Serum case study · ChatGPT visibility

From absent to named in 4 of 9 buyer searches in one month.

An anonymized cosmetic clinic moved from 0 of 9 to 4 of 9 locked ChatGPT searches between July 2 and August 2, 2026.

The result in one sentence

The clinic appeared in four previously empty treatment and provider-fit lanes after targeted pages were published, but one month and one platform do not prove a permanent ranking or booked-patient impact.

The baseline

On July 2, Serum ran nine locked buyer searches in ChatGPT search mode. Each query named the clinic's market and a priority service or positioning angle.

The clinic was named in 0 of 9 answers.

What Serum completed before the next check

Three targeted articles were live before the August rerun. The content focused on thin, commercially relevant lanes where the clinic had a specific claim and the market was not already dominated by generic popularity lists.

One winning query cited a Serum-produced page directly. We do not claim that every movement was caused by the content because ChatGPT can change sources and wording between checks.

The month-one check

Anonymized buyer searchJuly 2August 2
Device-specific laser resurfacing in the clinic's cityNot namedNamed #2
Deep-plane facelift surgeon in the clinic's cityNot namedNamed #2
Ponytail facelift in the clinic's cityNot namedNamed #1
Plastic surgeon for Asian faces in the clinic's cityNot namedNamed #3
Five crowded, generic service or clinic searchesNot namedNot named

What the result suggests

The clinic moved first in lanes with a clear, specific claim. The crowded generic searches stayed at zero. That pattern matched the strategy, but it is still an observation rather than proof of a private ranking formula.

Specific claims moved first. Generic popularity lanes did not.

Source of every metric

  • 0 of 9 baseline: dated ChatGPT query archive from July 2, 2026
  • 4 of 9 month one: the same locked query set in fresh Temporary Chats on August 2, 2026
  • Work completed: published URLs and dates retained in the client workspace
  • No booking claim: the visibility logs do not prove consultations, patients, or revenue

Limitations

  • One month of movement
  • One platform
  • One reading of each locked query per monthly check
  • A method change to Temporary Chat at month one, disclosed in the log
  • No claim that content alone caused every change
  • No proof of booked patients or permanent placement

The next useful check is whether the same four lanes hold over time.

Two ways to start

Start with your own three buyer searches.

The free clinic snapshot gives you a point-in-time baseline before a larger program.