Introduction
Private equity returns are not distributed evenly, and the distance between the best deals and the rest has widened rather than narrowed. Gain’s 2026 analysis of 15,502 private equity investments globally puts median multiple on invested capital by deal performance quartile at 1.9x, 2.6x, 3.5x and 7.0x.[1] The top quartile is not marginally ahead of the field. It is close to four times the bottom.
What distinguishes those deals is not entry timing, sector selection or leverage. Top-quartile deals post a median revenue CAGR of 20.4%, against a market median that sits close to 10%.[1] They grow, and they grow more than twice as fast as everyone else.
That should be the end of the discussion, and in one sense it is. The industry has settled the question of whether revenue growth drives returns. What it has not settled, and what this article addresses, is the harder question underneath: whether the growth a business is producing is the kind that survives contact with a buyer’s diligence team, or the kind that reverses without warning the moment customer acquisition becomes more expensive. Where an individual deal sits against that spread, and what counts as a good MOIC once hold length, sector and region are taken into account, is set out separately.
What is the MOIC gap?
- The MOIC gap is the distance between median and top-quartile returns on invested capital. It is primarily a revenue growth gap.
What is MOIC optimisation?
- MOIC optimisation is the set of operational decisions taken during the hold period that improve the multiple on invested capital at exit. It is distinct from financial engineering and from multiple expansion, because it works on the business rather than on the capital structure or the market. Customer-base improvement, specifically retention in the highest-value customer tier, the quality of newly acquired customers, and concentration risk, is the primary operational lever available.
What drives the MOIC gap
Three structural changes have converged, and each one narrows the number of ways a deal can produce a return.
Multiple expansion has stopped paying. Revenue growth accounted for 57% of private equity value creation across the last decade, but its share rose from 44% in 2019 to 75% in 2025, while the contribution from multiple expansion collapsed from 46% to 8%.[1] In Europe the median EV/EBITDA multiple for PE deals now sits at 10.5x, down 23% from the highs, according to Gain’s H2 2026 European market analysis.[2] The re-rating that rescued a generation of deals is not available.
Entry prices have not come down and leverage has. McKinsey puts entry multiples at 11.8x EBITDA in 2025, with debt falling to 37% of the entry multiple from 44% in 2016.[3] More equity in, at a higher price, with less to sell it into.
The growth required has more than doubled. Bain’s arithmetic is that a deal targeting 2.5x MOIC over five years now needs roughly 12% annual EBITDA growth, against about 5% a decade ago.[4]
Fund-level returns show the compression. The Cambridge Associates US Private Equity Index returned 8.7% in 2025, against a ten-year figure of 15.2%. Within that index, buyouts returned 7.6% while growth equity returned 11.9%.[5] The strategy built around growth outperformed the strategy built around control and structure, in a year when both faced the same conditions.
Why the MOIC gap is a revenue growth problem
The obvious alternative explanation is margin. If growth is hard to come by, the argument runs, take costs out instead.
The European data does not support it. EBITDA margins for European PE-backed assets have been broadly flat for eight years: a median of 11.0% in 2017 and 11.4% in 2024.[2] Eight years of operational improvement programmes, procurement reviews and cost transformations, and the median asset has moved roughly 40 basis points. Margin expansion works in specific situations, most obviously in turnarounds and large carve-outs, but at the median it has not been closing anything.
Growth, by contrast, grades every other outcome in the deal. Companies growing above 30% revenue CAGR post a median MOIC of 4.3x against 2.3x for those growing 0% to 10%. Fast-growing companies command exit multiples 30% to 70% higher than slower-growing ones, a relationship that holds across every size band and sector. Businesses with positive revenue growth are also more likely to expand margins, because operating leverage does the work a cost programme otherwise has to do.[1]
The downside figure is the one that should change how deals are underwritten. Among investments with negative revenue growth, 26% returned less than the capital invested. Among those growing above 10%, that falls to between 1% and 2%.[1] Growth is not the upside case. It is what stands between a deal and a loss.
None of which makes growth easy to find. In Europe, median revenue growth for PE-backed assets was 9.7% in 2025, with a top quartile of 24.6% and a bottom quartile of 0.1%.[2] A quarter of European PE-backed businesses grew by essentially nothing.
