Video summary

The 30-Page Local SEO Hack Google’s AI Is Already Rewarding

Main summary

Key takeaways

Business

Who’s speaking (business context)

  • Caleb Ulku runs/teaches local & AI SEO.
  • He describes a process that treats a business’s website as a structured trust layer for Google Business Profile (GBP) and AI agents.

Core thesis: Local SEO that mirrors GBP + “AI-readable” trust

Caleb argues that as Google/LLMs increasingly summarize and recommend businesses (agentic search), local SEO should:

  • Match website structure exactly to GBP categories/services so AI systems can interpret clean “signals.”
  • Create informationally additive content (not keyword stuffing; not thin doorway pages).
  • Optimize not just for ranking, but for AI recommendations and call conversion.

Playbooks / frameworks / process (as described)

1) The “Core 30” website framework (GBP-mirroring)

Build a website where:

  • Primary GBP category drives the top-level page (often homepage / GBP landing page)

  • Secondary GBP categories become major sections (e.g., H2s)

  • Core services (the most important ones) get dedicated sections.

Typical local business output: ~30 pages (“Core 30”).

For each GBP secondary category, create internal links down to specific service pages.

2) Rank Map → Topical Relevance → Geographic Relevance sequence

Caleb describes staged local SEO:

A) Rank map sizing (correct territory/coverage first)

  • Use “percent top 3” as the main metric.
  • Goal: top competitors should fall roughly in 60–70%, up to 80–90% in top-3 rankings depending on competitiveness.
  • If your target business is too low due to map size:
    • If % top 3 is much below ~60% → rank map is too big → rerun with a smaller map.
    • If leaders are 100% top 3 → map is too small → rerun larger.
  • Tools: often uses LeadSnap for rank maps.

B) Topical relevance threshold

  • Confirm Google believes you match the service entity.
  • If the client’s top-3% is ~40% (or around half of market leaders), he considers topical relevance “met.”
  • Core 30 is step one; if still not enough:
    • Add supporting content tied to the same service entity
    • Avoid unrelated “informational posts.”

C) Geographic relevance after topical relevance

  • Build hyperlocal content primarily where you rank just outside top 3 (positions 4–6).

  • Use landmark-based geo pages (e.g., “near River Oaks / near Montro / near 610 West & I-10”).

  • Scale tactics:

    • Use heavy research sources (e.g., US Census Bureau, Reddit, CRM conversation data) to make pages meaningfully different, not redundant.
  • Uses Google Places API to ensure landmarks are recognized by Google.

3) “Attribute matching” for AI agents (newer concept)

AI agents may prefer exact attributes/terms.

  • Example: if users ask for “lead main drain line replacement”, AI may not infer equivalence from just “main drain line replacement.”

  • Recommendation:

    • Include the same service entity wording in key places (title tag/H1/first paragraph)

    • Add trust details (reviews, responsiveness, outcomes).

4) Internal linking hub model (for scale)

If you have many supporting content pages:

  • Create a supporting-content hub
  • Link from the GBP landing page (possibly footer) to the hub
  • Hub links out to supporting pages
  • Supporting pages link back to the service page they support

Goal: avoid too many direct links from the GBP landing page.

Concrete local SEO implementation example: Plumber site structure

Homepage / GBP landing page

  • Based on GBP primary category + location (e.g., “plumber Houston”)

Add sections for:

  • Secondary categories (H2 per secondary category)
  • Core services (H2 per core service)
  • Short supporting paragraphs under each (described as ~50–70 words)

Service page examples

  • “bathroom remodeling Houston”
  • “water heater replacement Houston”
  • “faucet replacement Houston”

GBP hierarchy mirroring:

  • The bathroom remodeling page links to specific services under that category.

Multi-location tactics (GBP landing pages + rank safety)

  • A “GBP landing page” might not be the homepage depending on performance/risk.
  • Do not change GBP landing pages if rankings are already good:
    • If the homepage is currently powering rankings, moving GBP to internal pages may cause rankings to “plummet.”

  • Approach:

    1. Run a rank map first.

    2. If rank is horrible or sites are new:

      • Give each GBP location its own GBP landing page (copy of homepage structure)

    3. If authority accumulates across locations:

      • Later locations may need less Core 30 expansion
      • Focus shifts toward more geo relevance content.

