Micro-SaaS sells for 2.86x SDE, not 4x ARR

A founder with a $40,000-a-month software product almost always arrives at the same starting number. Recurring revenue trades at four or five times ARR, the reasoning goes, so a product doing roughly $480,000 a year should be worth somewhere above $2 million. Then the offers come in between $500,000 and $700,000 and the founder concludes the buyers are lowballing.

They usually are not. The ARR multiple and the offer are not two opinions about one price. They are two different markets, quoted in two different units, with two different sets of buyers. I price private companies for a living, more than a thousand of them, and this particular mismatch destroys more small software deals than any other single misunderstanding. Here is what the transaction data actually shows, and how to convert between the two bases so you can tell a low offer from a fair one.

What small software actually clears at

Start with the cleared prices. In my Felpel Private Company Index, built on 7,601 small-business transactions with a usable EV/SDE multiple, the micro and online SaaS cell holds 523 deals. Its median is 2.86 times seller's discretionary earnings, with the middle half of the cell falling between 2.20x and 3.80x.

That number is barely above the 2.67x median across the whole index, and it sits below several distinctly unglamorous sectors: retail and hospitality at 3.37x across 479 deals, manufacturing at 3.36x across 66, healthcare services at 3.35x across 139, professional and B2B services at 3.09x across 452. Software beats e-commerce, at 2.42x across 4,584 deals, and that is roughly the extent of the premium. At Main Street scale the recurring-revenue premium that dominates software investing is worth about a fifth of a turn.

A second marketplace tells the same story from a different angle. Acquire.com's biannual acquisition multiples report, published in February 2026 and covering 2025, found that confirmed SaaS sales on its platform closed at a median profit multiple of 3.9x in both 2024 and 2025, with average seller-reported margins of about 71 percent and an average 81 days on market. Different marketplace, different earnings definition, same order of magnitude: a small software business is bought for a few years of its earnings.

Where the 4x ARR number comes from

The ARR multiples are not invented, and this is the part founders get right. SaaS Capital's 2025 work on private SaaS valuations put the SaaS Capital Index of public software companies at 7.0 times current run-rate annualized revenue at the end of 2024, and derived from it a predicted private multiple of 4.8x for bootstrapped companies and 5.3x for equity-backed ones.

Two words in that sentence carry all the weight. The first is predicted: these are model outputs fitted to public comparables and private growth and retention data, not a median of observed private transactions. The second is the population. SaaS Capital's 2026 benchmarking, drawn from more than 1,000 private B2B SaaS companies, describes the bootstrapped cohort as companies with $3 million to $20 million of ARR, growing 15 percent a year, with 103 percent net revenue retention and 91 percent gross revenue retention.

It is also worth noticing which direction the anchor itself has moved. Acquire.com's report tracks the public software comparables that sit behind every private quote, and puts them at roughly 17 times revenue in 2022, around 7 times through 2023 and 2024, and about 5.5 times by the end of 2025. The rule of thumb a founder repeats today was usually formed when the reference point was two or three times higher than it is now. Rules of thumb are sticky in a way that transaction prices are not.

A founder-run product at $480,000 of ARR is not a small version of that company. It is a different asset with a different buyer. The company in the SaaS Capital sample is bought by a fund or a strategic acquirer underwriting a revenue stream that grows without its founder. The $480,000 product is bought by one person who will run it themselves and wants to know what it pays them.

The margin bridge, and what it costs to ignore it

The two bases reconcile with one line of arithmetic: an earnings multiple times the earnings margin equals a revenue multiple. Nothing else is needed, and running it in both directions is the fastest way to sanity-check any quoted number.

Take an illustrative product with $480,000 of ARR that throws off $216,000 of SDE, a 45 percent margin after the owner's own compensation is added back. At the index median of 2.86x SDE it prices at about $617,800, which is 1.29 times revenue. Now run the ARR multiple instead. At 4.8x ARR the same product would fetch $2.3 million, which is 10.7 times its own earnings. That is not a negotiating gap. It is a different asset class.

Margin is why the Acquire.com figure and mine are closer than they look. At the roughly 71 percent average margin Acquire reports, a 3.9x profit multiple implies about 2.8 times revenue. At the 45 percent margin in my example, 2.86x SDE implies 1.29 times revenue. Most of the distance between those two revenue multiples is not disagreement about software. It is the margin in the denominator, plus the fact that a marketplace of screened, profitable, high-margin software businesses is a different sample from a broad index of small-business transactions. Whenever two multiples disagree, check the basis before arguing about the number. ARR, revenue, SDE, adjusted EBITDA and seller-reported profit are five denominators, and a multiple quoted without its denominator is not information.

