madewithlove ads

Your CTO just left, and everyone says the senior devs can manage until a replacement starts. They'll keep things running. They won't make the calls a CTO makes. Tell us the excuse you keep hearing and we'll tell you why it no longer holds.

Raising soon? Due diligence is no box to tick when it finds a single point of failure in your code. Tell us the excuse you keep hearing and we'll tell you why it no longer holds.

Mike Tyson put it better than most engineering methodologies ever have. Everyone has a plan until they get punched in the mouth. For years the standard answer to uncertainty in software was to plan harder. Map everything out up front, reduce the unknowns, and only then start building. That did not work then and it works less well now, because planning was never capable of removing uncertainty on its own. It mostly relocates it. What changed underneath is the economics. Coding agents made implementation cheap, while validation stayed exactly as expensive as it always was. Once implementation is the cheap part, the bottleneck moves to deciding what is worth building at all. The response that follows from that is more autonomy and faster iteration, rather than thicker documents. Software remains cheap to change, which means being wrong is survivable as long as you find out quickly. Plan for the punch instead of trying to dodge it.

AI has made the knowledge part of consulting cheaper. Understanding a codebase, mapping dependencies, planning a refactor: work that used to take senior engineers days now starts with a few prompts. The problems did not shrink along with the knowledge barrier. Teams still route around the legacy system, architecture decisions from three years ago still go unchallenged, and everyone still knows precisely where the skeletons are. What is scarce now is the structural courage to act on what the room already agrees on. That turns out to be much harder to buy than expertise. https://andr.as/4fScQwk

There is effectively one vendor for the hardware that matters right now. Whatever your view on that, it belongs in your risk register rather than your roadmap. A single-supplier dependency underneath your core capability is a strategic fact, not a procurement detail, and it is worth writing down before somebody asks you about it. https://andr.as/4g96Vmz

We have done more than 180 audits, and in four out of five of them one person is the single point of failure. Usually that is the CTO, and often the same person also carries the product decisions, the deployments, and the phone that rings at three in the morning. Founders read that as a staffing problem. Investors read it as a discount. Key-person risk almost never shows up as an outage in the diligence report. It shows up in the deal, as a lower number, a longer earnout, or part of the money held back until the team stops looking fragile. Writing down what only one person knows takes a few weeks of unglamorous work. Not writing it down gets priced as a percentage of the company.