New: Competitor Benchmarking Is Now on Every Plan

Why We Made This Available to Everyone

Competitor benchmarking used to be limited to higher tiers. The reasoning made sense at launch: we assumed smaller teams were focused on their own brand and less concerned with competitive positioning. We were wrong.

The feedback was consistent, and it came from exactly the users we’d reasoned ourselves out of serving: knowing your own score in isolation doesn’t tell you whether you’re winning or losing. A Unified Visibility Score of 68 sounds good until you find your main competitor at 84. A score of 52 feels discouraging until you realise every brand in your category sits in the 45–55 band and you’re above the middle of it.

Competitive context isn’t a premium enhancement to the data. It’s what makes the data interpretable. So it moved to every plan, including free.

The design decision underneath it

There’s a reason this was cheap to un-gate, and it’s the more interesting half of the story.

Competitor data doesn’t come from separate queries. When we scan a prompt like “best CRM for a small team,” the answer already names whoever the engine chose to name — that’s the whole point of the question. The analyzer records those brand mentions alongside yours, out of the same response.

So tracking a competitor costs no additional engine calls, no additional credits and no additional query budget. It’s a different reading of a scan you were already paying for. Gating it was never protecting a real cost; it was protecting a price tier, and once we looked at it directly that was hard to defend.

This also explains a property of the feature that surprises people, covered below: competitor share of voice is measured over your query set, because that’s the set the answers came from.

What You Can Now Do

Add your competitors. The Tracked Competitors card on the Competitors page takes the brand names you want followed, per project — up to 20. They’re the names the engines would use, not domains, because what’s being matched is how a brand is referred to in an answer. Competitor data appears with your next scan and accumulates from there: on Free you scan manually, and every paid plan refreshes weekly.

Head-to-head comparison. The benchmark table shows your score alongside each tracked competitor’s, with trends, so you can see who’s gaining and who’s slipping.

Share of voice by engine. For each engine, share of voice is the proportion of relevant mentions belonging to each brand in your competitive set. This is where platform-level differences show up — it’s common to hold a strong position on one engine and be near-absent on another.

Query-level breakdown. Drill into a comparison to see which prompts you win, which a competitor wins, and which neither of you appears in. This is where the actionable detail lives. A prompt a competitor wins consistently while you don’t appear at all is a specific, addressable content gap — see content gap analysis.

How to Set It Up

Adding competitors takes about two minutes:

1. Open Competitors in the dashboard sidebar
2. In the Tracked Competitors card, type a competitor’s brand name
3. Press Enter or click add — changes save immediately
4. Repeat for each competitor you want followed

Then run a scan, or wait for your plan’s scheduled one. Historical data for a competitor starts from the date you add them; there’s no backfill, because we can’t retroactively analyse answers for a brand we weren’t looking for.

How to Read the Data

Four things worth understanding before you draw conclusions. The first is the one most likely to get misquoted in a board deck.

Share of voice is relative to your query set. Your prompts are built around your category and your buyers. A competitor’s share reflects how often they appear in the questions relevant to your customers — not their overall AI presence across everything anyone might ask. That’s usually the more decision-useful number, and it is not the number the phrase “share of voice” implies to most listeners. Say which one you’re reporting. See share of voice explained.

Scores fluctuate, and some of that is sampling. Generated answers aren’t deterministic; the same prompt can return a different shortlist on consecutive runs. Add genuine week-to-week movement in what the web says about both of you and single-week swings carry very little signal. Read four-week trends for direction and treat weekly data as a prompt to investigate, not a result.

A gap can be legitimate and still worth tracking. A competitor with a decade of press and ten times the review volume will out-score you, and no content sprint closes that this quarter. The useful questions are whether the gap is narrowing, and which specific query clusters account for it. A uniform deficit across every prompt is an authority problem; a deficit concentrated in six prompts is a content problem, and only one of those is fixable this month.

Absence isn’t always evidence. If neither of you appears for a prompt, that may mean the engine answered without naming any brand — which says something about the question, not about you.

Pairing it with your Google data

Competitive AI data gets considerably sharper when you put it next to classic search data, because the two disagree in informative ways.

The Search↔AI gap view joins the queries Google Search Console reports for your site against the prompts your scans actually ran. Read alongside the competitor breakdown, it separates two situations that look identical on a single dashboard: prompts where you rank well in Google but a competitor owns the AI answer — a genuine AI-layer problem, since the demand is proven and you’re already visible in search — and prompts where neither surface shows you, which is a broader positioning problem that no amount of AI-specific work will solve on its own.

It’s also where the discipline of not over-claiming matters most. The gap analysis deliberately refuses to report a query with no matching prompt as “AI ignores you,” or an AI mention with no matching query as “you don’t rank.” Google withholds low-volume queries entirely — its own deep dive on Search Console performance data explains that anonymized queries are excluded from the query table while still counting toward the chart totals, which is why summing that table never reaches the total. Those rows are surfaced as coverage, never as findings. In a competitive context that restraint earns its keep: “our competitor beats us here” is a claim someone will act on, and it should never be built on a row that was missing for privacy reasons.

One thing to keep in mind on timing: the query and page rows that this join runs on carry roughly 90 days of history when you first connect Search Console, not the longer window that the daily totals chart covers. The comparison is meaningful from day one, but it isn’t a multi-year backtest.

What this doesn’t tell you

Two limits worth stating plainly, because competitive data invites over-reading.

It doesn’t measure market share, revenue or customer preference. It measures how often engines name each brand in a defined set of answers. Those things may correlate; we have no basis for claiming they do, and neither does anyone else selling you a number in this category.

And it doesn’t establish causation when a gap moves. If your share rises while a competitor’s falls, the plausible explanations include your content work, their content work, a model update, a shift in the third-party sources engines draw on, and sampling noise. Investigate before you attribute — see how to do an AEO competitor analysis for a structured way through that.

Where to start

Pick your three closest competitors rather than the twenty companies your category page lists. Run a scan. Then go straight to the query-level breakdown and find the prompts where a competitor appears and you don’t — that list is the most directly actionable output of the whole feature, and it’s usually shorter and more specific than teams expect.

Competitor benchmarking is live on all accounts. We’d like to hear what you find.