The job description you are about to write is for a job that stopped existing around 2024. A Head of SEO for a B2B SaaS company in 2026 owns six distinct jobs, two of which did not exist when most candidates built their track record, and the honest version of the hire costs $150K+ loaded, takes about six months to find, and still leaves a third of the work uncovered. This guide is the real requirements list: what the role actually contains, what to pay, the interview that separates operators from tool-drivers, and the 90-day scorecard that tells you within one quarter whether you hired well.
We are an agency, so you know the alternative we sell. But we have also sat on the hiring side of this table, and half of the founders who reach us have just made this hire badly. The most useful thing we can do is show you exactly what it takes. If you read this and hire well, the guide did its job. If you read it and realize the list is longer than one person, that is worth knowing before the offer letter, not after.
6
distinct jobs the role now contains
$150K+
loaded annual cost: salary, benefits, overhead and tools
~6 mo
typical time to find a strong senior hire, before ramp
What does a Head of SEO actually own in 2026?
Six jobs. Write your job description against these, not against a 2019 template with “AI” sprinkled in, and be explicit with yourself about which are must-have on day one and which the hire can grow into. Nobody arrives with all six at senior depth. The hires that fail are almost always strong in the two jobs the interviewer knew how to test and absent in the two nobody asked about.
| The job | What it means in practice | How you test it |
|---|---|---|
| 1. Technical SEO | Crawlability, indexation, site architecture, JavaScript rendering, and now AI-crawler access (GPTBot, PerplexityBot, ClaudeBot). Your docs and app subdomains are part of this whether they like it or not. | Give them a real crawl of your site an hour before the interview and ask what they would fix first and why. |
| 2. Content operations | Not writing. Deciding what gets written against revenue, briefing writers and subject-matter experts, running the review gate, killing the pieces that should not ship. | Ask how they decided what to publish at their last company. Volume-based answers are a no. |
| 3. AEO and GEO | Getting the brand named and cited by ChatGPT, Perplexity, Gemini, Claude and AI Overviews. Prompt mapping, entity work, quotable page structure. The newest job and the rarest skill on the market. | The work sample below. Almost nobody has shipped results here; you are testing whether they can reason about it. |
| 4. Off-page and citations | Earning presence on the sources engines cite: reviews, communities, roundups, digital PR with original data. The slowest work and the first thing a stretched hire silently drops. | Ask for one placement or mention they earned without paying, and how long it took. |
| 5. Measurement | Citation share across engines, GA4 AI-referral tracking, branded-search proxies, self-reported attribution, and a monthly narrative leadership actually reads. Rankings screenshots do not count. | Ask what they would report if sessions fell while AI answers increasingly named the brand. |
| 6. Cross-team orchestration | AI reads your G2 profile, docs, partner listings and your founder's Reddit replies. Six teams own those surfaces and none report to SEO. The job is briefs, influence and follow-through. | Ask for a time they changed another team's output without authority over that team. |
The uncomfortable arithmetic: jobs 1, 2 and 5 have a deep talent pool. Jobs 3, 4 and 6 are where AI-era results actually come from, and the pool of people who have shipped all three is close to empty. Every hiring decision in this guide flows from that gap.
What should you pay?
US-market bands as of mid-2026, and treat them as bands, not quotes, because comp data moves. A senior IC who can genuinely run jobs 1, 2 and 5, with real reasoning ability on 3 and 4, sits around $120K to $160K base. A true Head of SEO with management experience and shipped AI-search results, to the extent that person exists, is $150K to $200K+ base. Load either number with 25 to 35% for benefits, payroll and overhead, add $10K to $20K for the tool stack (we itemized that stack here), and the honest annual figure starts at $150K and climbs fast.
Two more costs that never make the budget sheet. The search itself: a strong senior SEO takes around six months to find, and your organic program is aging the whole time. And the miss: a bad senior hire discovered at month nine costs you the salary, the recruiting fees and a year of compounding you cannot get back. The scorecard section below exists to shrink that discovery time to one quarter.
