Software in Freefall: When Building Beats Buying

Software in Freefall: When Building Beats Buying

Something structural broke in software pricing over the last two years, and most vendors are still selling as if it did not.

For twenty years the logic was airtight. Building software required engineers, engineers were scarce and expensive, and therefore renting a finished product at a few hundred dollars a month was obviously cheaper than building it. That arithmetic held for so long it stopped being examined.

It no longer holds everywhere. And the places it stops holding are exactly the places where the fattest software margins live.

TL;DR — Five Takeaways
  • The cost of building narrow internal software has fallen by an order of magnitude.
  • The tools most exposed are thin workflow layers on top of data you already own.
  • The tools least exposed are systems of record, regulated software and genuine networks.
  • Seat-based pricing is the most fragile model in the market right now.
  • This is a repricing, not an extinction — but the repricing is severe.

What is actually exposed

Not all software is equally at risk. The useful question is what a vendor is really selling you.

What you are rentingExposureWhy
A workflow on top of your own dataSevereThis is now a weekend of work if you know your workflow
A dashboard over an API you already pay forSevereThe data was never theirs to begin with
A system of recordLowMigration risk, audit history and inertia are real moats
Regulated or compliance-bearing softwareVery lowNobody sane builds their own payroll tax engine
A genuine networkVery lowYou cannot rebuild other people’s participation

Notice that the exposed row is where an enormous amount of modern SaaS actually sits. A great many products are a competent interface over data the customer already owns, sold per seat, per month, forever.

If your product is a nice interface over my data, and I can now describe that interface to a machine in an afternoon, you were never selling software. You were selling the fact that building it used to be hard.

The seat is the weak point

Seat pricing assumes headcount and value move together. That assumption is dissolving. If one operator using good internal tooling covers what four used to, seat-based revenue falls even when the customer is thriving. The vendor’s revenue is indexed to precisely the thing its customers are now trying to reduce.

We looked at this inversion in more detail in Backwards — AI priced itself on consumption at the same moment legacy software ran the other way.

What we think actually happens

Not collapse. Compression, and a change in what gets paid for.

The long tail of thin tools gets absorbed into internal builds. Category leaders survive on data gravity and integration depth, but with far less pricing power on the second and third product in the bundle. And the winners are the ones that sell an outcome you cannot generate yourself — a network, a proprietary dataset, a regulatory shield.

The vendors who suffer most will be the ones whose renewal conversation has been, for years, essentially "you already have all your data in here."

How we operate against this
  • Buy systems of record. Build workflow.
  • Never rent something whose only asset is our own data.
  • Assume any tool priced per seat gets re-examined at renewal.
  • Keep an honest list of what we are not going to build.

The discipline this requires

It would be easy to read this as permission to build everything. That is a trap. Internal software has a real ongoing cost — it just does not arrive as an invoice, so it is easy to pretend it is free until it is not.

The correct posture is not "build everything." It is "stop paying rent on things that were only ever expensive because building was hard." That list is longer than it was two years ago, and it will be longer again next year.

Analysis reflects LAMPWORK’s operating view and our own experience replacing internal tooling. See our build log, "We Rebuilt Our CRM in a Weekend."

Own a brand this applies to?

We talk to founders at every stage — long before they're ready to sell.

Start a Conversation