GEO 30-60-90 Day Plan for a New Hire

A 30-60-90 day plan tells a new GEO hire and their manager what success looks like at each stage. In days 1–30 the hire learns and measures. In days 31–60 they fix and publish. In days 61–90 they expand, re-measure and set up reporting. This template gives goals, deliverables and check-in questions for each phase, so progress is judged on agreed outputs, not impressions.

Template by MentionX, reviewed 31 Oct 2026 (CET). Example is illustrative. "Our view" marks opinion.

On this page

• Days 1–30: learn and measure
• Days 31–60: fix and publish
• Days 61–90: expand and report
• What does a filled example look like?
• How should the manager use the plan?
• Template checklist
• From MentionX data: what your baseline will probably show
• Worked example: Ledgerline's new GEO lead uses the plan
• What should happen in the first week?
• How do you choose what to fix first?
• How do you work with other teams?
• What should the day-90 report include?
• What if the baseline shows you're already doing well?
• What if the plan slips?
• How do you adapt the plan for an agency?
• How do you adapt it for a small team?
• How do you hand over at day 90?
• What tools do you need in the first 90 days?
• How do you keep stakeholders informed between check-ins?
• How do you know the plan worked?
• Common questions
• Related
• Sources

Checked on 31 October 2026. "Our view" marks MentionX opinion. Ledgerline is a fictional company used for illustration.

Days 1–30: learn and measure

Goal: understand the business and set a baseline.

For the technical check, confirm you aren't blocking the AI crawlers you want. OpenAI documents its crawlers and how they use robots.txt (OpenAI), and Perplexity documents PerplexityBot (Perplexity). See tracking AI traffic in GA4.

Day-30 check-in questions: Is the baseline agreed? What are the three biggest gaps? What's blocking progress?

Days 31–60: fix and publish

Goal: fix the highest-value problems and ship first content.

Day-60 check-in questions: What shipped? Which gaps are being worked on? Where do you need help from other teams?

Days 61–90: expand and report

Goal: show early movement and set up a regular rhythm.

Day-90 check-in questions: What moved, and how confident are we, given that answers vary between runs? What should we stop, start and continue?

Our view: expect early, partly noisy signals by day 90, not settled results. Agree this with the manager on day 1. See statistical confidence in tracking.

What does a filled example look like?

Fictional example (fictional company): a new GEO specialist at a B2B software company spends week 1 interviewing sales and support and collects 120 real buyer questions. By day 25 they have a 70-prompt core set and a baseline across four engines. The baseline shows two competitors leading on comparison prompts, and an AI engine describing a retired pricing tier. By day 60, robots.txt has been fixed to allow the search crawlers the team wants, the pricing page has been rewritten, and six comparison pages are live. By day 90 the re-measurement shows modest gains on comparison prompts, and the first executive dashboard has been sent.

How should the manager use the plan?

• Agree the deliverables in week 1 and write them down.
• Hold short check-ins at days 30, 60 and 90.
• Judge the hire on deliverables and learning, not on single-run visibility swings.
• Clear blockers in other teams, especially development time for technical fixes.

Template checklist

• Deliverables agreed with the manager in week 1
• Baseline method written down (prompts, engines, runs, dates)
• Owners named for technical fixes
• Check-in dates in calendars
• Day-90 report compares like with like

From MentionX data: what your baseline will probably show

Method. Read-only aggregate queries over MentionX scans from 22 August to 29 October 2026 (UTC): 2,217 engine runs (2,154 completed) across 18 anonymised projects and six engines.

Your first baseline will look uneven across engines and question types. Our sample shows the range you can expect:

Our view: use the first 30 days to decide which engines and question types to focus on. Comparison and brand questions name brands most often, so they're usually where early work shows up first. Engines that rarely cite are better tracked for mentions than citations.

Limits. A convenience sample of mostly B2B software projects, run through engine APIs, which can behave differently from the consumer apps. Perplexity, Copilot and DeepSeek figures come from a single day (26 October 2026), so treat them as a snapshot, not a trend.

Worked example: Ledgerline's new GEO lead uses the plan

Ledgerline is a fictional accounts-payable tool. Figures are illustrative.

Ledgerline hires its first GEO lead. In the first month she interviews sales and support, agrees four competitors (Competitor A to D) with the CMO, builds a prompt set and runs a baseline. She finds that one engine describes an old pricing plan and that Competitor A leads on comparison prompts.

In month two she gets the pricing page rewritten, asks a review site to correct an outdated profile and ships four comparison pages. In month three she re-runs the same prompts with the same method, sends the first one-page report and proposes next quarter's goals. The CMO sees steady, explained progress rather than a single headline number.

