How small businesses are valued

Three figures are easy to confuse.

• The asking price is what the seller hopes to get. It is set by the seller, often with a broker, before any buyer has tested the figures.
• Value is what the business is worth to a particular buyer, given its earnings, its risks and what that buyer can do with it. A competitor who can cut shared costs may be able to pay more than someone buying it to run on its own.
• The sale price is what a buyer actually pays, after negotiation and due diligence, and often with part of it paid later.

A valuation estimates the second figure, usually as a range. It helps you judge the first and prepare for the third.

A buyer is paying for future earnings. The simplest way to express that is a multiple: price divided by earnings. Turn it upside down and it becomes a rough yearly return. A business bought at 3 times earnings pays back about a third of the price each year, before tax, loan repayments and reinvestment. A buyer who accepts a higher multiple is accepting a lower return, usually because the earnings look safer or are growing.

Other methods have their place:

• Asset-based values, built from what the business owns less what it owes, suit asset-heavy businesses and those with little or no profit. They often act as a floor.
• Discounted cash flow models, which forecast cash flows and discount them back to today, are more common for larger businesses with predictable cash.
• Rules of thumb exist in some trades, such as a multiple of recurring fees in professional practices. They are shortcuts, not substitutes for looking at earnings.
• Revenue multiples are used for software businesses and as a sanity check for everything else.

If a business has no earnings, an earnings multiple cannot value it. Loupe's valuation tool does not produce an earnings-based value when SDE is zero or negative. It explains why and shows a revenue multiple as context only.

The first decision is which earnings figure to multiply. SDE and EBITDA explained covers the measures in detail. In short:

• SDE (seller's discretionary earnings) suits smaller businesses where the buyer is likely to work in the business.
• Adjusted EBITDA suits larger businesses, and any buyer who will pay a manager to run the business.
• ARR (annual recurring revenue) is often used for software businesses, whose value rests on subscription revenue and its growth.

Brokers follow a similar split. In the Market Pulse survey by the International Business Brokers Association and M&A Source, deals under $2 million in purchase price are reported as multiples of SDE and deals from $2 million to $50 million as multiples of EBITDA. Online business marketplaces often quote multiples of net profit, sometimes monthly rather than yearly. A multiple of 30 times monthly profit is the same as 2.5 times yearly profit, so convert before comparing.

Loupe's valuation tool picks the basis by rule. SaaS businesses with ARR are valued on an ARR multiple and cross-checked against an SDE multiple. Every other business is valued on adjusted EBITDA if adjusted EBITDA reaches USD 500,000 or SDE is above the tool's SDE ceiling, which is USD 750,000 by default, and on SDE otherwise. A revenue multiple is calculated for every business as a sanity check only.

Bigger businesses tend to sell for higher multiples. The reasons are practical. A larger business usually has a management team, so it depends less on one person. It often has more customers, so losing one hurts less. More buyers can afford it, including private equity firms and strategic acquirers, and lenders are more willing to finance it. Its earnings have usually survived more changes in the market.

Broker survey data shows the pattern clearly. In the Market Pulse highlights for the fourth quarter of 2025, the multiples reported by deal size were 2 times SDE for deals under $500,000, 3.1 times SDE for deals of $1 million to $2 million, and 5.5 times EBITDA for deals of $5 million to $50 million. The survey covers United States deals reported by business brokers and M&A advisers, and its figures move from quarter to quarter.

Loupe's tool reflects this by sorting each business into a band before it looks up a multiple. The bands are set on the basis figure itself (the earnings, or ARR for SaaS), converted to a US dollar equivalent, rather than on the asking price, so the price you are testing does not decide the multiple used to test it. Businesses are also grouped by model, because the same earnings carry different risks in different models:

• main street and services, which covers every model without its own rows, including agencies, trades, retail, hospitality, healthcare practices, manufacturing, distribution and logistics
• ecommerce
• content, media and apps, whose multiples are set lower to reflect the risk of relying on search traffic
• SaaS

Each group and band has a low, likely and high multiple in a versioned benchmark set, with a source and date stored against every row. The methodology page explains the benchmark sources and their dates. Above the top band for any basis, the tool uses the top row and warns that the business is outside its range.

Two businesses of the same size and type can deserve very different multiples. The difference is quality, which in practice means how confident a buyer can be that the earnings will continue under a new owner.

Figures from sold online businesses show how far prices can fall when buyers lose confidence. Empire Flippers, an online business marketplace, publishes average sale multiples of trailing twelve-month net profit on its scoreboard page. It counts any sale at 1.5 times or less as distressed, which it says usually follows a recent problem. When checked in September 2026, distressed sales averaged 1.2 times, against 2.2 times for all other sales. The groups are defined by the multiple achieved, so this is not the cost of any one problem, but it shows how wide the gap can be.

