06 — Your first 90-day publishing cycle

Part 6 of 6 · ← Regulated content · Index

This page is the concrete plan. 90 days. One named owner. Three category investments. Four measurement targets. Skip the "build a content engine" framing — that's a year of work; you don't have a year before LLM retrieval matters.

Day 0: pick the lever

Run the observation tool against your business. Pick one lever where:

• The carriers winning today are smaller than you (so you have a credible path to displace).
• The lever has at least one category-saturated comparison-site editorial surface (so the citation graph exists).
• Your existing owned-domain content is closer to the winning shape than your peers'.

Examples of viable starting levers for different carrier shapes:

• Mid-market home carrier with luxury or coastal book: property-type ownership for "luxury home" or "coastal home" or specific-state pages where you have rate-filing approval and the incumbent winner is weak.
• Auto carrier with specialty book (rideshare, classic, EV, fleet): emerging-category presence on vehicle types where the insurtechs lost the editorial race.
• Regional mutual or farm bureau: regional-monopoly defense in your dominant state. Even if you're already cited, defend by publishing your own line-of-business pages before a national entrant tries to displace you.
• Commercial-cyber carrier or MGA: vertical-specific cyber pages (fintech, healthcare-SMB, manufacturer) where the comparison-site coverage is still light.

If you don't see a clear pick, you're early — work a baseline audit (next section) for two more weeks before committing.

Days 1-14: baseline audit

Do these in two weeks. One person, half-time.

Citation baseline. Run 10 buyer-shaped queries against ChatGPT, Claude, Gemini, Perplexity. Record: - Whether your carrier appears. - The first-named carrier (which is the highest-leverage slot). - The URLs cited as sources. - How the framing of your carrier compares to the winner's framing.

This is a Google Sheets task. 4 LLMs × 10 queries × 5 runs each = 200 prompts; budget a day.

Comparison-site placement audit. For each of the top 5 comparison sites in your category (NerdWallet, Bankrate, The Zebra, Policygenius, Insurify for personal lines; equivalent set for commercial), record: - Whether you're listed. - Where you rank in the editor's "best of" shortlist. - Which editorial sub-pages mention you (state-specific, demographic-specific, specialty-specific).

This is a manual audit. Half a day.

Owned-domain inventory. List your top 30 indexed pages by traffic. For each: - What buyer query does it answer? (If unclear, mark "doesn't ladder.") - Is it geo-tagged or specialty-tagged? Y/N. - Is the page shape comparison-friendly or prose-narrative? (See page shapes that win). - Is there a regulatory-disclosure issue? (See regulated content.)

This is a one-day spreadsheet exercise.

End of week 2: you have a 30-page inventory plus the citation baseline. You can see where you're cited and where you're invisible.

Days 15-30: pilot one lever

Pick one query shape. Write 3-5 owned-domain pages on it.

Example: if the lever is "luxury-home insurance editorial," write: - 1 hub page: "Luxury home insurance — what's covered, what's different." - 3 state-specific pages: "Luxury home insurance in Washington / Florida / California." - 1 narrowly-specialized page: "Luxury home insurance for waterfront properties" or "for historic homes" or whatever sub-niche your book actually has.

Each page follows the page shapes that win template: - Comparison table or named-entity list above the fold. - One buyer query in the H1. - Geo-tagged where applicable. - Disclosure-first ordering on coverage statements. - <1,500 words.

Don't over-build. The pilot tests whether your team can produce in this shape and whether the first 5 pages start moving citation share within 4-6 weeks. If you can't ship 5 pages in 15 days at this depth, your real bottleneck is process — fix that first.

Days 31-60: comparison-site outreach

Comparison sites set their editorial calendars quarterly. Your placement on NerdWallet "Best home insurance for luxury homes" updates roughly every 90 days. The window to influence Q3 is May-June.

For each top-5 comparison site, do this in May:

Identify the editor. Most comparison sites publish editorial bylines. The editor for the "Best home insurance" category, the "Best car insurance for X" category, etc., is reachable.

Send a category-specific brief. One page. What you offer in this category. Quantitative claims (premium ranges, J.D. Power scores, AM Best rating, complaint ratio). Source links to your owned-domain pages from the pilot. State availability table. Quote from your underwriting lead.

Don't pitch generic awareness. "Acme Insurance has been protecting families since 1922" is not editorially useful. "Acme covers luxury homes in 47 states with specific-rider coverage for [X], which most national carriers don't" is editorially useful.

Track what gets cited next quarter. Comparison sites refresh in cycles. If your brief lands in May, your citation appears in the September refresh, and your LLM citation share starts moving 30-60 days after that.

Days 61-90: measure, re-cut, plan v2

End of quarter:

Re-run the citation baseline. Same 10 queries. Did your carrier appear in more answers? Did the first-named slot move? Did the cited URLs shift toward your pages?

Audit the new pages. Are they getting search traffic? Are they being cited by any third-party site? Are they being LLM-cited?

Decide v2. Three scenarios:

1. The pilot moved citation share. Double down. Expand to 8-12 more pages on the same lever. Pick the next state, the next sub-niche, the next product variant.
2. Comparison-site placement moved but LLM citation didn't yet. Wait one more quarter. Comparison-site → LLM-citation lag is typically 60-120 days.
3. Nothing moved. The diagnostic question is whether the lever was wrong, the page shapes were wrong, or the comparison-site outreach didn't take. Audit each. Don't ditch the program; ditch one of the three components.

What to measure

Four numbers, every two weeks:

• First-named-carrier share on your top 10 queries. The most important number.
• All-named-carrier share: how often your carrier appears anywhere in the answer. Lagging indicator of brand awareness through the LLM channel.
• Citation-URL share: of the URLs LLMs cite when they answer your top queries, how many are yours? (Target: 1-2 of typical 4-6 cited URLs.)
• LLM-referrer inbound traffic: ChatGPT, Perplexity, Claude, Google AI Overviews appearing as referrer in your analytics. Volume is small; trend is what matters.

If three of four are flat at month 6, the lever is wrong. Pick a different one.

Common failure modes

• Trying to win all five levers at once. Most-common failure. Pick one.
• Spending the first month on infrastructure. Schema markup, llms.txt, RSS feeds, syndication automation — all useful, none of it ships pages. Build infrastructure in parallel with content; don't let it be the prerequisite.
• Outsourcing the editorial calendar. Agencies will produce content. They will not pick the right lever, or notice when comparison-site placement breaks, or rebut state-DOI complaints. Keep the strategic call inside.
• Measuring at week 4. Citation share moves on quarterly cycles. Measuring at week 4 will tell you nothing useful. Bake the patience into the plan.
• Treating GEO as an SEO sub-task. Some teams do; the metrics are different and the incentives diverge. Either give GEO its own KPI track or accept that it'll get optimised against SEO metrics it isn't actually built for.

The honest framing for executives: GEO is a 12-month program where the first 90 days is the table-stakes pilot. Month 1-3 is "do you have the editorial discipline to ship pages in the right shape." Month 3-6 is "does the comparison-site channel respond." Month 6-12 is "does LLM citation share follow." Carriers that try to compress this into 30 days are usually carriers that abandon GEO at month 4 because nothing moved.