Appear client case study
Copeland appeared in 2.4× as many monitored Gemini answers.
Copeland Home Services · HVAC & home services · Forney, Texas
On the same 15 local HVAC questions, Appear recorded Copeland in 12 Gemini answers, up from five. The questions described a service need without supplying the company's name.
Client results reported by Appear
Read the original studyGemini: June–August vs. September 2026. This is TRC's summary of Appear's client work, not a TRC engagement or an independently verified result.
Source reviewed . Results vary by question, platform and measurement period.
As many named Gemini answers
Named answers out of 15
Gemini appearance rate
Named answers out of 62, all models
The Challenge
Copeland provides heating, cooling, plumbing and electrical services. Its discovery challenge was appearing when someone asked for nearby help without already knowing the company's name.
Appear monitored local questions about emergency AC repair and combined HVAC and plumbing services. A website citation and a named appearance were counted as different observations.
Observed results by platform
Matched questions and recorded models, comparing the earliest pre-September answer with the latest September answer. Three models increased; OpenAI decreased.
Gemini 2.5 Flash Lite
Named answers
Perplexity Sonar
Named answers
Claude Sonnet 4.6
Named answers
OpenAI GPT-5.4
Named answers
Appear's approach
Appear compared retained answers for matched local-service questions rather than treating a single favorable answer as a permanent ranking.
Initial observations
Record a baseline
The Gemini baseline covered June 18–August 17, 2026: five of 15 matched answers named Copeland.
September 5–21, 2026
Repeat the same questions
The latest Gemini checks named Copeland in 12 of the same 15 question-and-model pairs.
Comparison
Keep the model differences visible
Across four models, named appearances rose from 18 to 28 of 62 matched answers. OpenAI's named appearances fell rather than rose.
Work described in the study
Reported results
Gemini's named-answer rate rose from 33% to 80% in the matched sample. Twelve divided by five gives 2.4 times as many named appearances.
Across all four models, Appear recorded 28 of 62 named appearances versus 18 of 62 initially: approximately 45% versus 29%, or a 56% relative increase.
The source also reports increased Perplexity source-list inclusion, but explicitly distinguishes that measure from named answers and notes that source lists changed in length.
Measurement context
- The Gemini comparison uses 15 matched questions. The four-model comparison uses 62 matched question-and-model pairs; two pairs without an earlier answer were excluded from both periods.
- These are historical API observations, not actual homeowner searches or consumer-app sessions. Some retained answers were truncated.
- The study documents an observed change. It does not establish that Appear caused it or measure leads, bookings or revenue.
More case studies
Local roofing
BYLTup
Appear reports a 73% ChatGPT mention rate and a 64% citation rate for the Long Beach roofer after eight weeks, alongside 18 published answer pages.
Read case studyInsurance claims
Imperial Claims
Appear's twelve-week snapshot reports 69% ChatGPT mentions and 58% citations. A separate seven-question comparison tracked the initial change after four answer pages went live.
Read case studyPlan your next step in AI search.
Discuss your website, customer questions and priorities with TRC. We'll help define the work and how to measure it. Results are not guaranteed.
Book a free strategy call