What actually moves the price is size

If the recurring-revenue label is worth a fifth of a turn, what is worth more? Scale, by a wide margin. On an EV/EBITDA basis the index shows multiples climbing from 2.9x below $5 million of enterprise value to 6.4x between $5 million and $50 million, and 12.3x above $50 million. That basis change matters and I state it deliberately: EBITDA removes the owner's compensation, so EBITDA multiples read higher than SDE multiples for the same company, and 98 percent of the EV/SDE sample sits under $5 million.

Even allowing for that, the shape is unambiguous. The same business is worth a materially higher multiple of its earnings when it is bigger, because scale is what converts an owner's job into an asset a fund can underwrite. Empire Flippers' own marketplace data says it too: across 161 businesses sold in 2025, deals under $300,000 sold at an average 22.42 times monthly earnings, deals from $300,000 to $1 million at 26.69 times, and the eleven deals above $1 million at 35.09 times.

This is the single most useful thing a micro-SaaS founder can know before they list. The path from 2.86x to something better does not run through the pitch. It runs through the size and the transferability of the earnings.

The ask is not the clear

There is a second gap sitting underneath all of this, and it is the one I spend most of my time on with buyers. Asking multiples are systematically above clearing multiples, and both of the datasets I have used here show it independently.

Empire Flippers reported an average listing multiple of 28.69 times monthly earnings against an average sale multiple of 23.93 times for 2025, with an average listing price of $310,037 against an average sale price of $270,802. That is a gap of about 17 percent in multiple terms, on an average 101.8 days on market. In my own index, active listings ask a median 3.26 times SDE while verified-sold deals clear at 2.58 times.

Both comparisons are between populations, not within deals, and the distinction matters enough that I state it on the index itself. The businesses that never sell stay in the asking sample forever. That is precisely why the gap is useful to a buyer: it is a measure of how far the market of offers sits above the market of closes, which is not the same question as how much a given seller eventually came down.

Thin liquidity is part of the price

One more piece of the picture rarely makes it into valuation posts. The buyer pool for a small software business is thinner than the enthusiasm around the category suggests.

Empire Flippers took 8,540 seller submissions in 2025, passed 843 through initial vetting, listed 490 and sold 161. Its five most-sold business models that year were Amazon FBA at 36.1 percent of deals, display advertising at 13.3 percent, direct-to-consumer e-commerce at 9.6 percent, YouTube channels at 8.4 percent and dropshipping at 5.4 percent. Software did not make the list. Structure shows up in the terms as well: a third of their 2025 deals carried seller financing or an earnout, with sellers receiving about 68 percent of the price in cash at closing.

There is a forward-looking argument in that report worth repeating even though it is opinion rather than data, because it comes from the people running the auctions. Empire Flippers' view is that the moat small software relied on, the difficulty of writing the software at all, is being eroded by AI coding tools, and that valuations on the smaller end are the ones most exposed if that continues. I would not price a specific deal off a prediction. I would treat it as one more reason to underwrite retention and switching costs rather than the code itself, because the code is the part of a micro-SaaS that is getting cheaper to reproduce.

None of that says software is a bad business to own. It says the exit for a small one is a thin, slow, negotiated market, and thin markets price at a discount to the enthusiasm of the people inside them.

How I price one

1. Pick the basis first, and write it down. Below roughly $1 million of earnings the buyer is an operator and the working number is SDE. Above roughly $2 million the buyer hires a manager and the number is adjusted EBITDA. Quote "2.9x SDE", never "2.9x", and restate every competing offer on one basis before you compare them.

2. Anchor on cleared prices, then sanity-check the implied revenue multiple. Price the earnings, then divide by revenue and ask whether that revenue multiple is one a real buyer has paid recently for a business of this size. If it is not, the earnings number or the multiple is wrong.

3. Underwrite retention, not growth. The SaaS Capital cohort holds 103 percent net revenue retention. A micro product usually cannot show anything comparable, and a growth chart is not a substitute. Churn by cohort, with the raw export behind it, moves my number more than any narrative on the deck.