- WATCH OUT
The cheapest-looking option, a mid-level hire "who can grow into it", is usually the most expensive. Jobs 3, 4 and 6 need judgment that only comes from having shipped, and a growing hire practices on your category while competitors compound.
- NO BUDGET?
If the loaded number stops the conversation, the realistic alternatives are a fractional senior operator two days a week, an agency, or training an existing marketer who already knows your buyer. For the third path, the free SEO + AEO Academy is the training program, and it costs nothing but their time.
The interview that separates operators from tool-drivers
Resumes in this field are indistinguishable. Everyone grew traffic some percent, everyone knows Ahrefs, everyone now says AEO. The interview has to force specifics, and the pattern that works is simple: every question asks what they did, to which property, and what happened after. Vague answers to specific questions are the single most reliable negative signal in this hire.
- “Pick one property you grew. What did you personally change, and what happened?” Good answers name the site, the constraint, the specific moves and the numbers, unprompted. Then verify: the property should be real, and the timeline should match their tenure on LinkedIn.
- “Our category's buyers ask ChatGPT for recommendations and we are not named. Walk me through your first two weeks.” You want: run the prompts across engines, record who wins and which sources get cited, check AI-crawler access, then prioritize. You do not want: “publish more AI-optimized content.”
- “What would you refuse to do in your first quarter?” Operators have a parked list; they know sequencing is physics. Candidates who promise everything in 90 days have not shipped a real program.
- “Sessions are down 20%, rankings are flat, and the CEO wants an explanation Friday.” This one tests whether they understand 2026: AI Overviews intercepting clicks, the measurement shift from sessions to citations, and how to say that to a CEO without hiding behind jargon.
- “Tell me about a piece of SEO advice you gave that turned out to be wrong.” The field changed under everyone. A candidate with no revised opinions since 2023 has not been paying attention.
[ RUN THE WORK SAMPLE, EXACTLY ]
- For your two or three finalists only: send five real buying prompts from your category and ask for a 45-minute readout one week later. Pay for it ($500 to $1,000). Paid samples get senior candidates; free homework gets whoever is unemployed.
- The brief: run the five prompts through ChatGPT, Perplexity and Gemini, record whether we are named, who wins, and which sources each engine cites. Then tell us the three things you would do first, and the one thing you would explicitly not do yet.
- Score the readout on prioritization and honesty, not polish. Did they notice the dead citations? Did they check whether your site blocks AI crawlers? Did they say “I don't know” anywhere? The best candidates always do, at least once.
- Compare their findings against your own baseline run of the same five prompts, which you did the week before. If you have not, the method is in how to measure AEO and takes an afternoon.
YOU DID IT RIGHT IF
You watched each finalist reason about your actual category with real engine output on the screen, you paid them for it, and at least one of them told you something about your AI visibility you did not know.
Red flags that survive a good resume
- Every win is a rankings win. If the whole track record is positions and sessions, jobs 3 through 6 are unstaffed. Ask about citations and watch for improvisation.
- Tool-driver syndrome. They can operate every platform and cannot tell you what they would do without one. The stack measures the work; it is not the work.
- Volume as strategy. “We published 60 posts a quarter” with no revenue framing. Engines are actively tuned against that flood now, and so is your budget.
- No opinion on AI search. Or worse, a confident wrong one, like “llms.txt is the priority” or “buy brand mentions.” The second one costs you real money; here is why.
- Solo-artist history. Never briefed another team, never got product or support to change a page. Job 6 is where B2B SaaS programs scale or stall.
The 90-day scorecard
Agree on this before the start date, in writing. It protects both sides: the hire knows what winning looks like, and you find out in one quarter, not three, whether the hire is working. Score shipped artifacts, never activity.