Loupe's tool turns quality into a set of percentage adjustments to the multiple. The main factors, and the direction each one moves the multiple, are:

• Revenue trend over the last 12 months against the 12 before. Growth of 5% to 20% adds 5%, and growth above 20% adds 10%. A fall of 5% to 20% takes off 10%, and a fall of more than 20% takes off 20%. See declining revenue or profit.
• Customer concentration. The larger the share of revenue from the biggest customer, the bigger the cut: 5% when that customer provides 10% to 25% of revenue, rising to 25% when it provides more than half. See customer concentration and why buyers discount for it.
• Recurring or contracted revenue. A share of 25% to 60% adds 5%, and more than 60% adds 10%.
• Owner dependence. A manager running the business day to day adds 5%. An owner who works more than 40 hours a week, or holds the key relationships, licences or skills, takes off 15%. See owner dependence and how to test it.
• Years trading. Under two years takes off 20% and two to five years takes off 10%. Over ten years adds 5%. A short record is weaker evidence that earnings will last, because a young business may not yet have faced a downturn, the loss of a major customer or a round of equipment replacement.
• Financial records. Owner-prepared records only take off 15%. Reviewed or audited accounts, or a completed quality of earnings review, add 5%.
• Dependence on one platform, channel or supplier, known legal, tax or compliance issues, and leases, licences or key contracts whose transfer is uncertain all reduce the multiple. Significant legal, tax or compliance issues also set confidence in the result to low.
• For SaaS businesses, annual growth replaces the general revenue trend, and monthly revenue churn and net revenue retention also count. For content businesses, heavy reliance on search traffic and a falling traffic trend count against the multiple. Where a rule for one business model overlaps a general rule, only the larger adjustment applies.

The adjustments are added together and the total is capped between minus 45% and plus 30%, then applied to the low, likely and high multiples alike. The upside cap is smaller because the size bands already reward larger businesses. These are starting weights that may change, so treat the percentages here as a guide to direction and scale. Each result's "Show the maths" view lists the adjustments actually applied.

Region matters too. Until a calibrated regional factor with a source is in place, results for businesses outside the United States carry a note that the benchmarks lean on US and online marketplace data.

Larkspur Landscaping is a fictional commercial landscaping business in the United States. All names and figures, including the multiples, are invented for this example. Its SDE is $300,000 on revenue of $1,500,000, and it is on the market at $900,000.

SDE of $300,000 is below the USD 750,000 SDE ceiling, and adjusted EBITDA is lower still, so SDE is the basis. Suppose the benchmark for a main street business of this size gave low, likely and high multiples of 2.0, 2.5 and 3.0. These round numbers are illustrative only and are not Loupe's benchmarks.

The quality factors:

The total is inside the cap, so each multiple falls by a fifth:

Loupe's tool rounds each value to two significant figures and shows stock at cost, where entered, as an addition rather than blending it in. If the records were incomplete or the business had known issues, confidence would be set lower, and low confidence widens the range by 10% either side.

The asking price of $900,000 is 3.0 times SDE, the top of the illustrative base range before any adjustment and well above the adjusted range. That does not prove the seller wrong. It tells you what the seller needs to show: perhaps a signed multi-year contract with the largest customer, or a manager ready to take over the relationships. If the evidence is not there, the gap becomes the basis for negotiation, or for a structure such as an earn-out tied to the largest customer staying.

Asking prices are useful context and weak evidence of value.

• They are opening positions. Sellers expect to negotiate, and many start high to leave room for it.
• They come before diligence. Problems found later, such as add-backs that do not hold up, usually reduce the price.
• The headline can include deferred money. A sale agreed at the asking price may include seller finance or an earn-out, so less is paid at completion than the headline suggests.
• Not every listed business sells. Asking prices of businesses that never find a buyer still sit in any collection of listings.

Final sale prices usually land below asking prices. That is why Loupe's valuation results show the median asking multiple of comparable live listings as a separate marker on the scale, labelled "Asking prices, not sale prices", with the number of listings behind it. Comparables share the model group, size band and region. Use the marker to see how a listing sits against its peers, not as a measure of what the business will fetch.

Remember also that a price only means something once you know what it includes. Stock, working capital, property and any debt can move the real cost a long way. Working capital, inventory and what the price includes covers that in detail.

The free valuation tool is built to be transparent and repeatable. It shows the earnings build-up, the basis and band, the base multiples, every adjustment and the final range, and it saves the engine and benchmark versions with each saved valuation so a result can be reproduced. It also flags results that need a closer look: an implied revenue multiple above 2.0 for a business other than software, a profit margin above 50% outside software and content, or a basis figure above the tool's top band, where it recommends a formal valuation.

Its results are indicative. They are not a formal valuation, and they are not financial advice. A formal valuation from a qualified valuer is worth considering when the business is larger than the tool's bands, when a lender or investor requires one, when tax, a dispute or a shareholder matter depends on the figure, or when the business is unusual enough that benchmarks say little about it.

Used well, an indicative range tells you quickly whether an asking price is in a sensible place, shows which factors are pulling the value down and gives you specific questions for the seller. The twenty-minute listing screen is a good next step once you have a range, and the methodology page explains the method, benchmark sources, dates and limitations in full.