The pattern is sharpest in consumer. Consumer businesses rely on revenue growth for 66% of their value creation, more than any other sector, yet post among the slowest median growth at 9.5% and one of the highest loss rates at 9%.[1] Cambridge Associates data points the same way from a different direction: consumer discretionary was the weakest-performing sector in the US PE index in 2025 at 3.5%, against 14.0% for industrials.[5] The sector most dependent on growth is among the least reliable at producing it, which is what you would expect in categories where customers are cheap to buy and expensive to keep. This is the territory where revenue quality does the most work.
Growth decides which assets can be sold at all
The clearest evidence that this is a revenue problem rather than a market problem is in what is currently moving and what is not.
Median three-year revenue CAGR is 15% for European assets in the live exit pipeline, 13% for those exited in 2025 and 2026, and 10% for those that have gone unsold for three years or more.[2] The whole distribution sits lower for the unsold: a top quartile of 21% against 28% in the pipeline, and a bottom quartile of 2% against 7%. Sponsors are realising their higher-growth assets and holding the rest. The unsold pile is not a random sample of the market. It is the low-growth end of it, and it is accumulating.
The scale of that accumulation is now a structural feature of the asset class. Bain reports roughly 32,000 unsold companies worth $3.8 trillion in unrealised value, with distributions as a percentage of net asset value holding below 15% for four consecutive years, an industry record.[6] Cambridge Associates, working from a separate database, reports the same condition: distribution yield remained below its historical average for the fourth consecutive year, with managers distributing $185 billion and calling $145 billion.[5]
Two houses, two datasets, one finding. The problem is not that buyers have disappeared. Buyers are active for assets that grow. The problem is a large population of assets that cannot demonstrate durable growth, and therefore cannot be priced.
Durable revenue and replacement-dependent revenue
This is where the measurement stops and the judgement has to start, because revenue growth is being treated as a single quantity and it is not one.
Two portfolio companies can post identical 12% compound growth across a hold. In the first, customers who bought three years ago are still buying, buying more often, and buying across more of the range, and newly acquired customers are settling into the same pattern. In the second, the customer base turns over almost completely every eighteen months, and growth is the arithmetic result of acquiring replacements slightly faster than the business loses them.
On the profit and loss account those two businesses are indistinguishable. In the data room they look the same. In the equity story they read the same. They are not the same asset, they will not survive the same conditions, and they should not trade at the same multiple.
The first is durable revenue. The second is replacement-dependent revenue, and it has one property that makes it dangerous in a buyout specifically: it works until acquisition efficiency turns. When paid acquisition costs rise, or a channel changes its rules, or a competitor bids the same audience up, the growth does not slow. It inverts, because the business was already running to stand still.
I have spent twenty years working on customer bases in ecommerce, subscription, retail and financial services, and this distinction is visible in transaction data in almost every one of them. It is almost never visible anywhere else, which is why it so rarely reaches an investment committee.
Three customer-base mechanisms that close the MOIC gap
The gap is not closed by a single action. It is closed by three compounding improvements to the customer base, each identifiable and measurable from transaction-level data, and each operating on a different part of the base.
Mechanism 1: retention in the highest-value customer tier
How much of a company’s revenue rests on its most valuable customers varies enormously, and it cannot be assumed from the sector. In a European ecommerce business with 5.4 million customers, 16.6% of customers generated 55.7% of revenue. In a pan-African remittance business, two customer groups holding 38.6% of the base generated 68.8% of fees. In a short-term insurance group, premium share tracked customer share almost exactly, and value concentrated in margin rather than in revenue.
That range is the point. A retention programme aimed at the most valuable fifth of customers is transformative in the first business and close to irrelevant in the third. Which one you are holding is an empirical question, and the answer is in the transaction data, not in a sector benchmark. All three figures are from our evidence.
Where a business does carry that shape, the arithmetic is unforgiving. Take a business with £50m of annual revenue where 55% is carried by the most valuable sixth of customers, the ecommerce pattern above. That is £27.5m resting on one group. A 5% improvement in annual retention for that group, sustained across four years of a hold, produces a materially different revenue base from the baseline trajectory. High-value customers who stay also tend to grow their spend, so retaining them delivers both a floor and a slope.
The decisions that improve top-quintile retention are specific to that tier’s behavioural profile: its purchasing frequency, breadth, recency pattern and sensitivity to product, pricing or service changes. Aggregate retention reporting and satisfaction scores do not produce that picture at the granularity required to act on it. Individual customer-level analysis does.
