Munger's Follies: Blue Apron (Part 2)

What Could Have Saved It

In Part 1, I walked through what killed Blue Apron: a $460 customer acquisition cost hidden behind a $94 blended average, 72% churn in six months, a product that required permanent behavior change, no moat, a discount spiral, and an incentive structure that rewarded the CEO for building faster instead of building better.

The company sold for $103 million in 2023. A 95% decline from IPO valuation.

Now the harder question: could they have survived?

The Hindsight Problem

I have to be honest about something before I play Monday morning quarterback.

It's easy to dissect this from 2026. I have perfect hindsight. I know HelloFresh won. I know Amazon bought Whole Foods. I know the unit economics never worked. Every insight I'm about to share comes from the survivor's playbook, read backwards.

SpaceX was one rocket failure from bankruptcy. If Falcon 1's fourth launch had failed, there would be no Elon-the-rocket-genius story. There'd be a case study about a tech billionaire who burned his fortune on rockets. The lesson would be about hubris, not persistence. We'd all nod and say we saw it coming.

So I hold Blue Apron's leadership with some grace. They were building in a category that didn't exist. They made real decisions with incomplete information under real pressure from investors who wanted growth numbers every quarter.

That said. I still think there were moves available.

The Survival Playbook

Cut the cook time to 15 minutes as the core product. I can't prove this one would have worked. But the competitive set tells the story. Blue Apron was competing with DoorDash, Chipotle, and a $5 rotisserie chicken. Those take five minutes and zero skill. A 45-minute gourmet experience is a Saturday night product, not a Wednesday night one. The 15-minute meal should have been the default from day one, with the 45-minute kit as a premium weekend upsell. That single change might have bent the churn curve, because the behavioral ask drops from "become a home chef" to "heat and assemble something better than takeout."

Kill the subscription rigidity. Every week a customer gets a box they didn't ask for is a week closer to cancellation. Make pausing effortless. Let people order on demand. HelloFresh figured this out. Blue Apron treated the subscription model as sacred because recurring revenue was the story investors wanted to hear. Guilt creates churn faster than boredom does.

Build the business around the best 30%. Blue Apron's data showed their top customer segment was profitable. The company knew who these people were. Instead of spending $460 to acquire another million coupon clippers, expand their options. Increase their order value. Grow revenue per customer instead of customer count. And take a page from HelloFresh: build sub-brands for the segments you're losing. A budget line. A no-cook line. Give people somewhere to land inside the ecosystem besides cancellation.

Fix retention before scaling. You don't pour water faster into a bucket with a hole in it. You fix the bucket. Blue Apron never fixed the bucket. They just bought bigger hoses. Every dollar spent on acquisition while 70% of customers were unprofitable accelerated the collapse. COVID handed them a second chance in 2020 when home cooking spiked and demand surged. They still couldn't capitalize. If a global pandemic that forced everyone to cook at home can't fix your retention problem, the problem is the product, not the marketing.

And the unsexy one that might have mattered most: fight for every cancellation. A phone call costs maybe $15. If it saves even 10% of canceling customers, and those customers generate $200 in additional lifetime value, that's a 13x return. Blue Apron had two fulfillment centers with about 80% of their workforce packing boxes. Fifty people calling canceling customers would have been the highest-ROI spend in the company. HelloFresh invested heavily in retention calls and outperformed Blue Apron on churn. Blue Apron never tried.

What Other Companies Did That Blue Apron Didn't

The hardest thing about the Blue Apron story is that churn at this scale had been solved before. Not in meal kits. In other subscription businesses facing the same structural problem: a product that requires ongoing behavioral commitment from customers who have every reason to quit.

Peloton is the closest parallel. They sell a $2,300 bike and a $44/month subscription that requires you to exercise regularly. That's an enormous thing to ask of a customer on a monthly basis. Their connected fitness churn rate dropped to 1.2% per month in Q3 2025. Now, Peloton has a structural advantage Blue Apron never had: a $2,300 bike creates switching costs that a meal kit can't match. But the retention tactics mattered even without the hardware. Peloton tracked workout frequency at the individual level and flagged customers who hadn't logged a session in 30 days. Those customers got personalized outreach.

Not a generic win-back email. A specific class recommendation matched to their stated goals. Peloton also found that members using two or more workout types per month churned at roughly 60% lower rates. So they invested in expanding beyond cycling into strength, running, yoga, and meditation. More reasons to open the app meant fewer reasons to cancel.

Netflix operates at about 2% monthly churn using a different version of the same principle: reduce the effort required to stay. Their recommendation algorithm does the work of choosing. When subscribers cancel, about half return within six months, according to Recurly's 2023 data, because the switching cost of rebuilding your profile and recommendations elsewhere is real even if it's invisible.

Blue Apron had none of this. No behavioral tracking at the individual level. No intervention before cancellation. No flexibility in delivery cadence. No effort to make the product easier to stay with over time.

The Deeper Lessons

I keep coming back to a handful of principles from this story that apply way beyond meal kits.

Your current CAC is the only CAC that matters. This is the one I want tattooed on my thinking. Blended averages are historical artifacts. I run apartment buildings, not subscription boxes. But the principle is identical. What does it cost me to fill a vacancy today? Not what it cost on average over the last five years. Today. That number tells me whether my marketing is getting more efficient or less. If I'm averaging it with data from 2019 when the market was different, I'm doing exactly what Blue Apron did.

High churn is a product verdict, not a marketing problem. If people leave fast, spending more to replace them is the most expensive response. I see this in apartments too. If turnover is high, the answer isn't better ads. It's asking why people leave and fixing that first.

Building a category without building a moat is crazy. Blue Apron spent the money to teach America what a meal kit was. They captured the initial demand too, holding 53% market share as late as 2016. The problem was they couldn't serve it profitably, and by the time that was clear, better-run competitors like HelloFresh were already scaling. HelloFresh reported €6.8 billion in revenue in 2025, though they're facing their own headwinds now. Active customers are declining, and the meal kit model itself is showing cracks. If you're spending to create awareness in your market, you need to be building switching costs or operational advantages at the same pace. I think about this with every dollar I spend on neighborhood reputation. Am I the one who benefits when the area improves, or is every landlord on the street getting a free ride on my marketing?

Growth and progress aren't the same thing. This one stings because I've made the mistake. Blue Apron's revenue grew 10x in two years. Forbes cover. Nearly $2 billion valuation. And they were losing money on seven out of every ten customers. Revenue was climbing while the business was dying. I've done my version of this, counting doors and units when I should have been counting cash flow per unit.

Incentives drive behavior, even when that behavior drives the company off a cliff. The more useful question than "who screwed up?" is whether the system around the CEO made any other outcome likely. Blue Apron's board was stacked with growth-stage VCs whose fund timelines rewarded a fast IPO, not a slow fix. If the board, the investors, and the market all reward the same behavior, and that behavior happens to be destructive, the problem is structural. Swapping the CEO changes the name on the door.

Blue Apron's founders weren't incompetent. They built a product millions of people tried. They raised capital from sophisticated investors. They executed an IPO. And the business was ultimately sold for the equivalent of one year's marketing budget.

The gap between their story and their economics persisted for years because growth is intoxicating. When revenue is climbing, nobody wants to ask whether each dollar of revenue costs $1.30 to produce. The question feels disloyal.

I'm trying to train myself to ask it anyway. These models (CAC vs. LTV, behavioral gaps, incentive structures, moat-building, the discount spiral) are just lenses. They don't guarantee I'll make the right call. But they increase the odds I'll ask the right questions before the math catches up.

And the math always catches up.