Somewhere in your portfolio there is a value creation plan with a customer growth line in it. Cross-sell, upsell, share of wallet, expansion revenue, an adjacent product. It was sized at entry. It is probably the largest single lever in the plan, and the market has moved in that direction: revenue growth accounted for 75% of private equity value creation in 2025, up from 44% in 2019, according to Gain’s Private Equity Value Creation Report 2026. And that line has almost certainly not been underwritten against the actual customer base.

If the number was built from averages and an assumed conversion rate, it is a value pool, not a value creation plan. The gap between the two is where private equity leaves money on the table.

Not in the way most people mean, though. The usual version of this question points at upside: customers who could be served better, accounts with a credible next purchase, valuable relationships drifting away unnoticed. All real. But the larger loss for a sponsor is the other one. It is the discounts, sales effort, product investment and management attention a portfolio company spends chasing growth that was never there, while the relationships the multiple actually rests on get less attention.

A weak customer growth assumption does not just miss its target. It contaminates the value creation case around it.

Why do customer growth assumptions fail in a value creation plan?

Most customer growth assumptions fail because they are sized from averages, and the average customer does not exist.

Whether the base is subscribers, accounts, policyholders or shoppers, the pattern is the same. Growth gets sized with average customer value, broad customer groups, blanket KPIs and top-down opportunity maths. These are fine reporting devices. They are poor decision devices.

Take an account that buys twice a year, in line with its own planning cycle. It may be using the business exactly as it wants to. Label it “low frequency”, assign a rep to increase contact and push products it has no use for, and the cost goes up while the relationship gets worse.

Another account with the same headline frequency might show a completely different pattern, with a genuine route into a deeper relationship. Put both in the same bucket and the business cannot tell the real opportunity from the false one. Not every gap in customer behaviour is an opportunity.

What does a value pool look like when it meets a customer?

A value pool is a theoretical upside sized from a spreadsheet. A value creation plan is that upside qualified against real customer behaviour.

Here is a common one. A business decides every active customer should be buying across three product lines. Count the single-line customers, compare their value with the multi-line customers, apply an assumed conversion rate, and a large number appears. It goes into the plan as a strategic target.

It is back-of-the-envelope maths. It does not establish which single-line customers have any propensity to broaden, which are naturally narrow users who should be left alone, which next product is relevant to each of them, whether the second purchase is incremental or cannibalised, or what it costs to activate and serve.

There may be a genuine opportunity in there. A recent purchase in one line often does point to a credible next need in an adjacent, repeatable one. But that is the right next product for the right relationship, where the economics justify the effort. It is not “plus one product line for everyone”. A KPI can be directionally right and still be commercially wrong when applied to every customer.

Can more customer data make a bad assumption worse?

Yes. Models are not mirrors. A model tells you what would be true if your assumptions were right. It cannot tell you what is actually happening in your base.

A second example, anonymised. A business invested heavily in a large customer modelling programme. The output identified a very large customer group as a major opportunity for an adjacent financial product. Big group, valuable product, assumed take-up rate, enormous headline number. Exactly the kind of line that survives an investment committee because it looks rigorous.

The observed behaviour said otherwise. These customers barely engaged with the business. Nothing in how they used it suggested the new product was part of the relationship, or that they saw this company as a trusted route to it. Realistic take-up was close to zero.

The modelling had not created an opportunity. It had made a weak assumption look more sophisticated. An addressable market is not the same as an addressable relationship. Customer data becomes dangerous when it gives a business confidence without giving it judgement.

Why does the wrong customer growth strategy cost more than the sales and marketing budget?

Because a customer growth initiative can look compelling in a model and still reduce the value of the customer base when it reaches real customers. This is the erosion case, and every growth plan has one whether or not it is written down.

Low propensity gets forced with discounts, incentives and contact volume. Revenue is cannibalised rather than incremental. Cost to serve runs higher than assumed. And the customers on the receiving end learn something: ignore the outreach, wait for the discount, treat the relationship as transactional.

One irrelevant approach will not make a good customer leave. A persistent pattern of not understanding them can. The loss rarely shows up as a churn event. It shows up as lower frequency, narrower spend, more price sensitivity, less trust. Bain’s 2026 midyear report identifies the middle of the hold as the point where value creation most often gets lost, as the initial plan runs out of steam. This is one of the reasons why. By the time erosion reaches the revenue line, it is a diligence finding at exit rather than a fixable problem in year two.

How do you underwrite customer growth in a value creation plan?

A customer growth number becomes an EBITDA lever only when customer-level behaviour supports it.

Underwriting it means asking, at customer level: which customers support the case, what in their observed behaviour suggests they will respond, what is genuinely incremental, what it costs to activate, serve and retain the new behaviour, and what should cause the business to stop. Net value is the expected upside less the erosion case. The quality of a plan depends on whether both are written down. The pressure to do this is rising from the LP side too: McKinsey’s Global Private Markets Report 2026 describes an industry where alpha is less likely to emerge from market dynamics alone and increasingly has to be made through deliberate choices about how early and consistently operational value is created.

This is where Keystone IQ sits, upstream of the execution machinery. Keystone IQ’s Revenue Quality Architecture™ uses the portfolio company’s own transaction data, not a model of it, to separate a real customer opportunity from a large theoretical number, and a hypothesis worth testing from one that should come out of the plan before anyone spends against it. For deal teams it is a question at entry, answered by a five-day Health Check. For operating partners it is a question in the first hundred days. For the investment committee it is the difference between a lever and a line item, and the difference between revenue growth and revenue quality.

The model can size the prize. Only the base can tell you whether you have a credible way to win it.

FAQs