No private company has a price until an acquirer writes one on a letter of intent. Until then it has a range of defensible values, and valuation work is the discipline of estimating that range honestly before someone else defines it for you. The distinction matters because owners tend to arrive at market carrying a single number, picked up from a peer’s rumored exit, an industry newsletter, or an online calculator, and a single number is almost always wrong in one direction or the other. Wrong high, it kills good deals. Wrong low, it funds someone else’s return.
Chapter 5 set the frame for this entire part of the guide: price is a multiple of earnings, so every valuation argument is a fight over two numbers. Chapters 5 and 6 settled the first, ending with Meridian Industrial Services, whose $2.4 million of reported EBITDA became $2.85 million of adjusted EBITDA after normalization. This chapter takes up the second: the four methods professionals use to estimate what multiple of those earnings the market will bear, what each method is for, where each breaks at lower-middle-market scale, and why the actual price, in the end, comes from none of them.
Key point. A valuation is evidence for a negotiation, not an appraisal to be defended. Its job is to predict what the market will pay, arm the owner to test offers against reality, and expose weak assumptions before an acquirer does. The market, not the model, sets the price.
Guideline Transactions: The Workhorse
The guideline transaction method, which advisors also call precedent transactions or private-market comps, asks the most natural question in valuation: what have acquirers actually paid for companies like this one? Find completed sales of comparable private companies, extract the multiples paid, adjust for the differences, and apply the result to the subject company’s earnings. In the lower-middle market this method carries most of the analytical weight, because it is the only one whose evidence comes from the same market the company is about to enter: real acquirers, real financing constraints, real closing conditions.
The evidence comes from three places. The first is subscription transaction databases. GF Data, an ACG company, tracks private-equity-sponsored transactions of $10 million to $500 million in enterprise value (Chapter 8 quotes its headline figure); note that its coverage begins at $10 million of enterprise value, a notch above where many lower-middle-market deals sit. Databases built on broker- and intermediary-reported transactions extend coverage into smaller deals, with the tradeoffs that self-reporting brings. The second source is the advisor’s own deal flow: the indications of interest and letters of intent from recent live processes, including the bids that did not win. This is often the best evidence in the file, because databases only ever record the winning bid of a completed deal, while an advisor who ran six processes in your sector over the past two years knows where the second and third bidders landed and which acquirers stretched. The third source is what leaks around private deals: strategic acquirers’ filings, lender publications, trade press. All of it fragmentary, all of it useful with care.
Handled naively, though, comps mislead in predictable ways, and there are five worth knowing.
Structure hides in headlines. A reported 6.5x may include an earnout that will never fully pay and a seller note at a below-market rate. The cash-at-close multiple, the number an owner should care most about, might be 5.2x, and the database will never tell you.
Nobody says whose EBITDA. A multiple is a fraction, and the denominator may be reported EBITDA, adjusted EBITDA, or a marketing number that did not survive diligence. Chapter 6’s gap between $2.4 million reported and $2.85 million adjusted is nearly a full turn of apparent multiple on the same price.
Survivorship filters the sample. Databases record deals that closed. The process that stalled at a disappointing valuation, the company that went to market and came home, the letter of intent that died in diligence: none of it reports. The visible record skews toward successes.
Size mismatch flatters. Larger companies command higher multiples, for reasons Chapter 8 explains, so a $40 million enterprise value comp says little about a $14 million company, even in the identical business.
Time decays. A comp struck in a different financing environment reflects debt costs and acquirer appetites that no longer exist. Multiples move with the cost of money, and a comp from two years ago can be an artifact of a different market.
None of this dethrones the method. It means comps are evidence to be cross-examined, not answers to be applied, and the cross-examination is most of the skill.
Guideline Public Companies: The Borrowed Ceiling
The second method borrows prices from the stock market: identify public companies in the same industry, read the multiples they trade at, and discount for everything that makes a private company different. The mechanics are easy. The comparability is the problem.
