Your CTO Left in the Middle of the Rebuild. Who Else Can Read the Map?

The platform rebuild is eight months into a fourteen-month plan. It sits on a slide in the value creation plan your sponsor showed its investment committee, and a large customer has been promised the new version by spring. Then the CTO resigns. The architecture decisions, the vendor relationships and the half-finished data migration live mostly in one person's head, and that person has four weeks of notice left. The reflex is to call a search firm that afternoon and ask for a new CTO in sixty days. In a PE-backed or founder-led company, the more urgent job is finding out who else can read the map.

Software engineers have a name for this risk: the truck factor, the number of people who would have to leave before a project stalls. In a 2016 study presented at the IEEE International Conference on Program Comprehension, researchers measured it across 133 popular open-source projects on GitHub. 65% had a truck factor of two or fewer. Losing one or two developers would put most of them in serious trouble.

Those projects have public code, open histories and many contributors. A private codebase built under one technical leader is unlikely to be better spread, though nobody has measured yours. Treat the figure as directional. The decision it points to is concrete: before you write a job description, find out how concentrated the knowledge is.

The second risk is the people who stayed. In AlixPartners' 2026 private equity leadership survey of 427 investors and portfolio executives, fielded in late 2025, 44% of portfolio company leaders said the risk of losing top performers had risen. When a CTO leaves mid-build, the senior engineers who understand the system are exactly those top performers, and they are now being called by recruiters too.

The decision that figure points to: in week one, the CEO or the operating partner, not HR, should sit down with each senior engineer and ask what would keep them through the end of the build. Some will want scope, some a clear reporting line, some money tied to the build's milestones. The conversation costs nothing, and having it in week six costs a great deal more.

The four doors

The labels are the ones we use for every open leadership seat, including a head of sales who leaves mid-quarter. Behind each one, for a technology leader:

1. Fix the plan. If the rebuild only worked with the departing CTO's personal knowledge, or the timeline was never realistic, cut scope to what the remaining team can deliver and reset the dates with the sponsor now. An independent architecture review through our advisory work tells you which parts of the plan survive. Telling the board in week two costs less than telling it in month twelve.
2. Bridge with a fractional CTO. A fractional CTO takes the architecture decisions, the vendors and the engineering cadence, reports to the board, and makes knowledge capture the first deliverable. A confidential retained search runs alongside. By month two the bridge can tell you whether you need a builder, a scaler or an integrator for add-on acquisitions, which is the brief the search needs. We described the same pattern across PE seats in The Fractional Bridge.
3. Stay fractional. Plenty of industrial, distribution and services companies run mostly configured software with a small internal team. There, a strong engineering manager plus a fractional CTO a few days a month can cover architecture, security and vendor oversight. A full-time CTO would spend much of the week managing contracts.
4. Search now. Go straight to a retained search when the platform is central to the equity story and the next 18 to 24 months need a full-time technical leader who can build a second layer of management. Check first whether your internal head of engineering is ready; sometimes the successor is already in the building. If the departure is not public, keep the search confidential, and ask any firm you interview which of your target companies it is barred from recruiting out of, the subject of The Off-Limits List.

Before you replace the person who held the map, find out who else can read it.

Three questions for the first two weeks

1. Who else can explain the system? List the five components that matter most: the core application, the data pipeline, the integrations customers depend on, the infrastructure, security. For each, write the names of people who could explain it and change it safely tomorrow. Any component with one name, or none, is your truck factor. Use the CTO's notice period for recorded walkthroughs and written decision logs on exactly those components, and pair a named engineer with each one.

2. Which dates are already promised to someone outside the company? A migration cutover, a customer commitment, a security audit, a sponsor milestone. If two or more fall inside the next six months, that is the Bridge case. If none do, you have time to fix the plan or let an internal leader step up while you decide.

3. What will a buyer's diligence team find at month 24? In a PE-backed company the technology seat is judged twice: by customers now and by the next owner at exit. Diligence looks at architecture, documentation and the depth of the team, and a single point of failure is easy to spot. The portfolio company leadership clock runs on the same calendar, so pick the door with that review in mind. If the honest answer is a platform that needs a full-time leader through the exit, you are searching, and the only question is whether to bridge first.

What this is not

This is not a claim that every CTO departure is a crisis. Some leave once the hard part is built, and a capable head of engineering steps up without much disruption. It is not an argument against hiring a full-time CTO either; for many software-led companies Search now is the right door on day one. And sometimes the right move is to change nothing for a quarter except the documentation, then decide with a clearer picture.

The rule is the same in every seat. Name the door within two weeks, write down why you chose it, and set a review date, usually the next board meeting. That page keeps a stopgap from becoming the architecture of the company.