On one search into a software business preparing for an exit, the product was the lever the whole thesis rested on. What the buyers were weighing by the end wasn’t whether the product was good, it was whether the customer could build something close enough without them.
The buyer’s own technology team is the threat now
The line I get from enterprise technology leaders has changed shape. It’s some version of: why would we buy this, we know our environment, we know the regulators, we can build it ourselves. That’s not a remark about AI being impressive. It’s the person who would have been the customer telling you why they’re not going to be.
What that moves is the question, from the product to the revenue. A diligence team can confirm the software works and the customers renew. What it can’t do from a checklist is work out whether the revenue survives the buyer’s own engineering function deciding to replace it, and that’s the question sitting underneath whatever multiple is being discussed.
A business growing 50% a year that can’t find a buyer
I’ve spoken directly to a PE investor who has tried to exit a horizontal SaaS business which has grown 50% annually for five years and can’t find a buyer for it. A number like that should make a business an easy sell, and it isn’t, which tells you the market has repriced something the growth rate doesn’t capture.
What explains it is the difference between software that owns a workflow and software that sells a capability. Vertical software, built for one industry and the way that industry operates, tends to retain better because it becomes the system of record. The customer has organised its operating rhythm around it and the data has accumulated inside it, so moving off is a project rather than a decision. The model was never the moat in those businesses, which is why AI arriving doesn’t take one away.
Horizontal software selling a capability sits at the exposed end, because capability is what a model now does natively and cheaply. Where the product was the capability, the buyer’s technical lead will ask what’s left once that capability is a commodity. A business that has competed for years on the quality of its engine doesn’t have an easy answer to that, and plenty of them are still growing fast enough that the problem stays invisible until a process starts.
The verticals aren’t automatically safe either. Where the product sits at the edge of a customer’s operation instead of owning the middle of it, handling one slice of a process, it doesn’t build the same lock-in. Buyers read those as point solutions rather than platforms, and price them that way.
Efficiency doesn’t move a price
AI that takes cost out of a business is worth having and it doesn’t change what the business is worth. I don’t see anyone paying close to a full headcount for efficiency at the moment, because the return is hard to attribute and it rarely survives the next planning cycle. What moves a valuation is attributable new revenue. Efficiency gains are useful and they aren’t what prompts someone to sell, so a portfolio company framing its AI programme as margin improvement has a real benefit and a much weaker investment case than it looks.
The questions in the room have got harder
More questions get asked in a modern process than used to, partly because of AI and partly because investors have become more sophisticated about technology generally. People expect the work to have been done, and they expect an answer in the room rather than a promise to come back with one.
When an investor or an executive raises something on the technical side, what they need in the same breath is what it means commercially for the point they’re making. A technical answer on Tuesday and a commercial one on Wednesday is a worse outcome than a single answer that does both, and two answers arriving apart is how a process loses its confidence.
Two situations from my own record
The two situations I keep coming back to are different ends of the same sequence. One was a business that exited for $540M, and the work that got it there was building the senior leadership team before the process started rather than during it, because what a buyer pays a premium for at that level is the conviction that the business runs without the founder in the room.
The other was earlier in the sequence: a business heading towards a sale where the brief was a cultural overhaul, and the product was the primary engine for improving the multiple. The hire that mattered in that one was a top tier operator who could carry the overhaul.
Frequently asked questions
Where this leaves both sides
If you’re selling, the useful preparation is knowing what the buyer’s own engineers will ask and having the answer ready, because they will ask it whether or not it appears on a diligence list. If you’re buying, the work is separating the model from the business, which usually means working out how much of the price is the workflow and the data rather than the engine.
I [run technology diligence and leadership work](/what-we-do/transformation-value-creation/) on both sides of a process, and the conversation is more useful before the price is set than after.