Scaling note:

  • Core pages scale linearly: 5 locations → Core 150 (Core 30 × 5).

Biggest GBP mistakes (per Caleb)

  1. Categories/services not optimized

    • Google allows up to 10 categories; many businesses underuse this.
    • Myth Caleb disputes: adding secondary categories “dilutes” primary ranking—he says testing shows positives.

  2. Service entries issues

    • Too few services
    • “Keyword stuffing” in services (he claims entity-based understanding makes it ineffective/harmful)

  3. Unfilled GBP fields

    • Example: holiday hours must be updated periodically (Google limits how far ahead you can enter them).

On-page keyword placement priorities (for local pages)

Least emphasis:

  • URL structure vs. on-page elements.

Most important:

  • Title tag
  • H1
  • First paragraph
    • Put the keyword early
    • Avoid “history lesson” content; satisfy the searcher’s intent immediately.

URL slugs:

  • Doesn’t care much about parent folders/subfolders.
  • Prefers keywords in slugs mainly to support anchoring/linking without spammy tactics.

SEO mistakes he says agencies make (esp. blog/informational SEO)

  • Local SEO agencies waste time on blogs targeting top-of-funnel informational keywords that don’t reinforce topical relevance.
  • Example:
    • “Top 5 ways to winterize your water heater” isn’t tightly aligned with a plumber’s core service entity.

Case study (Lasik client)

  • Downtown Chicago Lasik client:
    • ~80,000 hits/month
    • But rank map looked bad because traffic came from irrelevant informational queries (e.g., eye health food/exercises).
  • After rewriting homepage for Chicago and emphasizing neighborhood geographical relevance:
    • Monthly traffic dropped to ~30–40,000
    • But rank map turned green across Chicago and calls increased (traffic quantity mattered less than local call performance).

Link building + local “virality/trust” tactics

Still-working local backlink methods

  • Chamber of Commerce memberships
    • Depends on competitiveness
    • Example: personal injury lawyer in New Orleans joined 17 chambers within 80 miles
  • Sponsorships
    • Often “couple hundred bucks” for links from local orgs
    • Example: sponsor a festival/event (e.g., “Bayou Bash”) for a strong local link

Finding opportunities with AI

  • He uses Gemini prompts to find local sponsorship sites.
  • Example: an Austin house cleaner got a link from UT Austin (TEDx sponsor link) with minimum sponsor around $250, and claims ranking improved 3–4 positions overnight.

AI content production: videos + embeddings

  • To align with Gemini/video-native strengths:
    • Produce a YouTube video for ~1/3 to 1/2 of published local article topics.
  • If the business won’t film:
    • Generate videos with Pictory using short scripts (2–3 minutes).
  • Videos aren’t aimed at viewers directly:
    • The purpose is to feed Gemini and then embed the YouTube video on the corresponding article.

AI changes to operations/costs (how it affected the business)

  • Production costs: ~80% lower vs mid-2022
  • Output quality increased:
    • More research time via automation
    • Stronger outlines
    • Better assembly at scale
  • Future differentiation expectation:
    • Not only content types (category/service/geo/topical relevance)
    • Add a 5th type: “Trust”
    • Trust content includes:
      • Pricing info
      • “Things that went wrong” stories
      • Service outcome attributes (timeliness, what happened, resolution)

Trust + reviews strategy (AI-facing)

Reviews strategy changes

  • Avoid heavy review-management software because:
    • action/conversion rates are low (he cites ~5%).
  • Main tactic:
    • Text customers to request reviews using the phone number they already use
    • Claimed review take rate: ~40–45% (about 10x higher)

After “Map/Ask Maps” emergence:

  • Rotate review sources beyond GBP:
    • GBP, Bing for Business, Yelp, Angi, ChatGPT/Bing-driven references, etc.

Review content ask:

  • Instead of “leave a review,” prompt for outcome/attributes:
    • e.g., “talk about what happened and the outcome”
  • He notes Google won’t allow “tell us specific outcomes” explicitly, so he stays compliant with wording that elicits attribute detail.

