How to Sell a Software or SaaS Company — Conclave Partners

Almost every software owner begins the conversation with one number: annual recurring revenue. It is the right number to start from, and it is almost never the number a buyer ends up multiplying.

Between the ARR in the founder's deck and the ARR in the purchase agreement sits a series of deductions. Revenue that is recurring in name but cancellable monthly. Implementation and customisation fees counted as subscription. Contracts that renew at the customer's discretion and have not yet been tested. A large account that has quietly signalled it is leaving. Each of these is found in diligence, and each of them moves the multiple rather than the revenue line — which is why a €400,000 correction to ARR can cost seven figures of enterprise value.

This article sets out what buyers pay for in a software business, what the market is actually paying in 2026, how churn and custom development compress the multiple, and what an owner should fix in the year before going to market.

The gap between perception and reality is wider in software than in any other sector we work in, because owners anchor on 2021.

Software Equity Group's data for the second quarter of 2026 puts the median public SaaS company at 3.2x trailing twelve-month revenue, down from 5.7x a year earlier. Private M&A held up better: the median SaaS acquisition closed at 4.0x TTM revenue, against 4.2x a year before, with the average at 6.2x — the gap between median and average being the handful of premium assets that pull the mean upward.

Volume, meanwhile, is at a record. SEG tracked 2,784 SaaS transactions over the trailing twelve months to 2Q26, the most active period in its dataset, with 698 in the quarter alone, up 9.6% year on year. This is the important structural point for owners: there is no shortage of buyers or of completed deals. There is a shortage of buyers willing to pay 2021 prices, and the market has become markedly more selective about what quality means.

Two clarifications save a lot of wasted negotiation. First, a multiple of ARR and a multiple of trailing revenue are not the same thing, and the difference is roughly one year of growth — quoting an ARR multiple against benchmarks built on trailing revenue overstates value. Second, published private benchmarks for smaller companies sit well below headline figures: for bootstrapped businesses in the low single-digit millions of ARR, a 3x to 5x range is the realistic starting point, not the 8x to 12x that circulates in founder communities. At Conclave Partners we would rather set that expectation in the first meeting than discover it in week ten of a process.

A buyer's first task in diligence is to rebuild ARR from the contracts rather than accept it from the dashboard. Owners should do the same exercise first.

Contracted recurring revenue is what customers are legally committed to pay over a defined term. Implied ARR is a monthly figure multiplied by twelve. A book of month-to-month customers can be perfectly healthy and still be valued differently from annual contracts with auto-renewal, because the buyer is pricing what survives a change of ownership and a price increase.

Implementation fees, data migration, training, configuration and paid support are ordinary parts of a software business and legitimate revenue. They are not recurring, and buyers strip them out. Where onboarding fees have been amortised into the subscription line, expect the restatement to be uncomfortable.

Consumption revenue is real but volatile, and buyers discount it relative to committed subscription. Overages that arrived because one client had an unusual year are normalised away. Perpetual licence sales and hardware resale are valued as a separate, lower-multiple stream.

If ARR quality decides what gets multiplied, retention decides the multiple itself. Nothing else in a software business moves it as much.

SaaS Capital's 2026 survey of more than 1,000 private B2B SaaS companies with $3–20 million of ARR gives the current benchmarks: median net revenue retention of 103% and median gross revenue retention of 91%, with the top decile at 117.9% and 100% respectively. Median growth in the same sample fell to 15% from 20% the previous year, while retention held roughly flat.

Two things follow for a seller. First, both numbers get examined, and gross retention is the harder test. Net retention above 100% is often carried by expansion within a handful of large accounts while the base is quietly leaking; buyers separate the two precisely because a company can show 105% NRR and 82% GRR, and that combination prices as a churn problem, not a growth story. Second, being at the median is not a weakness — it is the benchmark. The mistake is presenting median retention as exceptional, which costs credibility on every other claim in the model.

Logo churn deserves separate presentation from revenue churn. A business losing 20% of its customers but only 8% of its revenue is losing small accounts, which is a manageable story. The reverse is not.

This is the single most common reason a software company gets repriced, and the one owners most often defend rather than fix.

Many profitable software businesses in the lower middle market earn a substantial share of revenue from bespoke work: features built for one client, integrations quoted by the hour, ongoing customisation that never quite becomes product. It pays well and it feels like software. To a buyer it is a services business attached to a product, and services businesses trade at a fraction of software multiples.

The effect shows up in three places at once. Gross margin falls below the 75–85% band buyers expect, and margin is one of the first screens applied. Revenue becomes headcount-dependent, so growth requires hiring rather than deployment. And the codebase fragments into per-client variants, which turns the technical diligence into a list of risks and makes the product harder to sell to the next customer without more bespoke work.

The remedy is not to refuse custom work but to separate and label it. Report product revenue and services revenue as distinct lines with their own margins. Move recurring customisation into configurable product features where possible. Where a bespoke module has value to more than one client, productise it and price it as a module. A business showing 70% product revenue at 82% gross margin and 30% services at 40% is valued on two comprehensible streams; the same business reporting one blended number is valued on suspicion. In our work at Conclave Partners this separation, done a year ahead, is frequently worth more than a year of ARR growth.

Customer concentration in software carries the same logic as in any other sector, with one aggravating factor: software contracts are usually easier to exit than supply relationships. Below roughly 10% of ARR from any single customer a business is treated as diversified; between 20% and 30% concentration becomes an explicit pricing issue; above 30% a meaningful share of buyers withdraw or restructure the deal around the risk rather than negotiate the multiple.

The contract file matters as much as the concentration. Buyers examine notice periods, auto-renewal mechanics, price-increase rights, assignment and change-of-control provisions, uptime commitments and any service credits, and unusual termination-for-convenience clauses granted to close a large deal years ago. Contracts that cannot be assigned without consent create the same problem here as supplier agreements do in distribution: in an asset sale they must be novated one by one.

