Compare yourself with rivals on Standing

Updated

All docsReading your numbers
On this page

Standing compares you with the brands you track. Its five tabs show where you stand against rivals, how that has moved over time, which engines, themes and kinds of buyer drive it, and which questions started or stopped naming you.

What are the five tabs of Standing? #

Standing has five tabs under a shared heading.

  • Against rivals (the default): every tracked brand on the five measures.
  • Over time: one measure, week by week, against a comparison.
  • By engine: your figures on each engine.
  • By group: how you do against the leader in each theme of questions and with each persona, the kind of buyer who asks.
  • Gained and lost: the questions that started or stopped naming you, with a count beside the tab name.

The five measures are Mention rate, Share of voice, Average rank, Your pages cited and Sentiment, and What the five measures mean defines each.

What is at the top of every Standing tab? #

Every tab opens with the edition line, then a figure and a headline side by side. They are the same on every tab.

The figure is Your place on mention rate, for example 2nd of 26: your place among the brands that have a Mention rate this week. The headline is worded from the leader's side, such as "Novaly leads you by 3.7 points." A second sentence compares you with the next brand below you: level with it, or ahead of it by so many points.

When a gap is small enough for chance to explain, the headline says so. Under it, a line says whether your place held and whether any of the top five moved beyond chance. If you lead, it reads You lead on mention rate.

What does the Against rivals tab show? #

Against rivals is a table of every tracked brand with five columns after the brand name: Mention rate, Share of voice, Average rank, Your pages cited and Sentiment. Your own row is tagged You.

It is sorted by Mention rate to start. Each column heading carries a small arrow: press it to sort by that column, and the arrow shows which way. A brand with no figure shows Not measured and is never ranked last. A brand whose Average rank or Sentiment rests on fewer than 10 questions shows Too few to rank on that measure, with its question count under it. When only one brand is ranked on Average rank or Sentiment, the line under its figure reads No rivals ranked instead of a place.

One short line under the table says what applies, for example Too few to rank: named in fewer than 10 questions. All figures use the same 40 questions. Where you are strong and weak, by engine then gives one card per engine with its Mention rate, with the highest marked highest and the lowest marked lowest. A last block, Gained and lost this week, links to the tab of that name.

What does the Over time tab show? #

Over time draws one measure across your recent readings. Choose the measure with the chips, which list the five in order, and the span with the Weeks shown dropdown: Last 6 readings, Last 12 or All.

A solid line is you. A dashed line is the comparison, chosen under Compare with:

  • Top rival: the best-placed other brand on that measure, which can change from week to week.
  • A named rival: the two best-placed rivals are offered by name.
  • One engine: you on that engine alone. A second row, Which engine, picks it.
  • One theme: you in that theme's questions alone. Which theme picks it.
  • One persona: you in that persona's questions alone. Which persona picks it, and a persona's name on By group opens this view.

A rival you added later has a line that starts in the week it was first tracked, and the weeks before are left empty. In the table, those weeks read Not measured, and a note says, for example, Novaly is tracked from August 31.

Show as a table swaps the chart for the same numbers, and Show as a chart swaps back. Before there are two weeks of readings to draw, the tab says Not enough readings yet.

What is the band on the Over time chart? #

The band is the likely range for each week of Mention rate, drawn around your line. It is drawn only for Mention rate, because that is the one measure with a likely range. Nothing is written under the chart: the end of each line names it, and Show as a table gives every figure.

Show as a table adds a note under the figures on how a change is called real. A final line, The newest reading is final., appears only when the newest week is complete. Why a number is blank or shows a range explains likely ranges.

What does the By engine tab show? #

By engine is a table with one row per engine that has answered, with the columns Mention rate, Share of voice, Average rank, Your pages cited, Sentiment and Answers. It is sorted by Mention rate, and the highest engine is marked strongest and the lowest weakest.

The last row, All six, blended, is your figure across every engine (the number follows how many engines answered). It is the same figure This week shows, so the two screens cannot disagree.

A note under the table names any engine that has not answered yet, with a link, Engines, to the page that lists the engines. If no engine has answered, it says No engine has answered yet. Which AI engines BotSay asks says what each engine is.

What does Why an engine is low show? #

When one engine names you less than the rest, By engine adds a card named for it, such as Why Perplexity is low.

The card gives a count: "Perplexity named a rival and not you on 14 of 40 questions".

The card carries the line Read what it says when it answers from them. and a Read button, which opens Evidence filtered to that engine. The card is a count, never a cause. It tells you where to look, not why the engine prefers a rival. It appears only when at least two engines have answered and one is clearly lowest, and a tie draws no card.

What does the By group tab show? #

By group shows how you do against the leader in each group of questions. Once you have a persona, two pills choose the groups:

  • Themes: the themes you set in Setup.
  • Personas (shown first once you have any): the kinds of buyer who ask your questions, such as IT directors or clinic managers. BotSay finds them for you from your questions.

Each row has the columns Persona or Theme, Questions, You and the leader, Gap and Change. Gap says who leads and by how much, for example Novaly by 4 points or You by 2 points, or Close when the two are near. Change carries an up or down mark, gray when the move is small enough for chance to explain. Show data table gives the same figures with last week's beside them.

The line under the table reads Found for you from your questions. Change any in Setup. on Personas, or Change themes in Setup. on Themes. Manage your questions and themes covers both. A persona's or a theme's name opens Over time for that group alone.

Why does a By group row show words instead of a gap? #

A row shows words when there is not enough to compare yet. The words are Too few questions yet, Not read yet, Not enough to compare yet and No rival named. A group with fewer than 5 questions draws no figures.

Each is a fact about the evidence and says nothing is wrong with the group. If persona figures are a week behind, a line names the week they are for. With no persona holding a read question yet, Personas says No persona has a question yet.

What does the Gained and lost tab show? #

Gained and lost lists the questions where an engine's answer changed between the last two readings. A gain is a question that did not name you and now does, and a loss names you no longer.

A strip on top shows Started naming you and Stopped naming you, each counting question and engine pairs, and Net. Net carries a color only when the difference is bigger than chance explains.

The chips Lost and Gained filter the list, and the selects Engine and Theme narrow it. Losses come first. Each row gives the question, the engine, the theme and the dates, for example "Named you on Sep 7, 2026, not on Sep 14, 2026.", with an Open link to the question.

Why does Gained and lost say "One answer changing is often noise."? #

An engine's answer to a question varies from reading to reading, so a single gain or loss does not mean much. The note under the list reads: One answer changing is often noise. Look for the same question lost on several engines.

A loss that repeats across engines, or a Net that clears the color test, is the signal. Until an engine has two readings, nothing can flip on it, and the tab says it needs two readings on the same engine. If nothing flipped between the last two readings, the tab says so and notes that this is a result, not a gap.

Tessera, Novaly, Pinefield, Kestrel Pay and Orbita are invented brands, used here as examples.