Rank Tracking and AI Prompt Tracking

Ranking used to be the whole scoreboard: where you sit for a keyword, tracked over time. That scoreboard still matters, but a second one now runs beside it, measuring whether AI engines name you at all. This chapter covers both, and how to read them without fooling yourself.

Rank tracking

Rank tracking monitors your position for a fixed set of keywords over time, by location and by device. You might ask why you need a separate tool when Search Console reports average position. The reason is that GSC average position is blended across every query, page, and place you appeared, so it smooths over real movement: a term can climb in one city and fall in another and show a flat average. A dedicated tracker checks specific keywords at a chosen locale and device, and records which SERP features are present alongside your listing. Because rank is personalized and localized, no tool reports a single true position. It reports a representative one for the locale you set, which is the honest way to read it.

Location and device are not optional settings; they change the answer. A term that sits in the top three on a phone in one country can rank very differently on a desktop in another, and the SERP features around it, from local packs to featured snippets, shift too. A good tracker records those features alongside your position, because losing a click to a snippet above you is a different problem from losing the ranking itself. Set the locale and device to match the audience you actually serve, and treat the tracked position as a stable reference line rather than a promise of what any one searcher sees.

Building the keyword set

A tracker is only as good as the keywords in it. Seed the list from sources that reflect real demand: your own Search Console Queries, Bing's Keyword Research volumes, and the suggestions from autocomplete and People Also Ask. Expand those seeds with a keyword tool to attach volume and difficulty, then classify each term by intent so you know whether it wants an article, a comparison, or a product page. Map one primary keyword plus its supporting cluster to each page, so pages do not compete for the same term. Then re-check monthly, because volumes drift, intent shifts, and new questions surface.

Your own Search Console is the richest seed source because it shows terms you already earn impressions for, including the striking-distance ones sitting just off page one. Bing's Keyword Research adds real query volumes from a second index, and the autocomplete and People Also Ask suggestions surface the exact phrasing people use. Classifying by intent keeps the set honest: a transactional term and an informational one that share words still need different pages, and mapping them to one page each is what prevents the cannibalization that splits your own signals.

Prompt tracking

Prompt tracking is the AI-era counterpart. Instead of positions for keywords, it monitors whether and how your brand and pages get cited by AI engines across a defined set of prompts. The reason it exists is that AI Overviews resolve a growing share of searches with no click at all, and Search Console does not report those resolutions as clicks, so a whole layer of visibility passes by unmeasured. You build a set of buyer-intent prompts, run them across the engines on a fixed cadence, and record for each response whether you were mentioned, cited with a link, or named first among the options. Prompt-level citation share →

Watch for a specific tell in your existing data. If a query's impressions hold steady but its clicks fall, an AI answer may be intercepting the click before it reaches you. That gap between stable impressions and sinking clicks is one of the clearest signs that answers, not competitors, are taking your traffic. On prevalence, be careful with numbers: the share of searches that trigger an AI Overview is a moving target and the reported figures vary by method and by month, so present it as a range and re-check it before you quote a single percentage.

Reading the two boards together

The payoff comes from reading classic rank and AI citation share side by side. A page can slip in the blue-link results while its citation share in answers rises, or the reverse, and only the paired view tells you which. Sometimes ranking gains and citation gains track each other, since AI engines pull heavily from pages already ranking well; sometimes they diverge, and that divergence is the signal worth investigating.

Never credit a move to your own work without ruling out a core update first. Google runs broad ranking changes several times a year, and a shift that lines up with one of those dates is often the update talking, not your optimization.

That discipline applies to both boards. Before you attribute a ranking climb to a content refresh or a citation jump to a new page, check the update calendar and rule out a broad change. Rule out a core update first → Correlating a link campaign with ranking movement takes the same care, since link effects lag and can coincide with unrelated changes. Correlating link gains with rankings → Read both scoreboards on a steady cadence, keep your attribution honest, and you have the measurement foundation for the final step: turning everything in this Academy into one audit and one plan. The Apply capstone next →