- Day 30. A crawl-verified technical fix list ranked by revenue impact, with the blocking items already fixed or ticketed with owners. AI-crawler access confirmed in the logs, not assumed.
- Day 45. A baseline: your citation share across the four engines on a frozen prompt set, plus the cited-source map. This is the number every future report moves against.
- Day 60. A revenue-ranked question list (ten questions tied to deals, not volume) and the first two pages shipped against it.
- Day 90. The first monthly report in the format they will run forever: what moved, why they believe it moved, what did not, and next month's three bets. If you cannot understand it, that is a finding about the hire, not about you.
- PRO TIP
The single best predictor at day 90 is whether the off-page work started by week three. It pays off last, which is exactly why weak hires postpone it and strong ones start it immediately.
- WATCH OUT
A beautiful strategy deck at day 30 with nothing shipped is the classic failure pattern. Decks feel like progress and compound into nothing. Artifacts or it did not happen.
When the right answer is not this hire
Run the math honestly. One person, however good, covers perhaps four of the six jobs at real depth, and the two that slip are usually AEO/GEO and off-page, the two that decide AI visibility. Below roughly $5M ARR, a full-time Head of SEO rarely makes sense: there is not enough surface area to keep six jobs busy, and the money buys more as an agency engagement or a fractional operator plus the $200-a-month stack. The full comparison, including when in-house genuinely wins, is on our in-house page.
And yes, the other path is us. Everything in this guide, the six jobs, the baseline, the scorecard artifacts, is what an Enginekick engagement ships without the six-month search, the loaded salary or the single point of failure, on the system we proved by making Swydo the most-cited brand in its category with zero outreach. Some companies should hire the head; this guide is genuinely for them. The rest should not pay $150K a year to find out.
Frequently asked
What does a Head of SEO do at a B2B SaaS company?
Six jobs: technical SEO including AI-crawler access, content operations decided against revenue rather than volume, AEO and GEO (getting the brand cited by ChatGPT, Perplexity, Gemini and AI Overviews), off-page citation earning, measurement built on citation share rather than rankings alone, and orchestrating the other teams whose surfaces AI reads, like docs, reviews and partner listings.
How much does a Head of SEO cost for a SaaS company?
US bands as of mid-2026: a senior IC runs roughly $120K to $160K base and a true head with management experience $150K to $200K+, before loading 25 to 35% for benefits and overhead plus $10K to $20K of tooling. The realistic loaded floor is about $150K a year, and the search itself typically takes around six months.
What should I ask an SEO leader in an interview?
Force specifics on shipped work: one property they grew, what they personally changed, and what happened after, then verify the timeline. Then test 2026 readiness: how they would diagnose being absent from ChatGPT answers, what they would report when sessions fall while AI mentions rise, and what they would refuse to do in the first quarter. Vague answers to specific questions are the strongest negative signal.
Should I give SEO candidates a take-home test?
Yes, but paid, real and only for finalists. Send five real buying prompts from your category, have them audit how the major engines answer and which sources get cited, and score the 45-minute readout on prioritization and honesty rather than polish. Paying $500 to $1,000 keeps senior candidates in the process and gets you a genuine audit of your AI visibility either way.
When should a SaaS company hire its first SEO?
A useful threshold is roughly $5M ARR, or earlier if organic is already a proven channel or competitors are visibly winning AI answers in your category. Below that, one full-time head rarely has enough surface area, and the budget usually produces more through an agency or a fractional operator plus a lean tool stack, with the option to hire in-house once the program has scale.
Is it better to hire in-house SEO or an agency for B2B SaaS?
It depends on scale and coverage. In-house wins when you can staff the full function and feed it protected capacity; a single hire covers perhaps four of six jobs, and the two that slip, AEO/GEO and off-page, are the ones deciding AI visibility. An operator-led agency covers all six from week one for less than one loaded headcount. The honest tiebreaker is whether you can keep the calendar protected, not the org chart.
