Original Study · 18 Aug 2026

OpenAI Recommended Plumbers 3.25× More Often Than Electricians

Recommendation rates ranged from 8.7% for plumbers and auto repair shops to 2.7% for electricians. More than half of the 300 market-and-industry cells produced no recommendation at all.

Study sample

3,000 business-category records

Measured outcome

Business recommendations

AI platform

OpenAI

What this means for marketers

Benchmark AI visibility inside the client’s niche and market. Prioritize Maps visibility and review growth before weaker signals, then measure the same fixed prompt panel over time.

• Use a niche-and-market baseline instead of comparing every client with one site-wide average.
• Put Maps visibility first because it showed the largest measured relationship with recommendations.
• Treat review growth as a supporting priority and raw referring-domain count as a weaker direct AEO signal.
• Track a fixed prompt panel over time so campaign reporting separates improvement from normal answer variation.

What we tested

Do local businesses have the same chance of being recommended by OpenAI across industries and markets? The study covered 3,000 business-category records.

What the result does not prove

The study measured one OpenAI surface during one collection period. It found differences between niches but did not prove that industry itself caused them.

• The balanced panel gives each market and niche equal weight rather than estimating the population of US businesses.
• This was an observational study, so the niche differences do not prove that industry caused recommendation visibility.
• Each market-and-niche cell used one primary answer; repeat answers can return a different business mix.
• Twenty-nine physical business identities appeared in more than one category record, although all 3,000 market-and-niche candidate rows were unique.