Below roughly $10 million of EBITDA, the method mostly breaks down, for two reasons. The first is that the peers are not peers. A $4 billion public facility services company shares an industry classification with a $15 million regional contractor and almost nothing else: not the customer concentration, not the management depth, not the access to capital, not the consequences of losing one key person. The second is that the discounts stack until the answer is assumption all the way down. Start at a 12x public multiple, subtract something for size, something for illiquidity, something for key-person risk, and you can land anywhere from 5x to 8x depending on judgments no one can verify. When a method requires you to subtract your way to the answer, the answer was chosen, not derived.
The method still earns its place in the workup, in two specific roles. It is a ceiling: a lower-middle-market company will rarely clear the multiple of its genuinely best public comparable, so the public number bounds the conversation from above. And it is a window into the strategic acquirer’s math. A public strategic trading at 11x EBITDA that acquires a private company at 6x has, measured on its own trading multiple, created value the moment the deal closes. That arithmetic explains why strategics can stretch when a target genuinely fits, and why this method, even when it cannot price your company, helps you understand the acquirer across the table.
Discounted Cash Flow: Right in Theory, Fragile at This Scale
A business is worth the sum of all the cash it will ever generate, discounted back to the present at a rate that reflects the risk of actually receiving it. That is the discounted cash flow method, and it is less one method among four than the theory underneath all of them. A multiple is a compressed DCF: 5.5x is shorthand for a bundle of assumptions about growth, risk, and reinvestment. When acquirers argue about multiples, they are arguing about DCF inputs without saying so.
The trouble is not the theory. The trouble is that a DCF demands three inputs the lower-middle market cannot reliably supply.
The first is a credible multi-year forecast. A defensible model needs five years of projected revenue, margins, capital expenditure, and working capital. Most companies at this scale do not budget past the current year, and the forecasts produced for a sale have a familiar shape: flat history, then a hockey stick. An acquirer will not underwrite that forecast, and the owner should not either.
The second is a discount rate. For private companies the rate is typically assembled by a build-up: a risk-free rate, plus an equity premium, plus a size premium, plus a company-specific premium. That last item is a judgment posing as a parameter. And the output is brutally sensitive to it: move the discount rate by a point, or the terminal growth rate by half a point, and the calculated value commonly swings more than a year of genuine operational improvement could move it.
The third is terminal value, the lump sum that stands in for every year beyond the forecast horizon. It is computed off the final forecast year, which is the number you know least about, and it is routinely the largest single component of the total. A valuation dominated by its least knowable input is not evidence. It is an opinion with a spreadsheet behind it.
Put together, a DCF at this scale usually confirms whatever its author already believed. Advisors know this, sophisticated acquirers know it, and a valuation presentation that leads with a DCF invites a fair question: which came first, the model or the number?
None of that makes the method useless. It makes it a specialist’s tool, and there are jobs where it genuinely earns its keep.
First, testing what a price implies. Run the model backward: at the offer on the table, what growth and margins must the company actually deliver to justify the number? If an offer only makes sense on performance the company has never demonstrated, the acquirer is either underwriting synergies or planning to renegotiate later, and it is worth knowing which. This reverse use of DCF is more valuable to an owner weighing offers than any forward valuation it produces.
Second, valuing contingent consideration. An earnout is a stream of uncertain future payments, and there is no honest way to compare a $15 million all-cash offer against $16 million with $3 million contingent except to discount the contingent piece for time and probability. That is DCF thinking, applied where nothing else works.
Third, high-growth companies. A trailing multiple punishes a company growing 40% a year, because last year’s earnings say little about next year’s. For genuinely high-growth businesses, the DCF and its cousin the forward multiple are how trajectory gets priced instead of history.
Fourth, capital-intensive businesses. EBITDA multiples flatter companies that must spend heavily just to stand still, a problem Chapter 5 flagged. A DCF cannot avoid subtracting capital expenditure, which makes it a useful corrective in both directions.