The lead time is the point. Recency deterioration in the highest-value customer tier typically precedes revenue impact by 6 to 12 months. It is visible in transaction data throughout that period. It is not visible in an aggregate retention rate or a management dashboard until the revenue event has already started.
Mechanism 2: the quality of newly acquired customers
Acquisition volume is the metric most portfolio companies track. Acquisition quality is the metric that determines whether this year’s growth becomes next year’s revenue base or next year’s attrition problem.
It is measured by comparing the initial spend, frequency, breadth and retention trajectory of each successive acquisition year against the incumbent base at the equivalent point in its lifecycle. A business where each successive year of new customers arrives at progressively lower initial value is accumulating a structural weakness the revenue line will not reflect for 18 to 30 months. A business where that quality is improving is building an advantage that will be visible in the exit revenue base.
In our ecommerce case study, the business was acquiring 6.5 low-value customers for every high-value one. Revenue was growing. Acquisition quality was deteriorating. The revenue base that would actually be available three years later was already a different asset from the one being priced.
The implication is that acquisition spend has to be optimised for quality, not volume, which requires knowing at the point of decision what a high-value customer looks like behaviourally, and whether the current channel mix, positioning and pricing are likely to attract that profile or a different one.
Mechanism 3: concentration risk, reduced before the exit window
High customer concentration is one of the most consistent sources of exit multiple discount, and one of the least consistently priced. Buyers and their advisers apply it informally, embedded in the multiple rather than itemised in the model, which means it is rarely negotiated on its merits and almost never quantified against the actual shape of the base.
The direction is well evidenced. Acquirers of customer-concentrated targets earn lower returns and worse long-run operating performance, and the effect is sharper where customers face low switching costs.[7] What no source supports is a single reliable figure for the discount, because the adjustment varies by buyer, sector, contract quality and whether the concentration is improving or worsening. Anyone quoting a fixed percentage is quoting a rule of thumb, which is the problem rather than the answer.
Reducing concentration during the hold is therefore a value creation lever in its own right, separate from and additive to revenue growth. A business entering its exit window demonstrably less concentrated than at entry has created value the profit and loss account does not capture but the multiple will.
There are three routes. Grow the mid-tier, so the top quintile’s share falls because the rest of the base has grown rather than because top-tier revenue has fallen. Broaden the purchasing breadth of existing mid-tier customers, which raises their contribution and improves their retention at the same time. And diversify acquisition deliberately towards profiles similar in value but distinct in sector, geography or product dependency from the existing concentration.
None of these is visible in management accounts until the concentration figure itself has moved. The trajectory, meaning the direction and speed of change, is visible in customer behaviour data throughout.
Why extended holds widen the MOIC gap
Assets are being held longer, and the effect on the gap is not symmetrical.
Bain puts the average holding period at exit at around seven years globally, up from an average of five to six years between 2010 and 2021, with almost 40% of all companies held for more than five years against 29% in 2019.[6] In Europe, Gain reports a median holding period at exit of 5.8 years in 2026, and one-third of European PE assets now held for more than seven years.[2]
Longer holds do not simply mean more time. They mean the return depends more completely on the customer base. Assets held more than seven years derive 71% of their value creation from revenue growth, the highest share of any hold length.[1] They also produce the widest spread of outcomes: the highest top-quartile MOIC at 5.8x, and the highest loss rate at 13%.[1]
Time also works against the return in ways that cannot be operated around. Internal rate of return starts to stagnate around year seven and declines after that, and median total value to paid-in capital at fund level begins to flatten after year eight.[6] Limited partners are, for now, willing to wait: about two-thirds lean towards holding out for an improved multiple rather than near-term liquidity.[6] That patience is conditional on the fund having produced appreciation to date and having a credible plan for continuing it, which is a different thing from patience itself.
There is a compounding argument and a decay argument, and they are the same argument seen from either end.
A retention opportunity identified at year 2 of a seven-year hold has five years to compound into the exit revenue base. The same opportunity identified at year 5 has two. And value creation is already back-loaded: McKinsey finds roughly 6% of final EBITDA margin is generated in the last year of the hold, 4% in the penultimate year, and about 1% per year before that.[3] Sponsors are, in aggregate, doing the work late.