What should happen in the first week?

The first week is for listening. Meet the people who know what buyers ask and what goes wrong:

• Sales: the questions prospects ask, the competitors they mention and the objections that lose deals.
• Support: common confusions, wrong assumptions customers arrive with and outdated facts they've read.
• Product: what's changing soon, so you don't optimise pages that are about to change.
• SEO and content: what exists, what performs and how publishing works.
• Leadership: what success looks like to them and how they want updates.

Write short notes after each conversation. They become the raw material for your prompt set and your priorities.

How do you choose what to fix first?

Rank problems by how much they matter to buyers and how easy they are to fix. Wrong facts about your own product, such as old prices or retired features, come first: they mislead buyers and are often quick to correct on your site. Technical blockers that stop crawlers reaching key pages come next. Content gaps on high-intent prompts follow. Our view: a short list of three to five fixes, finished, beats a long list started.

How do you work with other teams?

You'll need help from developers, writers, PR and product. Ask early, with specific requests and a reason tied to a buyer problem. "The pricing page shows a plan we retired, and assistants repeat it" gets action faster than "we need to improve AI visibility". Track requests in the same tool the other team uses, not in a separate sheet they never see.

What should the day-90 report include?

Keep it to one page:

• Baseline and current values for the main measures, with the method and dates.
• What shipped: fixes, pages and outreach.
• What moved, what didn't and your honest read on why.
• The plan and budget request for next quarter.

Add a note on confidence. Answers vary between runs, so say how many runs you did and whether the change is larger than your normal range.

What if the baseline shows you're already doing well?

Then your job shifts from fixing to protecting and extending. Check accuracy regularly, extend the prompt set into new topics and markets, and watch competitors for gains. A strong baseline also makes a good case for keeping the investment steady.

What if the plan slips?

Plans slip, often because other teams are busy. At each check-in, say what slipped and why, and agree what to drop or delay. Don't quietly shrink the scope. If one blocker, such as developer time, keeps recurring, raise it with your manager with a specific ask.

How do you adapt the plan for an agency?

Agencies can use the same structure for a new client. Days 1–30 become onboarding and audit, days 31–60 the first delivery sprint, and days 61–90 the first full report and renewal discussion. Agree the prompt set and competitor list with the client in writing, because changing them later makes comparisons hard.

How do you adapt it for a small team?

If you're the only marketer, cut scope rather than stretching dates. Use a smaller prompt set, focus on one or two engines and fix the top three problems. The rhythm of measure, fix and re-measure matters more than the size of each step.

How do you hand over at day 90?

Write down how the work is done: where the prompt set lives, how you run it, how you tag results and who to ask for fixes. This short playbook means someone else can cover holidays or take over later, and it helps you spot steps worth automating. Store it next to the data, not in a personal folder.

What tools do you need in the first 90 days?

Very few. A spreadsheet for the prompt set and results, access to the engines your buyers use, analytics access for AI referral traffic, and read access to your CMS and robots.txt. A tracking tool helps once you know what you want to track, but buying one on day 1 often locks you into the vendor's prompt list rather than your own. Our view: run the first baseline by hand or with a simple script, then choose a tool in month two or three with a clear list of requirements.

How do you keep stakeholders informed between check-ins?

Send a short note every two weeks: what you did, what you found and what you need. Three bullets each is enough. People who hear from you regularly are far more likely to help when you ask, and nothing in the day-90 report comes as a surprise.

How do you know the plan worked?

By day 90 you should be able to answer yes to four questions. Is there an agreed baseline with a written method? Have the most damaging wrong facts been fixed? Are the first priority pages live? Is there a plan and budget for next quarter that leadership has seen? If all four are yes, the first 90 days did their job, whatever the visibility numbers say. If any is no, make it the first item of next quarter's plan. Agree these four questions with your manager on day 1, so you both judge the plan by the same standard. Write each answer down and keep the notes with the plan so later reviews can refer back to them.

Common questions

Can I use this plan if I'm not new? Yes. It works for restarting GEO work on any team.

How many prompts should the core set have? Enough to cover your main topics and question types; many teams start with a few dozen.

What if the manager wants results by day 30? Agree on day 1 that day 30 is for the baseline, not results.

Related

See hiring your first GEO specialist, the GEO strategy deck template and the 90-day GEO plan for SaaS.

Sources

• Google, Custom channel groups in GA4
• OpenAI, Overview of OpenAI crawlers
• Perplexity, Perplexity crawlers