There is a quieter fifth use. An owner deciding whether to sell at all is implicitly comparing the offer against the present value of continuing to own. Making that comparison explicit, with honest inputs, is one of the most clarifying exercises in the whole decision.
Asset and Cost Approaches: The Floor
The fourth method values the company as a collection of assets rather than a stream of earnings. Take the balance sheet, restate everything to market value (the fleet at what it would actually fetch, receivables at what will actually collect, obsolete inventory at zero), subtract liabilities, and the remainder is adjusted net asset value. A variant asks what it would cost to recreate the business from scratch.
For most profitable lower-middle-market companies this method tells you little, because most of the value lives nowhere on the balance sheet. Customer relationships, a trained workforce, a reputation that wins bids, systems that run without the owner: all of it is goodwill, and goodwill is precisely what earnings-based methods price and asset-based methods ignore.
The method matters in three situations. Asset-heavy businesses, first: contractors with large equipment fleets, transportation companies, manufacturers with owned real estate, where the assets are a meaningful fraction of any sensible price and lenders will finance against them. Underperformers, second: when a company’s earnings-based value falls below its adjusted asset value, the assets set the price, and the honest conversation is whether to sell the business or the balance sheet. Floor-setting, third: in every valuation, the asset approach defines the downside, the number beneath which the question stops being what the business is worth and becomes why it is being sold as a business at all.
One distinction to keep clean: going-concern asset value assumes an orderly sale into continuing use, while liquidation value assumes speed and duress, and it is lower, sometimes dramatically. An acquirer quoting liquidation-flavored asset values for a healthy operating company is not valuing. It is negotiating.
The four methods compared: each earns a place in the workup, and none of them alone is the answer.
| Method | What it is | Best for | Weakness at LMM scale |
|---|---|---|---|
| Guideline transactions | Multiples paid in completed sales of similar private companies | Primary evidence for the valuation range in most LMM deals | Reported comps hide structure, earnings definitions, and failed processes; close comps are scarce |
| Guideline public companies | Public peer trading multiples, discounted for size and liquidity | Ceiling check; understanding a strategic acquirer’s math | Peers are not comparable below roughly $10M EBITDA; stacked discounts become guesswork |
| Discounted cash flow | Present value of forecast cash flows plus a terminal value | Earnout valuation, high-growth cases, testing what a price implies | Forecast quality, discount-rate judgment, and terminal-value dominance make outputs fragile |
| Asset/cost approach | Net asset value restated to market, or cost to recreate | Asset-heavy or underperforming companies; establishing a floor | Ignores goodwill, where most of a profitable company’s value lives |
Triangulation: Building the Range
A valuation workup at this scale is not four methods run in parallel and averaged. It is a weighing of evidence, and the weights are the judgment.
In practice, guideline transactions anchor the analysis. The advisor assembles the closest available comps, adjusts for what the databases cannot show, and forms a core multiple range. Public comparables, where genuine ones exist, bound that range from above. A DCF runs as a cross-check: if it lands far outside the comp-based range, one of the two rests on a bad assumption, and finding out which is the point of running it. The asset approach contributes the floor and, in asset-heavy cases, a real input. What emerges is a range with the evidence behind each end made explicit: the bottom defensible on conservative comps and reported earnings, the top requiring the adjusted EBITDA to survive diligence and the right acquirers to show up.
The width of the range is itself information. A narrow range says the comparable evidence is deep and the company fits a well-traded pattern. A wide range says the market will decide more than the analysis can, which usually means the company is unusual, for better or worse, or its earnings quality is unsettled.
Common pitfall. Treating a formal appraisal as a market price. Appraisals prepared for estate planning, gifting, buy-sell agreements, or option pricing answer a legal question: what a hypothetical willing buyer would pay a hypothetical willing seller, often after discounts for lack of control and marketability. They are built to be defended to a tax authority, not to predict what the strongest actual acquirer will pay under competition. Owners who anchor on an appraisal number, in either direction, misprice their own expectations before the first conversation with the market.
