaviors, goals, and preferences to customize the customer experience—through product development, brand messaging, and exclusive services—for this valuable segment.

 This echoes a comment that Eric and I made on page 3 in the Teaching Note which accompanied our case study:

Philosophically, I agree with this strategy. Find what the acquisition characteristics are of the very best customers to find more like them. Don't over-invest in customers who are unlikely to stay around for the long term. In my view, though, Blue Apron will also want to be mindful of the following as they go about the rightsizing of their business:

1. By expanding into the Linden NJ facility, Blue Apron took a step towards "going big" and operating at scale. This will limit their ability to downsize efficiently and profitably, raising the question of how they will fill all of the excess capacity they currently have.
2. While Blue Apron continues to speak positively about expanding into grocery stores on a non-subscription basis, I do not believe they are a panacea. In fact, this move may destroy value. For more on this, see my article on eMeals versus Blue Apron.
3. While grocery stores and other players in the food space may not make for the best partners, these players are aggressively moving into meal kits. These players could operate at a scale which could more easily absorb Blue Apron's excess capacity, benefit from the brand's value, and from Blue Apron's expertise. I would look for a buyer before being forced to do so because of a strained balance sheet.
4. Over time, customer data may matter less as and when Blue Apron does expand its CPG-like business. CPG businesses don't have nearly the visibility to the end customer that direct-to-consumer businesses do.

To be continued.

If you're interested in reading more like this, or getting a "CBCV check-up" of your own, check out my company, Theta Equity Partners. We would love to hear from you.