Fake reviews and gating (risks)

  • Fake reviews are against the law and can lead to deindexing/scrubbing.
  • Avoid review gating (also against Google terms; if caught, reviews can be scrubbed).
  • Conversion insight:
    • Ratings around 47–49 may outperform a perfect 50
    • He suggests adjusting behavior to avoid an “too perfect” rating to improve conversions.

Call-handling as part of “SEO” (operations + KPI)

Caleb treats responsiveness/phone behavior as a ranking factor because Google tracks whether users complete the “goal” (calling).

  • Example behavior impact:
    • If a plumber doesn’t answer for ~2 weeks, rank map worsens after ~2 weeks.
  • Operational requirement in contracts:
    • Minimum pickup rate to avoid missed calls.
  • If AI phone agents are used:
    • Monitor recordings/data:
      • if callers hang up and call competitors, it can harm ranking.

Agency growth tactics (business operations)

Scaling by revenue band

  • Six figures (often ~5–6 clients)
    • Often manageable as a one-person show
    • Growth driven by consistent prospecting + improving phone closing
  • $83k/month for seven figures
    • Example math: ~42 clients if $2k average per client
    • Requires learning to build/train and avoid being the bottleneck
  • Beyond seven figures
    • Needs stronger team management systems (you can’t personally handle all clients)

Client acquisition playbook (local)

  • Start a YouTube channel (he claims it’s the best source once it works)

  • Attend every Chamber of Commerce meeting

    • Don’t pitch; build relationships
    • Can share rank-map findings and quick on-page fixes (title tags, H1)
  • Run ads (Meta)
    • Can work but requires:
      • offer testing
      • split testing
      • significant budget + sales capability
    • He cites spending $5–10k before an offer works
    • Operational mismatch note:
      • ad leads often require better sales-call coaching

Business KPIs and targets explicitly mentioned

  • Rank map KPI: % of locations/pages in “top 3”
    • Benchmark range: ~60–70% up to 80–90% depending on space
    • Topical relevance threshold example: around ~40% top three
  • Traffic vs ranking vs calls (case study):
    • From ~80,000 monthly hits → down to ~30–40,000
    • But rank map green and calls increased
  • Production cost KPI:
    • content production cost: ~80% lower vs mid-2022
  • Reviews take rate:
    • 40–45% via SMS review request vs ~5% with typical review management software
  • Agency math:
    • six figures: about 5–6 clients
    • seven figures: about 42 clients at $2k average revenue per client
  • Video production ratio:
    • create YouTube videos for ~33%–50% of local articles

Key actionable recommendations (condensed)

  • Build your site to mirror GBP categories/services (“Core 30”).
  • Use rank maps to set coverage and run in sequence:
    1. map sizing
    2. topical relevance
    3. then geographic relevance
  • Stop blog SEO that doesn’t reinforce service entities.
  • For geo content, ensure pages are informationally additive (real local differences; avoid near-duplicates).
  • Fix GBP categories/services first; ensure all fields (including holiday hours) are filled.
  • Improve call operations:
    • track pickup rate
    • ensure friendly, competent handling
  • Build local backlinks via chambers + sponsorships; use Gemini prompts to find sponsor opportunities.
  • Use reviews to supply AI-readable attributes; request via SMS for better take rates.
  • For AI/LLM recommendation readiness:
    • add a future-focused “trust” content type (pricing, failures/stories, responsiveness claims with supporting reviews).

Presenters / sources mentioned

  • Presenter (main guest): Caleb Ulku
  • Host: Edward (referred to as “Edward” throughout; no last name provided in subtitles)

Referenced/mentioned sources and entities

  • Google / Sundar Pichai interview (agentic search discussion)
  • ChatGPT-3.5 (timing around Nov 2022 / early GPT conversations)
  • Gemini (video-native model + maps discussion)
  • Claude (mentioned in prompt comparisons)
  • Exxon Mobile
  • Larry Page / Google search algorithm foundation (PhD dissertation reference)
  • US Census Bureau, Reddit
  • UT Austin TEDx (example sponsorship)
  • Ask Maps / Google News (press release strategy)
  • PR Underground and Press release ASAP (tools/services mentioned)
  • LeadSnap (rank map tool)
  • Pictory (AI video generation)
  • Google Places API (landmark recognition)

Original video