The gap between a Series A engineering organisation and a Series B one is rarely about headcount. It is about which architectural decisions were deferred and how expensive they have become.
Diagnose before you prescribe
A fractional CTO engagement that opens with a reorganisation is a failed engagement. The first four weeks should produce one artifact: a written, prioritised map of where the codebase resists change and what each of those points costs per sprint.
The modularity cliff
Most Series A codebases are a well-factored core surrounded by a growing shell of expedient integrations. The cliff arrives when a change to the core requires touching the shell in more places than a single team can hold in their heads.
- Measure change amplification: files touched per feature, tracked over time.
- Extract the seams that hurt most, not the ones that are easiest to extract.
- Resist the full rewrite; institutional investors read a rewrite as a stalled roadmap.
Handing it back
The engagement succeeds when the internal team can make the next three architectural decisions without you. Write the decision records, name the successor, and set a date.