Newsletter
Bain says CS is broken. They should know, they're the ones who broke it.
You’ve probably seen the article going around from Bain & Co that purports to explain why customer success is broken and how to fix it. It’s filled with intelligent, confident-sounding diagnoses and prescriptions. It even drops a few “AI” bombs to be down with the cool kids. They are certainly right that CS is broken, but they left out one important detail: they’re the ones who broke it!
A generation ago Bain set customer success off on the wrong track with a single, appealing, but catastrophically wrong idea: happy customers are loyal customers. To be fair, the idea wasn’t Bain’s invention. It had been around for a long time. But Bain hardened it into a full-blown religion in the business world, which, decades later, has come crashing down.
In 1990 Fred Reichheld of Bain co-wrote “Zero Defections” in Harvard Business Review. He was the founder of Bain’s Loyalty practice, and then in 1996 he published The Loyalty Effect. In 2003, as a Bain partner, he introduced Net Promoter Score in “The One Number You Need to Grow.” The claim was that “promoters” stay longer, buy more, and go on to bring in other customers, and that the score predicts growth. Bain’s research appeared to show NPS “correlates closely with sales growth, expanded share of wallet, sales force productivity, greater market share, greater employee engagement and higher profitability.”
The idea took the business world by storm. Two-thirds of the Fortune 1000 ultimately adopted NPS, and the score became a kind of premium corporate currency. Companies pursued customer experience initiatives with unbridled zeal, believing fervently that growth would follow. After all, Bain promised that improvements in customer satisfaction would translate directly into enterprise value. It is no exaggeration to say that the entire field of customer success was built on Bain’s idea: raise how customers feel about you and they’ll stay forever and buy more.
It didn’t work. Twenty years and tens of billions of dollars went into customer experience investments, satisfaction surveys, health scores, check-in calls, and QBRs. But the promised results never arrived. Bain’s own numbers say so: customer success headcount is up more than 50% in six years, and net revenue retention is down from about 115% to about 107%.
To those of us who went along for the ride, the article makes excruciating reading. It pronounces the failure of customer success as if Bain discovered it and Bain had nothing to do with what caused it. To be fair, the specific authors probably don’t even know. After all, that was a generation ago, and they clearly don’t understand what really caused it anyway. And that’s what makes this article and all the traction it’s getting so dangerous. Every one of its so-called “new” ideas is nothing more than a re-warmed version of the old ones that failed. The article gets literally every one of its premises wrong, and its proposals amount to nothing more than a fresh dose of old medicine. Here’s a quick tour of why every new Bain premise is dead wrong…
Ten False Premises
1. “Satisfied customers are more likely to renew.”
This is the founding false premise of customer experience, restated in an article about why the model failed. It was never true. In our data, promoters (NPS 9–10) stay 15 months on average. Everyone else stays 20. Across 130,000 customers, the correlation between a customer’s satisfaction score and how long that customer stays is zero. Happy customers leave, and some of the hardest to please stay for years. If you’ve spent any time in customer success, you’ve watched it happen. Customer satisfaction is a real thing. It’s just not the reason customers stay. By far the best predictor of retention is the measurable results customers get from the product. How they feel about the vendor has nothing to do with it.
2. “The problem wasn’t the investment; it was the model.” The mission hasn’t changed: help customers realize value.
Wrong again: the mission is the problem. “Value” is as mushy today as “satisfaction” was a generation ago. Customers with a measured result stay six times longer than customers without one. Satisfaction buys you a vanity score to brag about in the company report, NOT growth. Bain gives “measurable results” one mention in passing and clearly does not respect it as the central lever: they never say who measures the customers’ results, how, which results, or who owns the number. The new model still treats “value” as a glitteringly vague aspiration.
3. Net revenue retention is the right scoreboard.
Bain first measures the failure of customer success in NRR, then turns around and prices the rehabilitation in NRR. The problem with NRR is that it conceals the real rot: customer churn. By blending churn, downsell, and expansion into one figure, it hides how many customers left and when. Our data shows the median company has bonded only about 7% of its customers. Roughly 92% are still leaving, even if only very slowly. Retention and churn rates hide the problem too. Companies selling to large enterprises look great because they retain about 98% of their customers through the first year. Look closer and they bond essentially none. NRR is the preferred scoreboard because of what it hides.
4. “Much of that decline is a result of forces outside customer success’s control” (maturing markets, AI budgets).
Here we are taken on a ride back to the “unavoidable churn” excuse, with a squirt of AI hot sauce to spice it up. What does the data show? The actual behavior of millions of customers shows retention flat to declining for more than a decade, long before anyone had an AI budget. Bain’s only evidence for the AI threat is a survey of what buyers say they might do. External forces are real. They do not explain why the bonded fraction stays so low, or why the same companies keep losing customers at the same rate year after year.
