Fractional CTO KPIs and Deliverables: How Work Is Measured
November 20, 20266 min read
Fractional CTO KPIs and deliverables fall into four groups: delivery speed, reliability, cost, and team health. The role is measured by a short monthly scorecard (deploy frequency, lead time, change failure rate, uptime, cloud spend) plus written artifacts such as a technical roadmap, architecture decision records and incident reviews.
Why a fractional CTO needs both metrics and artifacts
A part-time technology leader is in the building one or two days a week. Nobody can judge the work by watching it. That leaves two kinds of evidence. Metrics show whether the engineering system is getting faster and safer. Artifacts show the thinking: what was decided, why, and what happens next.
Use metrics alone and you reward dashboards that look good while the architecture rots. Use artifacts alone and you get tidy documents and no shipped software. A sound engagement names a handful of each in the agreement. If you are still working out what the role covers, the overview of the fractional CTO role is the place to start.
The KPI scorecard, with what good looks like
Keep the scorecard to eight or ten lines. Each metric needs a source system, so nobody is assembling it by hand.
| Area | KPI | Where it comes from | What good looks like |
|---|---|---|---|
| Delivery | Deployment frequency | CI/CD pipeline | Rising, then steady. Releases are routine, not events |
| Delivery | Lead time from commit to production | Version control and pipeline | Trending down quarter over quarter |
| Delivery | Roadmap commitments delivered | Issue tracker | Most quarterly commitments ship, and misses are flagged early |
| Reliability | Change failure rate | Deploy log and incident log | Falling as test coverage and review improve |
| Reliability | Time to restore service | Incident tracker | Shorter each quarter, with a named on-call owner |
| Reliability | Uptime against the stated target | Monitoring tool | A target exists, is written down, and is met |
| Cost | Cloud spend as a share of revenue | Cloud billing and finance | Flat or falling while usage grows |
| Security | Open critical vulnerabilities and their age | Scanner and audit findings | None older than the agreed fix window |
| Team | Open roles, time to fill, regretted departures | Hiring tracker | Hiring loop runs on a schedule, no surprise exits |
Two cautions. First, do not measure lines of code, story points per engineer or hours logged. They are easy to inflate and they punish the right behavior. Second, a fractional CTO should not be held to a metric they have no authority over. If the founder overrides every roadmap decision, "roadmap commitments delivered" belongs to the founder.
Deliverables a fractional CTO should hand over
These are the documents and decisions a company should physically have at the end of each phase.
- Technical assessment. A written review of the codebase, infrastructure, security posture and team, usually in the first 30 days. It ranks risks, it does not list everything.
- Technology roadmap. A 6 to 12 month plan that ties engineering work to business goals and shows what is being deliberately postponed.
- Architecture decision records. One page per major choice (database, cloud provider, build versus buy on a core component) with the options considered and the reason for the pick.
- Technical debt register. A ranked list of known shortcuts, each with a cost of leaving it and a rough cost of fixing it.
- Incident reviews. A blameless write-up after each significant outage: timeline, root cause, follow-up actions with owners.
- Hiring loop. Role scorecards, interview stages, a take-home or pairing exercise, and a leveling guide.
- Vendor and agency review. An evaluation of outsourced development shops and major software contracts, with keep, renegotiate or replace recommendations.
- Diligence pack. When a raise or sale is coming, the system diagrams, security policies and dependency lists an investor's technical reviewer will ask for.
A monthly cadence that keeps the work visible
The rhythm matters as much as the list. Here is a cadence that fits a one-to-two-day-a-week engagement.
- Weekly: an engineering leads meeting (blockers, incidents, release plan) and a 30-minute one-on-one with the CEO.
- Every two weeks: an architecture review where engineers bring proposed designs before building them.
- Monthly: a one-page engineering report to the leadership team. Scorecard at the top, three things shipped, three risks, one decision needed.
- Quarterly: a roadmap reset with product and sales, a debt register review, and a board slide on technology risk.
Here is a worked example of the monthly report's risk section, for a hypothetical 12-engineer software company: "Risk 1: only one engineer understands the billing service. Action: pairing rotation starts this sprint, runbook due by month end. Risk 2: the database is at a size where the next plan tier doubles the bill. Action: archiving job scoped, decision on build versus managed service needed from the CEO by the 15th." That is the level of specificity to expect. Vague lines such as "improve scalability" are a warning sign.
What to expect in the first 90 days versus steady state
Early on, artifacts lead and metrics lag. In the first month you should see the assessment and a baseline scorecard, even if the numbers are ugly. A baseline that reads "we deploy twice a month and nobody tracks failures" is useful. It gives you a starting line.
By month three, the roadmap, the debt register and the hiring loop should exist, and at least the delivery metrics should be moving. Reliability and cost usually take longer because they depend on work that has to be scheduled around feature delivery. The guide to the first 90 days of a fractional engagement lays out that sequence in more detail.
In steady state, the question changes from "what did the CTO produce" to "can the team run without the CTO in the room." A healthy sign is the engineering leads presenting the monthly report themselves.
How to write these into the agreement
Put the scorecard and the deliverables in the statement of work, with dates for the first version of each. Name the review point (commonly day 90) where both sides check the list and reset it. Tie the retainer to time and outcomes you both control, not to a single number like uptime that one bad vendor outage can wreck. This is general guidance, not legal advice, so have an attorney review the contract itself.
If you are the executive, propose the scorecard yourself in the proposal. It signals that you expect to be measured, and it stops the client from inventing a metric later. Guidance on boundaries lives in the piece on scoping a fractional engagement. One last point for both sides: agree in writing who owns production access and on-call. A part-time CTO designing the on-call rotation is normal. A part-time CTO being the on-call rotation is a scope failure waiting to happen.
For experienced executives
Get found and hired for fractional work.
- A profile companies can find
- A calendar they can book
- Paid to your bank, you keep 90%
Frequently asked questions
What KPIs should a fractional CTO be measured on?+
Measure a fractional CTO on a short scorecard covering delivery, reliability, cost and team: deployment frequency, lead time to production, change failure rate, time to restore service, uptime against a written target, cloud spend relative to revenue, aging critical vulnerabilities, and hiring progress. Skip lines of code and hours logged.
What deliverables does a fractional CTO produce in the first 90 days?+
Expect a written technical assessment and a baseline scorecard in the first month. By day 90 there should be a technology roadmap, a ranked technical debt register, a working hiring loop, and the first architecture decision records. Delivery metrics should be moving by then, while reliability and cost usually take longer.
How often should a fractional CTO report to the CEO?+
A weekly 30-minute one-on-one with the CEO plus a one-page monthly engineering report works for most part-time engagements. The monthly report carries the scorecard, three things shipped, three risks with owners, and one decision needed. Add a quarterly roadmap reset and a board slide on technology risk.
Should a fractional CTO be on call for production incidents?+
Usually no. A fractional CTO should design the on-call rotation, the escalation path and the incident review process, then make sure full-time engineers staff it. Agree in writing who holds production access and who gets paged. Making a part-time leader the only responder is a scope problem.