How Software Buyers Assess ARR, IP and Engineer Retention

What to take away

Why San Francisco software targets price on ARR, not EBITDA

San Francisco software companies rarely sell on EBITDA. Most sit in SoMa, Mission Bay or the Financial District, funded by Sand Hill Road rounds, and they spend that money on sales teams and engineers at Bay Area wage levels. The result is little or no profit.

Buyers in the Bay Area instead anchor on annual recurring revenue, or ARR, because it is the clearest signal of a repeatable customer base.

EBITDA still matters for mature software businesses, but it sits behind ARR in the pricing order. A buyer looking at a $20 million ARR company with 15 percent EBITDA margin will usually start from the revenue line and then adjust for profitability, not the other way round.

That is a deliberate choice: recurring revenue is easier to forecast than a profit figure that swings with hiring.

This is where the broader logic of valuation methods matters. Two definitions of profit, accounting profit and economic profit, can lead to different answers, and software sellers should know which one the buyer is using before arguing about the multiple.

San Francisco also has a dense buyer population: venture-backed strategics, private equity firms with software platforms, and Bay Area acquirers that buy to fold a product into an existing suite. That competition supports revenue-based pricing, but it does not remove the need for clean books under GAAP.

What moves software ARR multiples

ARR multiples vary widely by size and quality. A small San Francisco software company with a few million in recurring revenue might attract a low single-digit multiple, while a larger, faster-growing target with strong net revenue retention can command a double-digit multiple. These are ranges, not rules, and the spread is wide.

Bay Area acquirers build their starting range from comparable company analysis, drawing on public software companies based in California and on announced private deals in the city and on the Peninsula. Classifying software targets properly matters here, and the SEC's Standard Industrial Classification code list is a standard reference for grouping companies into the right industry codes.

You can find it at SEC.gov | Standard Industrial Classification (SIC) Code List.

Public deal and valuation research is also available through SEC.gov | Search Filings, which lets a seller study how comparable public software companies disclose revenue and metrics. That research feeds the multiple, even for a private San Francisco target.

A simple worked example shows the arithmetic. Take a Bay Area software company with $10 million ARR, 30 percent growth, 80 percent gross margin and no customer above 8 percent of revenue. A buyer might apply a 6x multiple, giving $60 million enterprise value.

If growth falls to 15 percent, the same buyer might apply 4x, giving $40 million. The $20 million gap is the price of slower growth, not a change in the product.

For a structured walkthrough of how these figures combine, see the valuation methods comparison.

Valuing source code, patents and other intellectual property

Source code is the core asset in most San Francisco software deals, but it rarely appears as a separate line in the multiple. Buyers usually treat it as part of the operating business and test it during technical due diligence. Quality, documentation, dependency risk and security flaws all affect whether the buyer trusts the ARR forecast.

Patents are different. A patent portfolio can be valued separately using relief-from-royalty or cost-based approaches, and the result is often added to the enterprise value rather than baked into the ARR multiple. A Bay Area acquirer buying a software target with defensible patents may pay a premium, but only if the patents cover something the buyer needs.

Other intellectual property includes trademarks, copyrights in code and content, confidential know-how, and data sets. Each has its own valuation logic. The American Society of Appraisers sets business valuation standards that appraisers follow when these assets are valued for tax or deal purposes, and the IRS takes an interest when intellectual property is transferred as part of a sale.

Sellers should separate the intellectual property schedule from the revenue story. A buyer will ask which assets are owned, which are licensed, and which were created by contractors.

Missing assignments are a common defect in San Francisco software companies, and fixing them before a sale is cheaper than negotiating after the letter of intent. The due diligence process is where these gaps surface.

For a set of invented businesses that show how asset-based and income-based approaches produce different numbers, the valuation methods examples are a useful reference.

Engineer retention packages and how they enter the price

Software companies in San Francisco compete for engineers against large Bay Area employers, and buyers know that losing key engineers after closing can destroy the ARR they just paid for. Retention packages are therefore part of the deal, not an afterthought.

A typical package combines cash bonuses, equity or phantom equity, and sometimes earnout payments tied to product milestones. The cost is real. Buyers often fund retention from the purchase price, which reduces the cash the seller receives, or from an earnout, which shifts the risk to the seller.

Benchmarking those packages requires local wage data. The Bureau of Labor Statistics publishes industry employment and wage data through the Quarterly Census of Employment and Wages, available at Quarterly Census of Employment and Wages : U.S. Bureau of Labor Statistics.

Sellers can use that data to check whether their engineer compensation sits above or below the Bay Area market.

The methodology behind those statistics is explained in the Handbook of Methods : U.S. Bureau of Labor Statistics, which helps when a buyer challenges the numbers.

Retention usually covers founders and a named list of engineers for a set period, often two to three years. The size of the pool depends on how concentrated the technical knowledge is. If one engineer wrote most of the source code, the buyer will demand a larger package and a longer lock-up.

Due diligence on churn, concentration and contract terms

Due diligence in a software deal tests whether the ARR is durable. Buyers look at churn, both gross and net, by cohort. A San Francisco software company with strong gross churn but strong expansion can still show good net retention, and buyers will separate the two.

Customer concentration is the second test. A target whose top five customers make up half of revenue is riskier than one with a long tail, and the multiple reflects that. Buyers also read contract terms: renewal dates, termination rights, price escalation clauses, and whether customers can leave after an acquisition.

Financial due diligence covers revenue recognition under GAAP, deferred revenue, and the treatment of professional services revenue. Buyers often discount services revenue because it is not recurring. Legal due diligence covers intellectual property ownership, open-source licenses in the source code, and employee agreements.

Regulatory items can appear too. The Hart-Scott-Rodino Antitrust Improvements Act requires pre-merger notification for deals above certain thresholds, and FINRA oversees broker-dealers involved in M&A. Sellers should expect questions about IRS treatment of the transaction, especially where intellectual property is transferred.

Where the multiple breaks down in a down round market

A down round market hits San Francisco harder than most places, because so much of the local software base depends on venture money from Sand Hill Road and the Peninsula. When that funding tightens, buyers turn selective and multiples compress.

Several things break the multiple in this environment. Slower growth, higher churn, and a shorter cash runway all reduce buyer appetite. So does a cap table with liquidation preferences that leave common shareholders with little. Preferred stock terms can consume most of a lower sale price before founders and employees see anything.

Buyers also use down markets to push for earnouts and retention-heavy structures. The headline multiple may look unchanged, but the cash at closing is lower and more of the price depends on future performance. Sellers should model the after-tax proceeds under each structure, not just the enterprise value.

Automated tools that produce a number from revenue and growth inputs often miss these adjustments, so sellers should check the acquisition financing points before relying on them. They cannot price liquidation preferences, retention costs or contract risk. In a down round, those items decide the outcome.

Common questions