A service business has a ceiling.A platform does not.

Five companies that were operating businesses first, encoded what they knew, and found out what happens to the ceiling. Every figure on this page carries a source.

Verified September 2026. Where a number could not be confirmed against a primary or major-trade source, it is not here.

Why the ceiling moves

The work is the same. What you can charge for it is not.

A service business sells hours, and hours are finite. It grows by adding people, which adds cost at nearly the same rate it adds revenue. That is the ceiling, and it is a real one — it is not a failure of ambition.

Service business

Revenue scales with headcount

Each new customer needs more people
Margin is capped by what people cost
The knowledge lives in those people
Growth and cost rise together
The asset walks out at 6pm

Platform

Revenue scales with customers

The next customer costs a fraction of the last
Margin improves as you add customers
The knowledge is encoded and stays
Growth compounds instead of adding cost
The asset is on the balance sheet

The catch is that getting from one to the other is a different kind of engineering, and it is the part these five companies had to do before any of the numbers below happened.

Five that made the transition

Operating knowledge, encoded.

Ordered by how closely each one matches the pattern: an internal system, used in production, that outsiders asked for.

Illustration of a university quadrangle beside a stack of contractor documents

Internal system → spinout → acquisition

Highwire

Started inside: Harvard University

The complete arc, and it started at a university.

Harvard needed a better way to assess contractor risk and reduce insurance costs, so it built one. Contractors who went through the assessment started asking for the system themselves. In 2008, with Harvard's backing, it was spun out as an independent company. The platform now covers more than 60,000 contractors, and in October 2025 Highwire was acquired by Veriforce.

Internal operating problem→Internal system→Real production use→Outside demand→Spinout, 2008→60,000+ contractors→Acquired, 2025

Sources: Highwire, Our Story · Veriforce acquisition release, Oct 2025

Illustration of a partly shingled roof with a ladder and a phone

Roofing company → software company

CompanyCam

Started inside: White Castle Roofing, Nebraska

They went looking for the software first. It did not exist.

In 2015 Luke Hansen was trying to solve a job-site photo and record-keeping problem inside his family's roofing business. He could not find an app that did it, so he hired a local development studio in Lincoln to build one. The insight did not come from interviewing contractors — it came from being one. CompanyCam was reported at a valuation of close to $2 billion in February 2026.

Roofing business→Problem nobody sold a fix for→Built it locally→Other contractors wanted it→Independent software company→~$2B valuation, 2026

Sources: Fast Company, Feb 2026 · Startland News, Feb 2026

Illustration of a brokerage back office with contract folders and house keys

Back office → industry platform

Side

Started inside: Real-estate brokerage operations

Years of running other people's back offices, opened to the industry.

Side spent years operating the back offices of its partner brokerages — compliance, transaction coordination, payments, reporting, agent support. In July 2026 it opened that platform to independent and franchise-affiliated brokerages nationwide. The infrastructure already powered 600+ back offices and around 30,000 transactions a year before anyone outside could buy it.

Brokerage operations→Internal operating platform→600+ back offices→Refined in production→Opened to the industry, 2026

Sources: Inman, Jul 2026 · HousingWire, Jul 2026

Illustration of an accountant's desk with documents and a calculator

Ran the service to find the product

Soraban

Started inside: An AI-first accounting firm

He started the firm to learn what the software needed to be.

Enoch Ko set out to automate accounting workflow, and began by running an AI-first accounting firm himself. Operating it revealed that the real bottleneck was not the accounting — it was document chasing, client intake and interruptions. That production learning became the product. Soraban reports supporting more than 500 accounting firms across six tax seasons.

Ran the firm→Found the real bottleneck→Encoded the workflow→500+ firms→Series A

Sources: Soraban, About · PR Newswire, Series A

Illustration of a service van beside a toolbox and a heating unit

Domain knowledge → vertical platform

ServiceTitan

Started inside: Two contractors' sons

Not a spinout. The ceiling.

Ara Mahdessian and Vahe Kuzoyan grew up watching their fathers run residential contracting businesses, and built software around the scheduling, invoicing and back-office work they had seen done by hand. This one is not an internal-tool spinout — it is domain-derived platformization, and it is here to show where the ceiling sits. ServiceTitan went public on Nasdaq in December 2024.

Lived the operating problem→Built for the trade→Vertical platform→IPO, December 2024

Sources: CNBC, Dec 2024 · ServiceTitan 10-K, FY2026

What they share

None of them set out to build software.

The business came first

A roofing company, a university, a brokerage, an accounting firm. The operating problem was real before any code existed.

Outsiders asked

In every case demand arrived from outside rather than being sought. That is the signal, and it is the one worth waiting for.

Production came before product

Each system was carrying real work before anyone tried to sell it. That is what made the abstraction obvious rather than theoretical.

And the part nobody puts in the story

Between "it works for us" and "it works for them" is multi-tenancy, a core-versus-configuration boundary, and production ownership. That is the part we do.

Being honest about it

Most internal software should stay internal.

These five are survivors, and survivors are a biased sample. For every one there are many systems that were genuinely specific to one company, or whose owners did not want to run a software business — which is a legitimate answer and often the right one.

Two of these five are not internal-tool spinouts at all. ServiceTitan and, to a degree, Side are domain-derived platforms — founders who lived the operating problem rather than software that escaped a company. They are here because they show where the ceiling sits, and it would be dishonest to file them alongside Highwire. The test that matters is not whether it is possible. It is whether the problem you solved is a problem other businesses in your industry actually have.

Is yours one of these, or is it specific to you?

Seven days, $2,500. A map of what you have built, where the core-versus-configuration boundary really falls, the assumptions that break when a second business signs, and a 90-day plan. If it is not ready to be productized, we say so and you keep the map.

Get a Platform Readiness Review See what the transition involves