Deterioration behaves the same way in reverse and is harder to see. A customer base that is eroding beneath the surface, with high-value recency deteriorating and newly acquired customers arriving at lower value, can produce four or five years of flat or modestly growing revenue before the revenue event surfaces. By then the hold is nearly over and the options are limited. The assets where MOIC shortfalls concentrate are rarely the ones that deteriorated quickly. Fast deterioration is visible and gets a response. They are the ones that deteriorated slowly, across an extended hold, because nothing in use was capable of showing the early signal.
How to maximise MOIC through customer-base improvement
The sequence matters more than any individual intervention, because the value of each finding depends on how much hold period remains to act on it.
Commission a diagnostic at year 2. This is the earliest point at which 24 months of post-acquisition transaction data exists. It shows whether the customer-base assumptions made at entry were correct, which tiers are performing as expected and which are not, and what would most improve the revenue trajectory across the remaining hold. It is the moment to recalibrate the value creation plan against observed data rather than entry assumptions, which came from the seller. A customer base diagnostic is built for exactly this point in the hold.
Act on the single most important finding. The output carries one primary recommended action. Retention improvement in the top quintile usually requires a combination of product, service and pricing decisions taken with that tier’s behavioural profile as the brief. Acquisition quality requires changes to channel mix or targeting. Concentration reduction requires a deliberate growth programme in the mid-tier. A value creation map turns the finding into a sequenced set of decisions.
Commission a second diagnostic at year 4. Two years on, the first round of actions has had time to compound. The second diagnostic shows whether the three mechanisms have moved in the right direction, what the forward revenue range looks like as the exit window approaches, and what the customer evidence layer in the equity story will be able to say.
Build the equity story from the evidence, before the process starts. This is where the hold-period work converts into price. Among general partners surveyed by EY-Parthenon, 86% said exit preparation initiatives improved their exit valuations, and around half of those who began 12 to 24 months before sale reported substantial improvement, while those starting less than six months out reported materially weaker results.[8] The same study found 60% cite developing a robust set of data and KPIs as their hardest finance-function challenge at exit, and a clearly defined equity story ranked second both for impact on outcomes and for difficulty.[8] Two diagnostics covering four years of observed customer movement answer the first problem and evidence the second. How that material is used in a process is set out in exit readiness customer evidence.
What the instrument requires and what it produces
The input is an anonymised transaction extract. Transaction ID, customer ID, date, value, category. No personally identifiable information. Most portfolio companies can produce it from their accounting system or ERP within 24 hours, which is what makes it repeatable at year 2 and year 4 without the process cost of a formal diligence engagement.
The output is year-on-year cohort movement across the full observation period, typically 12 to 48 months, covering how the high-value tier has moved on retention, recency, spend and breadth; how the quality of newly acquired customers compares across the last four to six years; whether the revenue base is becoming more or less concentrated; twelve-month forward scenario projections based on observed migration rather than management assumptions; and the single intervention that would most change that forward range.
It is written to be legible to an operating partner, a deal team partner, a CFO and a CMO without specialist interpretation. It is a commercial verdict expressed in the language of the strategic review, anchored in individual customer data. The full approach is set out in Keystone IQ’s methodology, and the underlying framework is Keystone IQ’s Revenue Quality Architecture™.
For assets at entry rather than mid-hold, the same questions are answered faster and more narrowly by a health check, which sets the baseline the hold-period work is measured against.
The gap is operational
Limited partners have already reached this conclusion. In a January 2026 survey of 300 LPs reported by McKinsey, 53% ranked a general partner’s value creation strategy among their top five manager selection criteria, displacing sector expertise.[3] The question is no longer whether a fund can buy well. It is whether it can demonstrate what it did to the business afterwards.
Growth now carries three quarters of the return in private equity, and the industry has responded by getting much better at measuring how much of it a deal produced. It has not started measuring what kind. The assets sitting unsold at the bottom of the growth distribution were not badly bought. Most of them were bought on a growth rate that turned out to be replacement-dependent, and nobody could tell at the time, because nothing in the pack separated a customer who came back from one who passed through.
The test I would apply is short. If you cannot say what proportion of this year’s revenue came from customers who were already there, you do not know whether you own a compounding asset or an expensive treadmill. Both grow. Only one of them survives the moment acquisition gets harder, and that moment is where the 1.9x deals and the 7.0x deals part company.
