A fractional CTO engagement that has not produced anything concrete by day 90 is unlikely to recover. The pattern is consistent enough to be worth stating plainly: engagements that drift in the first quarter tend to keep drifting, because the political capital to make hard changes is highest at the start and decays.
Here is what should actually be delivered at each mark, and the specific signals that indicate an engagement is going wrong while there is still time to correct it.
Days 1–30: an honest assessment, written down
The first month is diagnosis. Not roadmapping, not restructuring, not tool selection - diagnosis. A fractional executive who arrives with a plan in week one is selling a template, because they cannot yet know what is wrong.
By day 30 there should be a written assessment covering the state of the architecture and its genuine constraints, delivery throughput measured rather than estimated, team capability and where the single points of failure sit, the vendor and contract landscape, and the three to five things that are actually blocking progress. Written down, shared with leadership, specific enough to be uncomfortable.
- Good signal: the assessment names things people already privately knew but had not said in a room together.
- Bad signal: the assessment is a generic maturity model with your company name on it.
Days 31–60: sequencing, and one visible fix
The second month turns diagnosis into a sequence. Not a two-year roadmap - a defensible ordering of the next two or three quarters, with an explicit statement of what is deliberately not being done and why.
Equally important, something should visibly improve during this month. A build pipeline that took forty minutes now takes eight. An on-call rotation that was one person is now four. A vendor contract that renewed automatically now has an owner and a review date.
This is not theatre. The point is establishing that recommendations from this person result in changes, which is what earns the authority to propose harder things in month three. A fractional executive who spends sixty days producing only documents has taught the organisation that their output is documents.
By day 60, the team should be able to point at something and say: that is better now, and it is better because of this engagement.
Days 61–90: the structural change, and the exit plan
The third month is when the harder structural work starts - the thing identified in month one that requires real organisational change. Reshaping teams around ownership rather than technology layer. Ending a vendor relationship. Killing a project with sunk cost attached. Changing how estimates are made.
By day 90 there should also be a written handoff plan: what the permanent CTO or CIO role needs to look like, what the hiring profile is, and what the fractional executive will have completed before that person arrives. If nobody has raised the topic of the permanent hire by the end of month three, that is a warning sign about the engagement's incentives.
Signals the engagement is not working
- Everything is in progress and nothing is finished. Fractional time is scarce; a wide, shallow work-in-progress spread means priorities were never actually set.
- The engineering team has not changed how it works. Leadership can be delighted while the people doing the work notice no difference - that means change is happening in meetings, not in the system.
- Nothing has been stopped. Every honest assessment finds work that should not continue. If nothing has been cancelled, the hard conversations are being avoided.
- The fractional executive has become a dependency. If decisions now queue on their two days a week, the engagement is building fragility rather than capability.
- There is no succession conversation. A fractional engagement with no defined end is a staffing arrangement, not a leadership intervention.
What to agree before day one
Most failed engagements were mis-specified at the start rather than executed badly. Before the first day, agree on decision authority - specifically what this person can decide alone, what needs the CEO, and what needs the board. Ambiguity here surfaces during the first genuinely contested decision, which is the worst possible time.
Agree the day count and protect it. Two days a week that reliably happen beats three days that get eroded by other demands. And agree what success looks like at 90 days, in writing, in terms specific enough to disagree about.
