Sector benchmark — what normal looks like in your sector

The problem it solves

Your Sources and Source Authority pages tell you which sites the models cite when they talk about you. What they cannot tell you is what normal looks like: your impressions drop 12% — is that a failure, or the season? A single figure says nothing. The page answers by setting your project against its market. It draws on data of different origins — and it matters not to confuse them.

If the page shows nothing

Everything hangs on the project's sector, inferred from the SEO keywords, the prompts and the name. When that inference fails, the project is « unclassified » — and it loses the demand curve, the market comparison in Impact proof and its briefs' publication dates, all at once.

• The fix lives in Settings → Project sector: a dropdown lets you set it by hand. Your choice always beats the inference.
• That is the intended way out for trades no list will ever cover — a funeral home, a niche retailer. Twenty-five families are offered; if none really fits, leave automatic detection on: no benchmark beats a wrong one.
• The setting also shows what PivotSEO inferred on its own, so you know whether to step in.

Two benchmarks, two origins

• Sector demand comes from Google Trends, for a whole category ("Home Improvement", "Restaurants", "Real Estate"…). It holds no customer data: it is a public Google statistic. It therefore shows from your very first project, with no conditions.
• Your sector's most searched queries also come from Google Trends, for the same category. Same nature: a public statistic, available from your first project.
• Your position against that market comes from your own Search Console: your impressions and your queries. They are shown only to you and never leave the project.
• Sources cited by AIs come from the scans of projects belonging to other PivotSEO subscribers in the same sector and market — not to your own projects. That is third-party data: it obeys the threshold described below, and this comparison is inactive today. The page says so instead of going blank.

Reading the demand curve

• The index runs from 0 to 100 and is never a number of searches. It gives the category's relative popularity, week by week, over twelve months.
• At the top, the next season to prepare for: the date of the next peak, the deadline to publish ahead of it (six weeks earlier, the time Google needs to index the page) and the days left. A sector with no marked season gets no date.
• Three figures: this week's demand against its median (high, low or ordinary season), the year's range, and the twelve-month trend (first quarter against last).
• The vertical axis doesn't start at zero: it runs from the year's minimum to its maximum, and it is labelled so that a 76-to-100 swing doesn't read as a collapse. Hover over the curve to read each week.
• The strongest months give you your editorial calendar. Aim to publish six to eight weeks before the month you care about.
• The Google category observed is always shown. If it doesn't match your business, the curve doesn't apply to you — that is your call to make, and it is named for that reason.
• An amber "partial match" warning flags the sectors Google carves up differently than we do. For example, "Law, accounting and finance" follows the Finance category, not law firms.

Your visibility against the market

Your Search Console impressions, grouped by week, are laid over your sector's curve. A Google Trends index and an impression count share no unit, so each curve is scaled to its own median (100). We compare movements, not volumes.

• Gap to the market over the year, in points: your trend minus the market's. +8% while the market falls 6% is +14 points. An 8% rise in a market climbing 12% is really a 4-point decline.
• These last eight weeks: where you stand against your usual level, next to the market against its own. That is the figure for right now.
• Link to the season: *strong* means your traffic follows the season — a dip in low season isn't a failure; the gap is what judges the work. *Weak* means your swings come from elsewhere: rankings, pages, brand.
• It takes 26 complete weeks of shared history. Below that, the card says so and shows nothing: a gap computed over a few weeks would give a coincidence the appearance of a fact.
• The history comes from the Search Console archive. If the card says it is empty, open the project's Search Console page once: it retrieves sixteen months in one go. The Google account must be connected.

What your market searches for most

The Google category's 25 most searched queries over twelve months, matched against your Search Console queries. A sector query counts as covered when one of your queries contains all its words: « rug » is covered by « persian rug montreal ». The best variant and its position are shown.

• Presence: how many of these queries you get impressions on. Top 10: how many are in the first ten results. To push: visible but outside the top 10 — the quickest wins, with a link to strengthen the page.
• Searches for your brand are excluded. They measure your awareness, not your place in the market. They are recognised from your brand names and the root of their domains (those declared in the project settings); the screen says how many were set aside.
• Queries where you don't appear are folded away. The Google category is broader than your offer: pick what concerns you. Some are store names, not worth targeting.
• Singular and plural count as one query.
• Rising fast, close to your business: queries climbing in the category, filtered by the vocabulary of your SEO keywords. Without that filter, the list mixes trivia with real opportunities.
• The bar under each query shows relative interest (100 = the category's most searched), never a volume.

What is shared, and what is not

• Shared: the *domains cited* by AI engines in your project's answers. These are public sites (a news outlet, a directory, an encyclopedia), and the fact that an engine cites them is public by nature.
• Never shared: your project's name, your domain, your keywords, your prompts, your rankings, your traffic. None of it goes into the aggregate and none of it comes out.
• What you get back: the list of domains cited in your sector, their weight (how many projects in the sector they appear for), and a "cites you" or "doesn't cite you" marker.
• What you don't get: the identity of the other projects. It simply isn't in the response — no setting can make it appear.

Why the cross-subscriber comparison is inactive

Pooling one subscriber's data for another's benefit means stating so in the terms of service first. That is a publisher's decision, not a technical setting: the feature therefore ships switched off, and nothing is shared between subscribers today. The settings panel shows its real state at any time.

The five-project threshold

The comparison of cited sources is only computed from five projects in the same sector and market onwards. That threshold is deliberate and not negotiable: an aggregate built on two projects would point straight at those two projects, and "the sector's sources" would become a roundabout way of reading a neighbour's data. Until it is met, the page says so plainly — but the demand curve stays on screen: it comes from no other project.

How to read the page

• 1. Open the External presence hub, then the Sector benchmark tab.
• 2. Under the title: the sector and market used, with a link to change them. If either doesn't match your reality, nothing below applies to you.
• 3. The next season to prepare for, then the demand curve with its three figures and strongest months.
• 4. Your visibility against the market: the gap in points, over the year and over the last eight weeks.
• 5. What your market searches for most: your presence on the sector's main queries, and the briefs to start.
• 6. At the bottom, folded: the comparison with other subscribers — inactive today. If it is ever switched on and the threshold is met: the sector's sources that don't cite you yet, then those that already do.

Opting out, or back in

In Settings, the "Withdraw this project from the sector benchmark" checkbox removes the project from the cited-sources aggregate. The trade is symmetrical: a withdrawn project no longer contributes and no longer has access to that aggregate — there is no observer seat. The Google demand curve remains available: it shares nothing, so there is nothing to withdraw. The choice takes effect immediately, without going through the "Save configuration" button, and can be undone the same way.