Deal Autopsy: We Overpaid. Here Is Exactly How It Happened — and the Investment Model That Would Have Stopped It
All details in this account have been anonymised. The sector, the individuals, and the specific figures have been changed to prevent identification. The financial mechanics, the reasoning errors, and the lessons are drawn from a real transaction.
Most deal autopsies are about things that went wrong that nobody anticipated — a hidden liability, a contractual complication, a relationship that broke down. This one is different. This is a deal where the fundamental error was visible in the numbers before completion, if only we had built the right model and looked at it honestly. We did not. We paid too much, and the consequences were felt for the better part of three years.
Writing this is not comfortable. But the lesson it contains is one of the most practically useful in any acquisition education — because overpaying for a business is a mistake that is almost always preventable, and the prevention comes from one thing: a rigorous investment return model, stress-tested against realistic assumptions, before heads of terms are signed.
The Business and the Context
The business was a specialist distribution company operating in a sector we knew reasonably well. Strong brand recognition within its niche. A customer base that was genuinely diversified — no single customer above 12% of revenue. Revenue of approximately £4.2m with an EBITDA of £520k on a normalised basis. The owner had been running it for eleven years and was motivated to sell.
On the face of it, the business was attractive. The sector was growing. The brand was defensible. The customer base was clean. The financial history was consistent — EBITDA had grown from £380k five years earlier to £520k at the time of sale, a compound growth rate of around 6% per year.
The agreed price was £3.12m — six times normalised EBITDA. At the time, that felt justifiable. Six times was at the upper end of where similar businesses were trading, but the growth trajectory, the brand strength, and the sector dynamics felt like they supported a premium multiple.
The deal structure was straightforward: £2.4m on completion from a combination of equity and senior debt, and £720k of vendor finance over three years at 6% interest. No earn-out. Clean.
What the Investment Model Said: When We Built It Honestly
The investment model we built at the time of acquisition was optimistic. It assumed EBITDA growth of 8% per year over a five-year hold period, based on the historic growth rate and a reasonable assumption that the sector tailwinds would continue. On that model, the exit EBITDA in year five would be approximately £765k, and at a conservative exit multiple of 5.5x, the exit enterprise value would be £4.2m. After repaying the acquisition debt, the equity return looked attractive.
The problem — which became apparent only when we rebuilt the model with more conservative assumptions eighteen months after completion — was that the base case model had embedded an optimism at every stage that, individually, seemed reasonable but which compounded into something that bore little relationship to what was actually achievable.
Let me walk through the assumptions that were wrong, in order of impact.
Assumption 1: EBITDA growth of 8% per year
The historic growth rate of 6% per year had been achieved in a period of strong sector tailwinds. Those tailwinds began to moderate in the year of acquisition — something that was visible in the market data if we had looked for it, but which we had not specifically investigated as part of the commercial due diligence. A realistic base case growth assumption, built on current sector dynamics rather than the prior five-year trend, was closer to 3% to 4%.
Over a five-year hold period, the difference between 8% growth and 3.5% growth compounded into an EBITDA gap of approximately £120k in year five. At a 5.5x exit multiple, that is a £660k difference in exit enterprise value. Not catastrophic in isolation. But this was the first of several compounding errors.
Assumption 2: An exit multiple of 5.5x
We had paid 6x. The model assumed we would exit at 5.5x — a modest multiple compression that felt conservative. What we had not adequately considered was the impact of scale on exit multiple. At £520k EBITDA, the business was at the very bottom of the range where 5.5x to 6x multiples apply. At £400k EBITDA — where a slower growth scenario would land us if the sector growth moderated — the applicable market multiple in most comparable transactions was 4x to 4.5x. The multiple compression from a lower earnings base was more severe than the model assumed.
Assumption 3: Free cash flow closely tracking EBITDA
The model used EBITDA as a proxy for cash generation. In year one, this assumption proved significantly wrong. The business had a capital expenditure requirement — replacement and upgrade of distribution infrastructure — that was significantly higher than the depreciation charge in the accounts suggested. Approximately £85k per year of additional capex above the depreciation run rate was required to maintain operational capability. Over five years, that is £425k of cash that the EBITDA model had not accounted for.
The business also had a working capital cycle that consumed cash as it grew. Every percentage point of revenue growth required additional debtor book funding. At 3% growth, that was manageable. It had not been specifically modelled.
The compounded result
Running the model with corrected assumptions — 3.5% EBITDA growth, a 4.5x exit multiple at the lower earnings base, and realistic free cash flow conversion — produced a picture that was materially different from the original model. The five-year equity return on the corrected model was approximately half what the original model projected. The deal was still marginally positive — we were not looking at a loss — but the risk-adjusted return was significantly below what we should have required to justify the capital commitment.
We had paid 6x for a business that, on realistic assumptions, warranted closer to 4.5x. The premium we paid — approximately £780k above where the deal should have been struck — was not recovered over the hold period.
The Three Errors in Our Analytical Process
Error 1: We built the base case from the seller's narrative
The growth assumption in our model was anchored to the historic trajectory that the seller had presented — and that we had accepted without sufficiently interrogating whether the conditions that drove it would continue. Good investment modelling starts from the current reality of the market, not from the seller's recent history. The seller's history tells you what the business has done. Market analysis tells you what it is likely to do. We let the first substitute for the second.
