Hey Startup enthusiasts,
This is Read Startup. Twice a week, we pick one company's boldest call and take it apart so you can see the machinery underneath.
- Today's case: Blue Apron
- The move: Going public as a subscription company while barely using the word "subscription"
- The puzzle: The acquisition cost in the IPO filing looked cheap, and the real number was several times higher
- Your takeaway: How to calculate marginal CAC and payback so blended averages can't hide a leak
- Read time: 5 minutes
Let's see what we got.
In June 2017, Blue Apron filed to go public with a number that made its model look lovely. The company said it spent an average of $94 to acquire each customer.
For a business doing nearly $800 million in yearly revenue, $94 sounded like a bargain.
That average was quietly blending two very different companies together. Over the twelve months before the IPO, Blue Apron was spending closer to $463 in marketing per new customer. Same company, same product, a cost five times higher once you looked at the right window.
Blended CAC is a comfort blanket
CAC (customer acquisition cost) is total marketing spend divided by new customers. Simple enough.
The trap sits in the time period. Blue Apron's $94 figure averaged 2014 through 2016, including the early years when word of mouth and cheap Facebook ads did most of the work. Its marketing budget climbed from $14 million in 2014 to $144.1 million in 2016. Each extra dollar was reaching people who were harder to convince.
Founders call the number that matters here marginal CAC: what your next customer costs, today. A three-year average flatters every scaling startup, because early adopters were the cheapest buyers you will ever get.
The bucket had a hole
Expensive acquisition is survivable if customers stick. Blue Apron's did not.
Daniel McCarthy, a marketing professor at Emory, modelled the S-1 data and estimated that close to two-thirds of customers left within six months. His math put gross profit at roughly $25 a month per customer, which meant about four and a half months just to earn back a $94 CAC.
Put those two facts side by side. Payback took around four and a half months, and the typical customer churned at around six. That leaves a sliver of profit per customer at the old CAC and a deep loss at $463.
But here is the rough back-of-envelope version. Six months at about $25 of gross profit is around $150 per customer. Subtract a $94 CAC and you keep roughly $56. Subtract $463, and you are over $300 in the hole on every new signup, before warehouses, staff, or software.
No amount of top-line growth fixes that equation. More spending simply buys more losses, faster.
Investor Nikhil Basu Trivedi spotted a telling detail in the filing too. The words "subscription" and "churn" barely appeared, and the S-1 offered no cohort retention chart.
Why meal kits churn by design
Netflix fits into an evening you already had. A meal kit asks you to plan, cook, and clean on a schedule every week.
The product sold novelty: new recipes, new skills, and a fun box on the doorstep. Novelty has a half-life. Once you have cooked your fourteenth Korean beef bowl, the reason to stay fades, and the price per serving starts to look high next to a supermarket run.
Blue Apron's best channel hinted at the answer. Its referral program brought in 39% of new customers in late 2016. People who arrived through a friend tended to be cheaper to acquire, yet paid TV and direct mail made up the biggest share of spend.
The market did the math
Two weeks before the IPO, Amazon announced it would buy Whole Foods for $13.7 billion. Blue Apron cut its price range from $15 to $17 down to $10 and is listed at a valuation of about $1.89 billion.
Amazon was the headline. The unit economics were the slow burn. When the company pulled back on marketing in late 2017, its customer base shrank fast, which told investors the growth had been rented.
In 2023, Wonder Group agreed to buy Blue Apron at an equity value of roughly $103 million. That works out to around 5% of the IPO valuation.
The 2025 plot twist
In August 2025, under Wonder, Blue Apron dropped its subscription requirement and let people order à la carte, with an optional autoship discount and a paid membership.
That is a company finally pricing for the way customers actually behave. Plenty of households wanted meal kits in some weeks and takeaway or groceries on others. Forcing a weekly plan raised churn and made every cancellation feel like a breakup.
Founder takeaways
Every founder has a lot to learn from Blue Apron’s move. To summarize it in a nutshell:
Track marginal CAC every month. Divide this month's spend by this month's new customers by channel. If it climbs faster than revenue per customer, slow down before the averages notice.
Put payback next to median lifetime. If payback takes five months and half your users leave by month six, your growth engine is barely break-even. Show both numbers on the same slide.
Run the six-month LTV test. Multiply monthly gross profit per customer by the average months a customer stays. If the answer is under three times your marginal CAC, fix retention before raising your ad budget.
Build the cohort chart before investors ask. Plot what share of each monthly cohort is still active at months one, three, six, and twelve. A flattening curve is your real moat.
Double down on the cheapest loyal channel. Referred customers often cost less and stay longer. Fund the channel that brings sticky users before buying reach.
Forward this to a founder who just told you their CAC is "only" $50.
600 1st Ave, Ste 330 PMB 92768, Seattle, WA 98104-2246
Unsubscribe · Preferences