5. “Customers want deeper product and technical expertise.”
Actually, customers do not want your product at all. They want results. Customers using the exact same product get wildly different results, so the product cannot be what separates the ones who stay from the ones who leave. The expertise that matters is knowing the specific behaviors that produce good business results with the product. Product specialists and forward-deployed engineers know the software. Nothing in the job Bain describes says they know how customers get results with it in their business. And no, adoption is not the same thing (see the next point). Bain is sending more people who know the product to solve a problem the product does not cause.
6. Adoption, usage, and advocacy are leading indicators worth paying on.
False. None of these predicts who stays. They are lagging indicators of customer success, not leading ones. In our data the heaviest users churn as fast as the lightest. Successful customers do end up with higher adoption, but the causality runs the other way: high adoption is far more an outcome of results than a cause. Advocacy is merely the NPS question again (“would you recommend…”). And once compensation depends on these metrics, they will invariably be gamed. This “new” scorecard is just the old scorecard with better software and higher stakes.
7. One executive-facing “success architect” with C-suite relationships secures the renewal.
Bain’s answer to the failure of customer success is to turn it into sales: hire and pay people like account executives, manage them like a sales team, and send them to talk to executives. It won’t work. Our data shows zero impact from executive QBRs and higher executive engagement. Renewal is decided by results, not relationships. Customer success exists to get customers measurable results and materialize them so the customer understands. Measuring those results is nowhere in the job description Bain wrote for the success architect.
8. AI agents should cover every account (outreach, renewal admin, QBR preparation).
Here Bain takes an old, worn-out idea and slaps AI lipstick on it. There has never been evidence that “touching the customer” through constant interaction improves retention. It is the essence of the method that failed. Every time we “touch” a customer without adding value, we train them to ignore us. That’s why I call it “inappropriate touching.” Replacing humans with bots so we can do more of it, faster and cheaper, will not change the game. AI makes a worthless engagement model cheap enough to do to every customer. If you think your customers are ignoring you now, just add AI and see how it goes.
9. “Growth now comes from customer expansion.”
Here’s where the article irredeemably jumps the shark. Bain says expansion has now become the leading driver of growth. Ahead of what? Expansion has always been the dominant driver of growth. It is literally the founding premise of the entire subscription economy. And it’s bigger than that: expansion is not just something you get from keeping customers, it is the reason they stay. Renewal is a forward-looking decision. Past results matter mainly because they make the promise of future results credible. In our data, the largest predictor of churn after the first year is the lack of expansion. Bain has been talking about customer loyalty since 1990, and here they are in 2026 presenting the basic arithmetic of the subscription model as their latest discovery.
10. “Post-sale is where the next decade of software growth is won.”
The final false premise is not merely empty, it’s fatal. The evidence shows what everyone in customer success already knows: you can’t solve churn after the sale. The primary drivers of churn are in place before the close. 88% of the variation in customer bonding comes down to factors customer success can’t even touch. Customer fit, results expectations, and commitment to change are settled before the ink is dry. Right-fit customers bond at about 92%. Wrong-fit customers have essentially no bonded core. Even how we sell matters: discounting and free trials are both huge drivers of churn. The consequences arrive fast: the median company loses about 47% of its customers in the first year, a clear sign of weak product-market fit. Companies buy more churn than they make. Owners, specialists, and agents added after the sale cannot change that math.
A New Cycle of Old Ideas
All of this is just a rehash of the mentality that got us here. The good news is that our research reveals a compelling alternative approach with real leverage. It starts with a correct understanding of what really causes customers to stay. Don’t call it satisfaction or loyalty. I call it Customer Bonding, which is a real and consistent condition we see throughout the data. Bonded customers lock in with a company indefinitely. But what is customer bonding? Why do some customers bond? And why are some companies so good at producing it?
The key to understanding customer bonding is a simple idea: bonding happens when customers cannot visualize their future success without your company in the picture.
Our research shows bonding is the result of five essential elements:
- The product has to produce a compelling and measurable business result.
- It must only be sold to customers who are able to achieve the result.
- Customers have to change how they operate in the specific ways the product needs to produce results.
- The results must be measured and materialized to the customer early and often.
- Customers must be continually led to their next result.
None of these elements is comprehended in Bain’s model. It’s wise always to consider the source when taking advice. Bain’s last big idea about keeping customers sent the business world down a wasted path for twenty years. Caveat emptor, dear reader.
Keep reading
One finding from the research, every Friday.
Thank you — see you Friday.