02 / Establish the local truth
Baseline Before Changes: Know What Moved
This check helps a business establish a dated baseline so later changes in visibility, engagement, leads, and booked work can be compared without rewriting history.

Record before you repair. Without a before-state, improvement stories become memory and correlation becomes attribution.
Scenario
Illustrative scenario · not a client result
A team updates categories, hours, location pages, review requests, internal links, and conversion tracking in one week. Ninety days later calls are up. Nobody can reconstruct which pages changed, which profile metrics improved, what seasonality looked like, or whether paid media also changed.
Evidence Lab
Collect the minimum evidence needed to answer this check. Keep the date, location, platform, route, role and source beside the observation.
- 01
Current profile information and verification state.
- 02
Representative local visibility samples using the market definition from Check 01.
- 03
Business Profile metrics actually available for this business, such as call-button clicks, website clicks, directions or bookings. Record the platform definition and whether paid activity is included.
- 04
Website organic/local landing-page traffic and conversions where tracking exists.
- 05
Calls, forms, bookings, qualified leads, and booked work from the business system of record where available.
- 06
Current review count/rating/recency distribution as factual observations, not magic thresholds.
- 07
Known concurrent campaigns, seasonality, pricing or operational changes that could affect results.
A manual search, profile screenshot, rank-grid point, website event and booked job are different kinds of evidence. Keep their definitions separate.
Principle
Without a before-state, improvement stories become memory and correlation becomes attribution.
The job is to help customers discover, understand, choose and contact the real business, then verify what happened. This field test is workflow guidance; it is not a claim that each action is a ranking factor.
Rebuild
- Choose a baseline date and preserve the raw observations.
- Record source, date range, location, query sample and metric definition beside each number.
- Separate Google profile interactions from website analytics and downstream CRM/job outcomes.
- Record tracking gaps explicitly instead of backfilling estimates.
- Define the review cadence before making changes: weekly operational checks, monthly comparison, quarterly strategy review.
Before editing a Business Profile, confirm the current platform requirements in the source trail. Keep the proposed change within the business’s actual operation and your account authority.
Field Test
The test
Create a one-page baseline with at least one visibility observation, one profile-engagement metric, one website/conversion metric, and one business-outcome metric—or explicitly mark the missing layer as unverified.
Record the result honestly
Verified pass · Verified fail · Partial · Not tested · Blocked · Not applicable
A pass needs reproducible evidence. If access, a definition or permission is missing, name the gap and the next person who can resolve it. Missing evidence never becomes a pass by default.
Use by role
- Owner / operator
- Supplies booked-work and lead-quality truth.
- SEO / analyst
- Captures repeatable visibility and platform baselines.
- Marketing / content
- Logs major campaign/content changes that can confound attribution.
- Engineer / analytics
- Documents event definitions and tracking gaps.
Checklist
Use these as reading markers. They stay in this page session and are not submitted, saved as an audit or sent to analytics.
Ask Your AI
Give the agent your expected state, relevant business facts and safe evidence. Keep secrets and private customer data out. Start read-only; account and production changes require their own authorization.
Read the full prompt
You are conducting a bounded local-search investigation for CHECK 02: BASELINE BEFORE CHANGES.
PAGE / BUSINESS JOB
This check helps a business establish a dated baseline so later changes in visibility, engagement, leads, and booked work can be compared without rewriting history.
EXPECTED STATE
State what should be true for this business, location, market and platform before proposing changes.
EVIDENCE
Collect current, reproducible evidence. Tie it to date, location, query/service, platform/surface, page/profile, and source. Separate platform policy, observed behavior, inference, hypothesis and unknown.
SPECIFIC CHECK
Build a dated local-search baseline before recommending changes. Separate visibility samples, Business Profile interactions, website analytics, calls/forms/bookings, qualified leads, reviews, and downstream business outcomes. Record source, date range, definitions, missing data, and concurrent changes. Do not infer attribution from correlation.
MORTEN A. GIRAFFE WORKING RULES
- Evidence before advice.
- The map is not the market.
- A local ranking observation is not a universal rank.
- A profile field is not permission to keyword-stuff.
- A review is customer evidence, not copy the business controls.
- A location is real because the operation is real.
- A local page earns its URL by serving a distinct audience, place, question, proof set or next action.
- Visibility is diagnostic; qualified demand and business outcomes determine whether the work mattered.
- Correlation is not clean attribution.
- Do not invent proof, rankings, review counts, customer data, testimonials, revenue, eligibility or platform behavior.
SAFETY / AUTHORITY
Begin read-only unless the user explicitly authorizes changes. Do not alter Business Profiles, listings, domains, analytics, production pages, review requests, customer records or tracking without the required account authority and a rollback/verification plan.
OUTPUT
For each finding report:
ID · market/location · platform/route · severity · confidence · classification · expected vs actual · evidence · root cause or labeled hypothesis · smallest recommendation · acceptance/retest · remaining unknowns.
Finish with the three highest-value next actions, not a generic SEO task list.
Reference:
https://www.mortenagiraffe.com/journal/local-search-field-guideVerify the result
- Did the evidence come from the intended business, location and market?
- Is every current platform claim tied to an official source?
- Are the recommendation, observed facts and untested hypothesis distinguishable?
- Are profile and website facts still consistent?
- Did the correction introduce an eligibility, review-policy, accessibility, routing or measurement problem?
- What observation would disprove the recommendation?
Questions
Is this check a ranking factor?
Not every useful practice is a ranking factor. This check addresses the business baseline, representation or geographic evidence. Use the cited platform documentation for platform claims and the field test for practical verification.
Does passing guarantee visibility?
No. Local search is competitive and contextual. Google describes relevance, distance and how well-known the business is; a checklist cannot guarantee a position. Google’s local-ranking guidance explains that framework.
What if part of the test is unavailable?
Record Not tested, Blocked or Partial and state the exact access, evidence or authorization needed. Continue independent checks that can be completed safely.
Source trail
Platform sources checked . Recheck policy before acting. Illustrations and scenarios teach the method; they do not show measured client outcomes.
- Google: improve your local rankingFirst-party platform documentation
- Google: understand Business Profile performanceFirst-party platform documentation