Presenting a range rather than a point is not hedging. A point estimate invites false precision and positional bargaining over a fiction. A range built on evidence does the opposite: it tells the owner which outcomes are realistic, signals to acquirers that the owner knows the market, and gives the negotiation factual terrain instead of two anchored egos.
Illustrative example. Return to Meridian Industrial Services from Chapters 5 and 6: $12 million of revenue, $2.4 million of reported EBITDA, $2.85 million of adjusted EBITDA after normalization. Guideline transactions for services businesses of this size and profile support, illustratively, 5.0x-6.0x adjusted EBITDA, a $14.3M-$17.1M enterprise value range. Public comparables (large facility and industrial services companies trading at roughly twice those multiples) contribute a ceiling, not a data point, once the discounts stack. A DCF built on management’s moderate-growth forecast lands near the middle of the comp range, which is treated as confirmation, not independent proof. The asset approach (fleet, equipment, and net working capital) supports roughly $6 million and matters only if the going-concern story collapses. The advisor presents $14M-$17M, anchored on adjusted EBITDA, and points out one number worth staring at: on reported EBITDA of $2.4 million, the same multiples produce only $12.0M-$14.4M. At the midpoint multiple, the add-back work from Chapter 6 carries roughly $2.5 million of enterprise value. Every figure here is illustrative; Chapter 8 covers what typical ranges look like and why they vary.
The One That Sets the Price
Everything above has a limit: valuation predicts, and the market decides. A valuation is a forecast of what a competitive process will produce. It is not a substitute for one.
The reason is that companies do not have one value. They have a different value to every acquirer who looks at them. A financial acquirer prices the standalone cash flows and the debt they will support. A strategic prices those cash flows plus the costs it can remove and the revenue it can cross-sell, sometimes with cheaper money than anyone else at the table. A private equity platform pursuing add-ons prices what the company does to its blended entry multiple. The same $2.85 million of adjusted EBITDA can carry meaningfully different prices in those three rooms, and no spreadsheet run in advance tells you which acquirers will show up, or what their alternatives look like that quarter.
What the evidence says about competition is more interesting than the folklore. In a study of 400 US takeovers, Boone and Mulherin (Journal of Finance, 2007) found that roughly half of targets were sold through multi-bidder auctions and half through single-bidder negotiations, with no significant difference in returns to selling shareholders between the two methods. The honest lesson for owners is subtler than “auctions win”: competition matters most as credible optionality. An owner with real alternatives, even in a negotiated deal, prices and papers better than one with nowhere else to go.
Read carefully, that finding is not an argument against competition. It pins down where competition does its work. What disciplines an acquirer’s price, and just as importantly the terms attached to it, is not the number of names on a buyer list. It is whether the owner can credibly do something other than accept. That credibility can come from a full auction, from two live bidders, or from a negotiated deal where the owner is visibly prepared, unhurried, and willing to keep the company. What it cannot come from is a valuation report, however well built. Optionality claimed is posture. Optionality demonstrated is price.
The two chapters that follow pick up the two halves of this conclusion. Chapter 8 shows what multiples actually look like across size bands and industries, so the range you build stands on market reality. Chapter 10 shows how a well-run sale process creates the credible optionality this chapter ends on, and turns a valuation from a prediction into an opening bid.
The Bottom Line
- Four methods estimate value: guideline transactions carry the most weight in the lower-middle market, public comps set a ceiling, DCF tests assumptions, and asset value sets the floor.
- Every comp deserves cross-examination, because headline multiples hide deal structure, undisclosed earnings definitions, and the failed processes that never reached a database.
- DCF is intellectually correct and practically fragile at this scale; use it to test what a price implies, value earnouts, and price genuine growth, not to defend a point estimate.
- A finished valuation is a range with evidence behind each end, not a number to be defended.
- Valuation predicts and the market decides: competition matters most as credible optionality, and an owner with real alternatives prices and papers better than one with nowhere else to go.