Related reading: customer-driven goodwill and value creation evidence for operators.
References
- The Private Equity Value Creation Report 2026. Gain (Jain, Raju, Zegar), 2026. Median MOIC by quartile and top-quartile revenue CAGR of 20.4% (p.9). Revenue growth 57% of value creation over the decade, 44% in 2019 to 75% in 2025, multiple expansion 46% to 8% (p.2, p.4). MOIC 4.3x above 30% CAGR against 2.3x at 0–10%, exit multiples 30% to 70% higher, loss rates 26% against 1–2% (p.5). Extended holds 71% revenue-driven, 5.8x top quartile, 13% loss rate (p.39–40). Consumer 66% revenue reliance, 9.5% median CAGR, 9% loss rate (p.31–32). Sample of 15,502 investments globally, Shapley decomposition. Gain notes an upward bias to MOIC figures because better-performing deals are reported more frequently.
- The State of European Private Equity Report, H2 2026. Gain (Jain, Raju, Shetty), 2026. Exit selection by three-year revenue CAGR (p.36). Median revenue growth 9.7%, top quartile 24.6%, bottom quartile 0.1% (p.57). EBITDA margins 11.0% in 2017 to 11.4% in 2024 (p.57). Median EV/EBITDA 10.5x, down 23% (p.56). Median European holding period 5.8 years, one-third held beyond seven years (p.53, executive summary). European data throughout.
- Global Private Markets Report 2026: Private Equity, Clearer View, Tougher Terrain. McKinsey & Company, February 2026. Entry multiples 11.8x EBITDA, debt 37% of entry multiples against 44% in 2016 (p.17). Back-loaded value creation (p.18). 53% of 300 LPs ranking value creation strategy top five (p.21).
- Global Private Equity Report 2026. Bain & Company, February 2026, p.32. Approximately 12% annual EBITDA growth required for a 2.5x MOIC target over five years, against roughly 5% a decade ago.
- US Private Equity & Venture Capital Benchmark Commentary, Calendar Year 2025. Cambridge Associates (Slotsky, Carneal, Sample), published 30 July 2026. CA US PE Index 8.7% in 2025, buyouts 7.6%, growth equity 11.9%, ten-year 15.2% (Figure 1). Consumer discretionary 3.5% against industrials 14.0% (Sectors). Distribution yield below average for a fourth consecutive year, $185bn distributed against $145bn called (LP cash flows). All figures are pooled horizon internal rates of return net of fees, covering US buyout and growth equity funds only. They are not multiples and must not be described as MOIC.
- Private Equity Outlook 2026: Gaining Traction. Bain & Company (MacArthur, Burack, Rose, Schmitz, Yang, Lamy), Section 1 of the Global Private Equity Report 2026, 22 February 2026. Holding periods at exit around seven years, up from five to six years between 2010 and 2021, and almost 40% held beyond five years against 29% in 2019. 32,000 unsold companies worth $3.8 trillion. Distributions below 15% of NAV for four consecutive years. IRR stagnating around year seven, TVPI flattening after year eight. About two-thirds of LPs favouring improved MOIC over near-term liquidity, from the ILPA LP Sentiment Survey 2025–26 Edition.
- Customer concentration and M&A performance. Yizhe Dong, Chang Li and Haoyu Li, Journal of Corporate Finance, volume 69, article 102021, August 2021. Acquirers purchasing customer-concentrated firms experience significantly lower stock market returns and worse long-run operating performance. The effect is more pronounced where customers face lower switching costs, where the target undertakes higher relationship-specific investment, where cash volatility is higher, or where the acquirer is less well known. The negative association is driven mainly by corporate customers, with government customers moderating the effect. Listed-company M&A rather than mid-market private equity, so it supports the mechanism rather than any specific discount figure.
- Global Private Equity Exit Readiness Study 2026. EY-Parthenon (Nardi, Lehon), 2 June 2026. 86% of GPs reporting improved exit valuations from preparation initiatives. Around half of those preparing 12 to 24 months ahead reporting substantial improvement. 60% citing data and KPIs as the hardest finance-function challenge at exit. Equity story ranked second for impact (51%) and second for difficulty (45%).
Related reading: customer-driven goodwill and value creation evidence for operators.
>