How Much of Software's Growth Is Bought?

Software companies buy revenue as well as build it, and the reported growth rate does not distinguish between the two. The industry grew 16.7% over the trailing twelve months. Set aside the revenue that the largest acquirers newly consolidated during the period and the rate falls to 15.8%. That gap - about one point, or roughly a twentieth of the reported rate - is smaller than most people expect, and it has been remarkably stable for a decade.

One ratio tells most of the story

Goodwill is what a company records when it pays more for a business than that business's net assets are worth. It is, in effect, the accumulated price of everything a company has ever bought. Divide the software industry's aggregate goodwill by its aggregate revenue and you get a measure of how acquisition-built the industry is.

That ratio has sat between 0.62 and 0.66 every year since 2017. It was 0.66 in 2017 and 0.62 in 2025, with nothing outside that band in between. Goodwill has grown at almost exactly the rate revenue has, straight through the 2021 deal boom, the 2023 drought, and the largest software acquisitions ever completed.

The stability is the finding. An industry that was progressively buying more of its growth would show a rising ratio. An industry weaning itself off acquisitions would show a falling one. Software shows neither. If anything the ratio has drifted down slightly since 2022, which means revenue is now growing a shade faster than goodwill is being added.

Why that means acquisitions barely distort the trend

Reported growth is organic growth plus bought growth. The useful consequence of a flat goodwill-to-revenue ratio is that the bought portion is roughly the same size every year - and anything present in equal measure in two years drops out of the comparison between them.

Take two years at a constant one point of bought growth. A reported acceleration from 12.2% to 13.7% is an organic acceleration from 11.2% to 12.7%. The gap between the years is 1.5 points either way. The acquisitions inflate both levels and neither difference.

Now suppose the bought share had been 0.5 points in the first year and 2.0 in the second. The same reported acceleration would conceal no organic acceleration whatsoever - the entire move would be deal activity. This is the scenario a flat goodwill ratio rules out.

So the answer to "should industry growth be restated net of M&A" depends entirely on what is being claimed. A level - software grew 16.7% - is overstated by about a point. A comparison between years is very nearly clean. A comparison between two companies is not clean at all, and that is where the real distortion lives.

There is no such thing as organic industry revenue

Organic growth is a rate, and it describes one period only. It cannot be accumulated into a level, and the reason is worth following, because it governs how the figures above should be read.

An acquisition permanently changes the base. Suppose an industry with $2,000 of revenue grows 10% to $2,200, of which nine points are organic and one is bought. The organic figure - $2,180 - describes that year and then dissolves. The following year's organic growth is measured against $2,200, because $2,200 is what the combined businesses actually have, and the acquired revenue grows from there alongside everything else.

Compound the annual organic rates for the software industry from 2015 and you arrive at $713 billion for 2025, against the $852 billion the industry actually recorded. The gap is $139 billion. But only $55 billion of revenue was acquired across the entire decade. The other $84 billion - three-fifths of the shortfall - is growth on acquired revenue that the combined companies genuinely owned and genuinely grew. Chaining organic rates denies them that growth in perpetuity, and the denial compounds.

So 15.8% is not a counterfactual. It does not say the industry would have grown 15.8% had nobody bought anything. It says the industry grew 15.8% excluding revenue newly consolidated during the period, measured against the same starting base.

The counterfactual would be meaningless in any case. Had Synopsys not bought Ansys, Ansys would still exist and still have revenue. Several of the businesses acquired over the past year were public companies in this very dataset. Their revenue does not leave the industry when they are bought - it moves under a different ticker.

This is also why the distortion does not build up. Because acquired revenue joins the base and counts as organic from the following year, a given deal lifts the reported growth rate for about two years and then stops contributing. It is a flow that washes through rather than a stock that accumulates - which is why the annual figure sits near one point year after year instead of climbing, and why goodwill and revenue have grown in step.

At the company level it is not subtle

WiseTech Global reported 73.9% revenue growth in its 2026 financial year. The company states that 7% of it was organic. The rest is e2open, acquired in August 2025, which contributed US$541 million over eleven months of consolidation.