The correction: build the growth assumption from the bottom up, using current market data, competitive position analysis, and a realistic assessment of what the specific business can achieve under new ownership — before applying any optimism from historic performance. If the historic growth rate cannot be supported by current market conditions, the model should reflect current conditions, not the historic trend.
Error 2: We treated the base case as the likely case
The investment model had a base case. It did not have a downside case. We reviewed the base case, confirmed it produced an acceptable return, and proceeded. We did not systematically ask what the return looked like if growth was half what we modelled, if the exit multiple contracted more significantly than expected, or if the capex requirement was higher than the maintenance depreciation suggested.
A properly constructed investment model always includes at minimum three scenarios: base case, downside case, and upside case. The downside case should be genuinely conservative — not the base case with 2% shaved off the growth rate, but a realistic assessment of what happens if things go meaningfully wrong. If the downside case still produces an acceptable return, proceed with confidence. If the downside case produces a return that is unacceptable — or a loss — the deal needs to be structured at a lower price or walked away from.
In this deal, a properly constructed downside case — which the corrected model three years later effectively represented — would have shown a return well below our hurdle rate. That should have been a signal to renegotiate, not to proceed.
Error 3: We did not model free cash flow — we modelled EBITDA
EBITDA is not cash. We covered this in detail in our Business Valuation guide and in the financial literacy content of this blog. When building an investment return model for an acquisition, the metric that matters is free cash flow — EBITDA minus maintenance capex minus the cash consumed by working capital growth minus tax. Using EBITDA as a proxy overstates cash generation in almost every business that has meaningful physical assets or a working capital cycle.
The correction: always build a free cash flow model, not an EBITDA model. Estimate maintenance capex independently from the depreciation charge — they are rarely the same, particularly in businesses that have been underinvesting. Model the working capital cash consumption at different growth rates. Apply a realistic tax charge. The free cash flow figure is what services the acquisition debt, funds investment, and ultimately drives the equity return. That is the number the model should be built around.
What a Proper Pre-Completion Investment Model Looks Like
The investment model that should have been built before heads of terms were agreed on this deal has the following components:
Revenue and EBITDA projections — built from current market conditions and competitive position analysis, not from the seller's historic trajectory, across three scenarios: base, upside, and downside
Free cash flow conversion — EBITDA adjusted for maintenance capex (estimated independently, not taken from the accounts), tax, and working capital movement at each growth rate
Debt service — the full debt service schedule across the hold period, including both senior debt and vendor finance, showing the cash available to equity after all debt obligations are met
Exit valuation — modelled at the appropriate multiple for the projected earnings level in each scenario, not at a single fixed multiple applied to all scenarios
Equity return — IRR and multiple of money on equity invested, in each scenario, with a clear hurdle rate that the deal must clear in the base case and must not fall catastrophically below in the downside case
Sensitivity analysis — showing the impact on equity return of variation in the three most sensitive assumptions: growth rate, exit multiple, and free cash flow conversion. This tells you which assumptions matter most and therefore which ones deserve the most scrutiny in due diligence
Building this model takes a full working day for someone who knows what they are doing. It is the most important day of work in any acquisition process. The cost of not doing it — or of doing it optimistically rather than honestly — is demonstrated by this deal.
The Harder Lesson: Knowing When to Walk Away From a Good Business
The most difficult aspect of this deal, with hindsight, is that the business was genuinely good. The brand was real. The customer base was clean. The management team was capable. The sector, while moderating, was not in decline.
What made it the wrong deal at that price was the combination of a premium multiple and a realistic growth outlook that did not support it. The business was a 4x to 4.5x business masquerading as a 6x business, partly because of the seller's presentation and partly because of our analytical optimism.
The discipline that this deal taught, and that I apply to every deal I have been involved with since, is this: a good business at the wrong price is still the wrong deal. The quality of the business determines whether it is worth owning. The price determines whether it is worth buying. These are not the same question, and confusing them is one of the most common and most expensive errors in acquisition entrepreneurship.
If the investment model — built honestly, with realistic assumptions, and stress-tested against a genuine downside case — does not produce an acceptable return at the proposed price, the answer is to renegotiate the price, not to find reasons why the assumptions might be more optimistic than they appear.
Build the Model. Be Honest About It. Then Decide.
The lesson from this deal is not that acquisitions are dangerous or that six-times multiples are always wrong. The lesson is that every acquisition requires a rigorous, honest investment model built before commitment — and that the model needs to be the basis for the decision, not a document constructed after the decision to justify it.
If you are looking at a deal and want a second opinion on the investment model — or if you want to build a proper acquisition return model for a business you are evaluating — get in touch directly. This is the work that prevents the most expensive mistakes in acquisitions.
Download the Dealwise Business Valuation Playbook for a detailed framework on how to assess EBITDA quality, free cash flow conversion, and the realistic multiple range for any UK SME business.
Download the Business Valuation Playbook at www.DealwiseAdvisory.co.uk
Contact Steve at [email protected] to discuss the investment model for a specific deal
WhatsApp Steve on +44 7930-857243
