"You need a technical co-founder" is the most repeated — and least examined — advice in startups. Sometimes it's right. Often it's a reflex, handed out because the person advising you got one, or wishes they had. Before you give away 30% of your company or commit to a $250k salary, it's worth separating the actual jobs hiding inside the word "CTO."
Job 1 — Building. Writing the actual product. This is an engineering job. Early on it gets done by a founding engineer, an agency, or a very hands-on technical co-founder — but it is not, itself, the CTO job, and conflating them is the root of most bad hires.
Job 2 — Deciding. Architecture choices, build-vs-buy, technology bets, hiring plans, "should we rebuild or refactor." These decisions are infrequent but expensive to get wrong, and they need senior judgment applied at specific moments — not forty hours a week.
Job 3 — Watching. Continuous oversight: is the work good, is velocity healthy, is spend sane, is risk accumulating, is the agency's invoice honest. This job is constant, unglamorous, and — critically — the one that most often has nobody doing it in a non-technical founder's company.
Almost every "do I need a CTO?" agony resolves instantly once you name which job you're actually missing. So let's price the options against those jobs.
Cost: 20–50% of your company, forever. Covers: all three jobs, if you find a great one.
The right technical co-founder is the best deal on this page and the wrong one is the most expensive mistake you can make — a bad marriage with a cap table attached. The honest test: do you want this specific person as a partner in everything, including the parts that have nothing to do with code? If you're seeking a co-founder because you need code written, you're overpaying for Job 1 with permanent equity. Hire for the job instead.
Cost: $200–350k salary plus meaningful equity, roughly $21k+/month all-in. Covers: Jobs 2 and 3 completely, Job 1 partially and decreasingly.
Genuinely necessary — later. The full-time CTO earns their cost when there's an organization to lead: multiple teams, real architecture at scale, a hiring pipeline, board-level technical strategy. Hiring one to oversee three engineers and an agency is buying a jet to cross the street. Most companies don't need this role until 15–30 engineers, and a premature CTO hire often creates the bureaucracy it was supposed to prevent.
Cost: $3,000–8,000/month for a handful of hours weekly. Covers: Job 2 well. Job 3 poorly — through no fault of theirs.
For the deciding job, fractional is superb value: senior judgment, exactly when decisions arise, no equity. The structural weakness is the watching job. A person who sees your operation four hours a month sees snapshots, and the expensive problems — velocity fades, quiet quality erosion, cost creep — live precisely in the space between snapshots. If you hire fractional (often the right call), be clear-eyed that you've covered decisions, not oversight.
Cost: $150–500/month. Covers: Job 3 completely. Job 2 not at all — by design.
The watching job is continuous pattern-recognition over machine-readable records, which is to say: it's the one job on this list that software now does better than a part-time human, because software doesn't sample — it sees everything, always. What software should not do is make your architecture bets or hire your engineers. (Full disclosure: this is what we build at NovaCTO — continuous oversight, plain-English briefs. We're biased about Job 3 and honest about the boundary.)
1. What's broken this quarter — building, deciding, or watching? "Product isn't getting built" is a hiring problem (engineers or agency), not a CTO problem. "I face technical decisions I can't judge" points fractional. "I can't verify what I'm paying for" is the watching job. Solve the one that's actually bleeding.
2. What does the next big technical decision look like, and when? If a genuine fork in the road — rebuild vs. refactor, platform bet, first engineering hires — is coming within two quarters, line up senior human judgment for that moment. If the road is straight for now, you're buying oversight, not strategy.
3. Would $250k of engineering beat $250k of management? At almost every pre-scale stage, the answer is engineering. The affordable stack — continuous oversight software plus a fractional advisor on call for the big calls — covers Jobs 2 and 3 for under $100k a year, and the difference hires you two more engineers for Job 1.
The stack most non-technical founders actually need before 15 engineers: hired builders for Job 1, a fractional CTO on call for Job 2, and continuous oversight for Job 3. Total cost: less than half of one full-time CTO — with better coverage of the job that was killing you.
Every option above can be right. The only reliably wrong answer is the default one: nobody watching at all. Building without oversight doesn't feel dangerous day to day — that's what makes it dangerous. The bill arrives later, as a rebuild, a breach, a blown diligence, or an agency invoice you funded for a year past its honesty date. Pick any option on this page. Just don't pick none.
Connect a repo, get your first plain-English brief on Monday. First two briefs free.