Synopsys reported 46.3%. Behind Ansys, it grew about 2% - and its Design IP segment shrank 6% over the same period. Constellation Software reported 17.7% and describes 4% of it as organic, or 2% after currency; for that company acquisition is not an event but the entire business model. Gen Digital reported 27.1% against 6.1% organic after MoneyLion added US$823 million.

These are not marginal corrections. A company placed on a growth-and-margin grid at 73.9% sits in a different quadrant from the same company at 7%, tells a different story, and drags the cell average it lands in. That grid has since been priced: among the forty-nine companies with $150M to $350M of revenue that carry a market capitalisation, splitting at the cohort medians gives median multiples of 4.10x revenue for the high-growth, high-margin cell and 0.62x for the low-growth, low-margin one. Misplacing a company is not a cosmetic error - it is the difference between the two cells the market separates most. The industry aggregate can absorb this; a chart of individual companies cannot.

The concentration is worth noting. Of the 0.87 points of bought growth identified in the trailing twelve months, Synopsys and Constellation account for 0.52 between them. This is not a broad industry phenomenon - it is a handful of very large deals. Which is also why the median company barely moves, from 12.20% to 11.97%, while the revenue-weighted aggregate moves four times as far.

What goodwill cannot tell you

Goodwill detects that an acquisition happened. It cannot size what that acquisition did to revenue, and the temptation to convert one into the other should be resisted.

Synopsys had the largest goodwill increase in the entire data set in 2025 - 586% of its prior year revenue - and reported 15.1% growth that year, because Ansys closed with three and a half months of the fiscal year remaining. Paychex increased goodwill by 83% of prior revenue and grew 5.6%. Workday and CrowdStrike both show goodwill increases large enough to flag and both grew organically, because their acquisitions closed too late in the year to contribute revenue or were too small to matter.

Goodwill also moves for reasons that have nothing to do with buying anything: currency translation, impairment, and reclassification between goodwill and identifiable intangibles. Converting a goodwill increase into acquired revenue requires assuming a purchase multiple, and the plausible range of that assumption is wider than the quantity being measured.

Every organic figure in this article is therefore the company's own published number. Where a company does not disclose one, no estimate has been substituted.

Acquisitions are not the only reason a reported growth rate can mislead, and the second reason has nothing to do with deals. The market-data vendor behind these figures carries ANNUAL statements that disagree with its own quarterly data for a meaningful share of non-US filers. D2L's fiscal 2026 annual statement is inflated 2.19% on every line, which turns 5.9% reported growth into 8.3%. Open Text's fourth quarter holds a copy of the full year, so its four quarters sum to 1.74 times the annual figure. Of twenty companies whose data permits the test, every one of the ten US filers reconciles exactly and eight of the ten non-US filers do not.

The two failure modes point the same way. An acquisition inflates a growth rate by a knowable amount that the company itself discloses; a vendor translation error inflates it by an amount nobody discloses and no internal consistency check will catch - D2L's income statement reconciles to the dollar on the wrong figures. For industry aggregates both wash out below a point. For any statement about a single company, both have to be checked, and only one of them can be checked from the data alone.

The rule of thumb

About one percentage point of software industry revenue growth each year is bought rather than grown. Expressed as a share of the reported rate that is between 6% and 12% in normal years, and about 5% in the trailing twelve months, because the reported rate itself was unusually high. It was higher in 2017, when a single wave of consolidation pushed it above 20%, and it has not returned there since.

For anyone comparing this year to last, that is a number to know and then set aside. For anyone comparing one company to another, it is the first thing to check.

Endnotes

[1] This analysis is based on 263 active public software companies · trailing twelve month (TTM) financial results as of 1 October 2026 · Market caps as of 30 September 2026 · Source: EODHD, Worldlocity analysis

[2] Method note: bought growth is measured by substituting each company's own disclosed organic growth rate for its reported rate, holding the prior-period revenue base unchanged. Fifteen companies were verified this way for the trailing twelve months, including every flagged company above $3B of revenue - the ones large enough to move a revenue-weighted aggregate; four of the fifteen proved fully organic. Twenty-four flagged companies remain unresearched, the largest of them Check Point at $2.8B. The longer series applies a calibration factor of 1/2.1 to a goodwill-based proxy, derived by comparing that proxy against the verified year, and should be read as an order of magnitude rather than a precise figure.