First, measure the actual gap
Write down the ten to fifteen questions your buyers genuinely ask, put each to the engines your market uses, and record who is named and which domains are cited. Verbatim, with dates. This half-day of work replaces months of guessing.
When we ran this for one service business, fifteen questions across five engines, the shape was stark: the brand was named once in 75 answers, and a single incumbent competitor was cited 27 times. The rest of the market barely existed either. Answers concentrate hard on whoever the sources agree on.
Then read where the winner actually lives
Take the answers that recommend your competitor and look at the citations. In service categories the pattern repeats: the cited domains are directories, review platforms and listicle pages, plus sometimes the winner's own question-shaped content. The competitor is winning upstream of their website, on pages that list and describe them.
This is the part most losing brands misread. They compare websites, conclude theirs is prettier, and stay confused. The engines are not judging websites side by side; they are following what the pages they trust keep saying, and those pages keep saying the incumbent's name.
The three gaps that explain almost every case
Gap one: presence. You are missing from the specific third-party pages the answers cite. Claimable listings, review volume, inclusion in the roundups that matter. Boring, decisive.
Gap two: answers. The winner has a page that directly addresses the question and you have a brochure. Engines quote specifics: prices, timelines, formats, named customers. Positioning language is invisible to them.
Gap three: entity clarity. The winner is described consistently everywhere; you are described three different ways in four places, or your site is technically unreadable to crawlers. The engine can pin them down and cannot pin you down, so it stays silent about you.
Close it in that order
Fix readability and schema first, because everything else depends on it. Then ship one page per losing question. Then work the cited third-party surfaces, honestly: accurate listings and real reviews, not manufactured ones. Then re-run your question set on a schedule and watch the per-question wins, because that is how the gap closes, one answer at a time, not all at once.
The encouraging part of the concentration effect: it works in both directions. Categories where one name dominates are categories where the engines follow their sources tightly, so when the sources start saying your name too, the answers move.