When a Client Asks Why a Competitor Is Named First

It usually arrives as a screenshot. A client asked ChatGPT for the best option in their category, a competitor came first, and the message under the screenshot is short and unhappy.

We have learned to respond in two steps. First, acknowledge it without defensiveness. Second, investigate in a fixed order, because the explanation is rarely the one the client suspects.

What the client usually believes

Common theories include: the competitor pays for placement, the engine is biased, our content is bad, the competitor has more backlinks, we were penalized. Some of these may contain a grain of truth. Most are guesses.

We do not argue with them in the first reply. We say we will look into it and report back with evidence.

Step one: is it repeatable?

A single answer is a single sample. We rerun the prompt on the panel and see whether the pattern holds across variations and over days. Promptwatch records daily prompt panels per model, country, and language, so we can look at the history rather than rely on a screenshot.

Often we find the competitor is first sometimes, and the client is first other times. That changes the conversation from "we are losing" to "we are inconsistent".

Step two: which engine and which prompt type?

We check whether the pattern is specific. The competitor may lead on Perplexity and not on ChatGPT, or lead on "best for small teams" and not on "best for enterprise". The heatmap that compares brands against models and topics makes this visible quickly.

Step three: what sources does the engine use?

We open the citations for those answers. Which pages and domains does the engine point to when it names the competitor first? Are they review sites, comparison articles, community threads, the competitor's own pages? Citation analytics lists these at page and domain level.

Frequently the reason is not something the competitor's own site did. A few third-party pages list them favourably, and the engine leans on those pages.

Step four: are we reachable?

We check our own crawl status. Did the bots that matter fetch our key pages successfully? Agent Analytics shows IP-verified crawler hits and errors. A competitor may be first partly because we were unavailable to the bot.

Step five: read the competitor's cited pages

We add their cited URLs to the Page Tracker and read them. First paragraph, structure, freshness, specificity. We compare against our equivalent page honestly. Sometimes theirs is simply clearer.

Step six: look at sentiment and positioning

Being named first may reflect position and sentiment as well as frequency. If the engine describes the competitor as the choice for a segment and describes us in more general terms, the gap may be about positioning clarity, not volume. We read the actual wording of the answers.

Step seven: decide what is in our control

Now we sort the findings:

• Access problems: fix.
• Missing or weak pages: write or refresh.
• Absent from key third-party pages: plan outreach.
• Outdated claims in sources: correct.
• Genuine product differences: feed back to the client's product and marketing teams.

Some of what we find is outside our control. We say so. If the competitor is simply the better-known name, engines may favour them for a while, and the response is a patient content and mention program.

What we do not do

• Promise to overtake them by a date.
• Suggest paying for placement as though it were the same as organic visibility.
• Make claims about their product that we cannot verify.
• Attack them in content.
• Treat a screenshot as proof.

How we report back

A short note with:

1. The pattern across the panel, with the screenshot's prompt in context.
2. The main reasons, ranked, each tied to evidence.
3. What we will do, what the client needs to do, and what we cannot influence.
4. When we will recheck.

We keep it factual and unemotional. Clients calm down when they see a method.

Using the answer to improve the program

An episode like this often sharpens the program. It may add prompts to the panel, add a competitor to tracking, or reveal a page to refresh. The Action Board in Promptwatch collects the resulting tasks, so the investigation ends with work in the queue instead of a report filed away.

Ads in AI answers

If the competitor appears as a sponsored result, that is a different matter. Promptwatch's Ads Radar, available from Business and on agency plans, captures sponsored ads in ChatGPT Search, Google AI Overviews, and Google AI Mode answers for tracked prompts, including the advertiser. It does not cover Perplexity. We check it before concluding the placement is organic.

Our recommendation

Do not answer with theory. Investigate in order: repeatability, engine, sources, access, content, sentiment. We use Promptwatch for this because the panel history, citation analytics, Page Tracker, and crawler logs are in one project. See it at promptwatch.com.