Renewal timing is worth checking before launch. A process that runs a large renewal cohort through diligence is exposed — if those renewals slip, the buyer sees the trend line bend at exactly the wrong moment.

Technical diligence in software is rarely where value is created, but it is frequently where value is lost.

Ownership of the intellectual property is the first question, and the most common defect in owner-managed companies is contractor work performed without a written assignment of rights. Freelancers, offshore agencies and early friends-of-the-founder who wrote foundational code all need documented assignments; in several European jurisdictions the default position without a written agreement does not favour the company. This is fixable, but it is much harder to fix once the buyer has asked.

Open-source licence compliance is the second. Copyleft components embedded in a distributed product can carry obligations that a buyer's counsel will treat as a material issue, and a software bill of materials prepared in advance turns a red flag into a checklist item.

Security and data protection form the third. Buyers now expect documented access controls, a breach history, subprocessor records and a defensible GDPR position, particularly where the product handles personal data across borders. Absence of certification is not fatal at this size; absence of any documented practice is.

Where AI coding tools have been used heavily, expect specific questions about provenance and licensing of generated code, and about which model providers hold what rights over inputs. Buyers ask this routinely now.

The buyer universe has shifted, and it favours prepared sellers.

Financial buyers dominate. SEG recorded private-equity and PE-backed strategic acquirers at 59% of SaaS transactions in the second quarter of 2026 — the buyer of a lower-middle-market software company is now more likely to be a fund or a fund's platform than an independent trade acquirer. These buyers are quantitative, fast when the data is clean, and unforgiving about revenue quality. They also frequently structure around risk with earn-outs and rollover equity rather than walking away.

Strategic acquirers pay for what they cannot build quickly: a customer base in a segment they lack, a compliance-heavy integration, a data asset, or a team. They tolerate weaker metrics if the fit is right, but they take longer and involve more stakeholders.

A third category, consolidators of small vertical software businesses, has expanded considerably. They buy at disciplined multiples, close reliably, and are often the right answer for a founder who wants certainty over maximum price. Knowing which of the three a business realistically appeals to — before writing the materials — is most of the work, and it is where Conclave Partners spends the first weeks of a software mandate.

Rebuild ARR from the contracts. Produce a customer-by-customer schedule showing contracted value, term, renewal date, notice period and any non-standard clauses. This document will be requested; having it before the process starts changes the tone of every subsequent conversation.

Separate product revenue from services revenue, with gross margin for each, and restate at least two years of history on the same basis. A one-off reclassification in the final year looks like preparation for a sale.

Present retention properly. Report gross and net revenue retention and logo retention, defined consistently, calculated the same way every month, with the definitions written down. Buyers care less about the level than about whether the calculation survives their own recomputation.

Close the IP and licensing gaps. Collect assignment agreements from every contractor who has written code, produce a software bill of materials, and resolve any copyleft exposure. Twelve months is enough; four weeks is not.

Reduce founder dependence. If the founder is the lead architect, the main salesperson and the escalation point for the largest accounts, the buyer is acquiring a job rather than a company. Documented systems, a second technical decision-maker and customer relationships held by the team change the risk assessment materially.

Clean the accounts on the same principle that applies in every sector: adjustments a buyer can verify are accepted, and adjustments supported only by the owner's word cause the buyer to discount the whole earnings figure. For software this includes capitalised development costs, which buyers routinely restate. Sellers who ask Conclave Partners where to begin are pointed at the ARR schedule and the revenue split first, because everything downstream depends on those two documents holding up.

A prepared software business typically takes six to nine months from launch to completion — slightly faster than physical-asset sectors, because there is no property, inventory or plant to inspect, and because financial buyers move quickly when the data room is complete.

Expect the diligence to be deeper in three specific places: the ARR rebuild, the retention recomputation, and the technical and IP review. Expect also that exclusivity is where leverage disappears. Once one buyer has it, competitive tension is gone and every negative finding is renegotiated in that buyer's favour. Enter it late, with the ARR schedule, the revenue split and the IP file already answered.

In the second quarter of 2026 the median SaaS acquisition closed at 4.0x trailing twelve-month revenue, with public comparables at 3.2x, according to Software Equity Group. Smaller bootstrapped companies typically transact below those medians, in a 3x to 5x range. Position within any range is set by retention, growth, revenue quality and gross margin.

Growing software companies are usually valued on a revenue multiple; slower-growing, profitable ones are increasingly valued on EBITDA. Around 10–15% growth many buyers will look at both and use the less flattering result, which is why margin discipline matters as much as growth at that stage.

Directly and substantially, because it changes the multiple rather than the revenue. Against 2026 benchmarks of 103% net and 91% gross revenue retention for private B2B SaaS, sitting materially below either figure moves a business out of the range most financial buyers will consider without restructuring the deal.

Yes, when it is undifferentiated from product revenue. Services trade at lower multiples than software, so a blended presentation invites the buyer to value the whole business closer to the services multiple. Separating the two lines, with margins, protects the product multiple.

Statistically, a financial buyer: private equity and PE-backed acquirers accounted for 59% of SaaS transactions in 2Q26. Strategic buyers and vertical software consolidators make up the rest, and each values the same company on different criteria.

Roughly six to nine months from launch to completion for a prepared business, with the ARR rebuild, retention analysis and technical diligence accounting for most of the timetable. Preparation before launch takes another six to twelve months where the contract file, revenue split or IP assignments need work.

Ildar Zakirov — Conclave Partners ildar@conclavepartners.com

Sergi Kosiakof — Conclave Partners sergi@conclavepartners.com