4. Price what transfers, and only what transfers. Documented infrastructure, support that does not route through the founder's inbox, contracted revenue rather than month-to-month, no single channel or customer carrying the business. Every one of those closes part of the gap between an owner's job and an asset. The rest of the gap is closed by being bigger.

The founder at $40,000 a month is not being lowballed at $617,800. They are being quoted accurately in the market they are actually standing in. The useful question is not why the offer is below 4x ARR. It is what it would take to be sold in the other market, and the honest answer is usually several years of growth and one fewer dependency on the person selling.

Common questions

What multiple does a micro-SaaS business actually sell for?

In the Felpel Private Company Index, the micro and online SaaS cell of 523 transactions cleared a median of 2.86 times seller's discretionary earnings, with the middle half between 2.20x and 3.80x. That sits barely above the 2.67x median across all 7,601 transactions in the index and below hands-on sectors such as retail and hospitality at 3.37x and healthcare services at 3.35x. Acquire.com's January 2026 multiples report gives a similar picture from a different marketplace: confirmed SaaS sales at a median profit multiple of 3.9x in both 2024 and 2025. Small software is priced on earnings, not on revenue.

Why is my SaaS not worth 4x or 5x ARR?

Because that multiple belongs to a different market. SaaS Capital's 2025 private SaaS valuation work put the SaaS Capital Index of public companies at 7.0 times run-rate revenue at the end of 2024 and derived a predicted private multiple of 4.8x for bootstrapped companies and 5.3x for equity-backed ones. Those are model-predicted values fitted to companies with institutional growth and retention, and the firm's 2026 benchmarking of more than 1,000 private B2B SaaS companies describes a bootstrapped cohort at $3M to $20M of ARR growing 15 percent a year with 103 percent net revenue retention. A $500,000 ARR product run by its founder is not in that population, and the buyer for it is an individual operator pricing what the business will pay them.

How do you convert an ARR multiple into an SDE multiple?

Multiply by the margin. An earnings multiple times the earnings margin gives the revenue multiple: 2.86 times SDE on a business earning 45 cents of SDE per dollar of revenue is about 1.29 times revenue. Run it the other way and the quoted ARR multiples get expensive fast. On an illustrative $480,000 ARR product with $216,000 of SDE, 2.86x SDE is about $617,800, while 4.8 times ARR would be $2.3 million, which is 10.7 times the same earnings. Any time two multiples disagree, check the basis before you argue about the number, because ARR, revenue, SDE, adjusted EBITDA and seller-reported profit are five different denominators.

Do small software businesses sell for their asking price?

Usually below it, and the gap is visible in two independent datasets. Empire Flippers reported that across 161 businesses sold in 2025 the average listing multiple was 28.69 times monthly earnings while the average sale multiple was 23.93 times, a gap of about 17 percent, with an average of 101.8 days on market. In the Felpel Private Company Index, active listings ask a median 3.26 times SDE while verified-sold transactions clear at 2.58 times. Both are population comparisons rather than within-deal discounts, and both say the same thing: the asking multiple is a marketing number and the clearing multiple is the one to underwrite.

What actually raises the multiple on a small software business?

Size first, then transferability. On an EV/EBITDA basis the index shows multiples climbing from 2.9x below $5 million of enterprise value to 6.4x between $5 million and $50 million and 12.3x above $50 million, which moves price far more than any category choice. After that it is whatever reduces the buyer's dependence on the founder: documented code and infrastructure, support that does not route through one inbox, retention evidence rather than a growth story, contracted revenue instead of month-to-month, and no single customer or acquisition channel carrying the business. Empire Flippers also reported that trademarked businesses on its marketplace sold at materially higher average prices in 2025, which is the same principle in another form: buyers pay for what transfers.

Related reading

• Is a business worth 3x profit? The honest answer is a range.
• What is SDE, and which add-backs actually survive.
• The AI-Native Valuation Benchmark: where comps break.
• How I price a company in 48 hours.

Go deeper

If you are buying or selling a small software business and want the number underwritten rather than asserted, I give an independent, defensible range with the reasoning attached.

Tomasz Felpel is an investor, founder, and advisor in private markets and healthcare, based in New York. He is a three-time founder of Value Alpha, an AI-powered private-markets valuation platform, Sonnerie VC, an early-stage healthcare venture firm, and Pond. Previously he led corporate development and M&A at Fortune 500 scale, pricing more than 1,000 private companies. Columbia Business School EMBA. Read the full story.