New clinic, 3 weeks live — 12 GSC impressions, 0 patients from Google

Week three: the screenshot that changed the internal meeting

The clinic owner sent a Google Search Console screenshot to the internal WhatsApp group: 12 impressions, 0 clicks, in 21 days since launch.

It was not because the area was quiet. On Google Maps, two competitors within 2 km already showed for the same queries — with reviews, waiting room photos, and pages that answered “Saturday opening hours” in the first paragraph. Their new patients came through neighbor recommendations and housing complex groups. Digital channel? Nearly zero.

The frustrating part: the website was not bad. A paid vendor had delivered proper design — treatment room photos, service pages, contact, WhatsApp button. From the outside, everything looked professional. From the search engine’s perspective, this clinic barely existed.

Two battlefields, both empty

Local patients now research in two places before booking: Google Search / Maps and generative answers (ChatGPT, Perplexity, Gemini). The clinic team tried a prompt: “Recommend a [service X] clinic in [city Y]”. Their name never appeared. Competitors with tidy GBP profiles and specific service pages? Named every time.

Without consistent entity data and pages that answer directly, this brand was absent in both fields at once. Not one — both.

What they had already tried (and why it was not enough)

The website vendor delivered “looks good.” Admin set up GBP alone — uploaded photos, picked a category, done. Instagram posted occasionally. SEO? An acquaintance said: “Wait 6–12 months, it will rise on its own.”

The problem was not lack of patience. The problem was no foundation that could be indexed correctly: conflicting business data, pages that did not answer specific questions, and zero measurement for the AI channel already sending traffic elsewhere.

Audit findings: small things that make Google hesitate

Once we audited, the pattern was familiar for SMBs going digital:

Inconsistent NAP. GBP hours differed 30 minutes from the website footer. Address written “Jl.” in one place, “Jalan” in another. For search engines, that is a weak identity signal. For patients comparing Maps vs website, it is small confusion that erodes trust.

Service pages = brochures. Local patients ask practical questions: open until what time, is service X available, rough pricing, online booking or phone only. Old pages focused on marketing copy — “we care about your health” — without extractable answers.

GA4 did not separate AI-referral. Without a segment for chatgpt.com and perplexity.ai, the owner would never know if signals were coming from the new field — even while organic was still zero.

The controversial internal decision: fix foundation first, do not produce 10 blog articles. The owner initially wanted “content so there is something there.” Data won: without NAP and answer pages, new content had no foundation either.

What we ran (Oura Starter + light Atlas)

Starter at Oura does not mean no system. Execution order:

1. Align NAP across website, GBP, and relevant local directories.
2. Schema LocalBusiness + Organization + FAQPage + sameAs to active social profiles.
3. Three core service pages rewritten answer-first: question in heading, answer in first paragraph.
4. GA4 with AI-referral segment from day one.
5. Light Atlas loop: GSC + GA4 → monthly brief → dry-run → verification. Owner reads the report, does not need to understand SEO.

Every production change went through dry-run and backup — important for a clinic selling trust to patients.

How to read the numbers below

Patterns we see in similar Starter engagements:

• GSC impressions start moving after NAP + answer pages stabilize (weeks 4–8).
• Maps becomes relevant for specific local queries after GBP alignment (months 2–3).
• AI-referral signals usually appear after structured entity + FAQ — separate from organic in reports.

Once foundation is right, the path to Oura Growth is clear: scaled answer-first content, OAVF, and data-driven iteration — not guessing.