Why a number is blank or shows a range

Reading your numbersUpdated

A blank, a range or a "cannot say yet" on This week or Standing is a statement about the evidence, never a hidden zero. Each section takes one thing you may see on screen and says what it means.

Why is a figure blank or shown as Not measured? #

Not measured means BotSay looked and found no figure. It is not a zero: a Mention rate of 0.0% means the engines answered and never named you, and it is ranked.

Common reasons for a blank:

Why does This week say it cannot say yet whether anything moved? #

The headline We cannot say yet whether anything moved. means there is no sound comparison with an earlier week. The line under it gives the reason:

The headline still states your place, and the figure is not changed.

What does "no real change" mean? #

no real change appears under a figure in The other four measures and means the move since last week is smaller than chance would produce, or is exactly zero.

It is a result, not an absence: BotSay measured both weeks and compared them. It does not mean the figure is fixed, because a real change smaller than the bar can be hidden by chance for a week. A larger move shows as up or down. Where BotSay cannot judge the move, it shows last week's figure and says nothing about the change.

What does it mean when a move is small enough to be chance? #

On This week, the headline can name your place and say that nothing moved beyond what chance would produce. The line under it ends: At this many answers, a change that size can happen by chance.

BotSay asks whether the change is bigger than the two weeks' own uncertainty put together, at the standard 95% level. If it is not, the move is described as chance, not as a result, even though it might turn out to be real. On Standing, By theme, a move like that reads as sitting inside the range chance explains, and a tie reads no change.

What does "It leans up, but not by enough to call a change" mean? #

It describes the size of a move that is not bigger than chance but is a good fraction of the way there. It is not a claim that something changed.

It leans up, but not by enough to call a change. (or down) is added under the headline, which stays on "nothing moved beyond chance". The word compares the move with the uncertainty, and the uncertainty also moves, so a row can change words while its rate stands still. Where the two figures print the same, the line reads the same as last week.

What is the size of change that would have counted? #

When nothing moved, the line ends with the bar BotSay used for its verdict, for example A change of 7.7 points or more would have counted.

The bar is rounded up to one decimal, so the number always agrees with the headline. It describes this week's reading and is not a target. With more answers the bar is smaller, and with fewer it is larger. If the week has too few answers for any change to count, the line says This week's reading is too small for any change to count as real. and prints no number.

What does a likely range mean? #

A likely range is the span the true figure probably sits inside, shown beside Mention rate as Likely 19.8 to 31.0%.

BotSay asks a sample of questions, not every question that could be asked, and the range shows how far the figure could sit from the truth because of that. It comes from how many answers the figure covers: fewer answers widen the range and more narrow it. It is worked out at the standard 95% level, and it is not found by asking a question again to see how it wobbles. On Standing, Over time, the range is the band behind your line.

Why do only some measures show a likely range? #

Mention rate is the only measure with a likely range, because it is a share of the answers read and a range has a clear meaning for it.

It shows on This week, on Over time and in the table beside the chart. Share of voice, Average rank, Your pages cited and Sentiment show the figure and its movement instead. Their movement is still checked against the evidence behind the figure, and a move shows as up 1.0 point, down 2.3 points or no real change only when that check can be made. Having no range on those four is the same for every workspace and is not a fault.

What does "One reading is a wide range" mean? #

On a first reading, the figure's note ends: One reading is a wide range: it narrows each week.

The likely range beside your first figure comes from how many answers that week covers, as every week's range does. It narrows when more answers stand behind a week. What builds up from week to week is the comparison: a first reading compares nothing, and from the second reading This week can say whether anything changed.

Why does a week say Provisional? #

Provisional marks a figure that may still move because the week is still being read. The edition line says Still reading: 5 of 6 in. and names the engines not finished.

Only figures the unfinished reading could still move are marked, and a provisional figure carries no likely range and no verdict, because a range on half a reading would claim more than the reading can back. Google AI Overviews and Google AI Mode send their answers after the others, and an answer that failed and is asked again can move a figure a little after the week is complete. When the last engine arrives the marks go, and Complete since and a day marks the finished week. When readings happen explains the timing.

Why does This week say it is missing answers? #

This week is missing answers: Perplexity 190 of 228 questions. names an engine that answered fewer of its questions than it should have. The sentence continues: It is not like for like with a full week, so a lower figure may not be a fall.

An engine is short when it is missing more answers than 2% of the questions it was asked, and one missing answer is always allowed. An engine that answered nothing while another did reads has not answered. BotSay does not change the figure. It withholds the comparison with last week, so a short week never reads as a fall.

Why does a theme row show too few or no change figure? #

On Standing, By theme, too few replaces a theme's rates when there is not enough behind them to judge a change. The Change column then says, for example, 2 questions cannot show a change.

Three other notes explain a row with no change figure:

Each is a fact about the evidence, and none says the theme's figure is wrong.

What is a new baseline? #

A new baseline starts when you change the country your questions are asked as, because a week asked as one country cannot be compared with a week asked as another.

This week says so in a line such as Your questions are asked as Canada from the week of 21 September, so that week starts a new baseline. The headline compares nothing across the change. On the Over time chart the line breaks where the country changes and a tick is labelled Country changed to Canada, so the chart never joins two weeks that were asked differently. The country is set in Setup, Readings.

Why does it say Last week or Week of instead of a verdict? #

Under a measure, Last week 13.1% or Week of 24 Aug 2026: 13.1% is a plain comparison with no colour and no claim. BotSay shows it where it cannot give a verdict.

A move that leans but does not clear the bar is not a claim either way. Your pages cited gives a verdict only where the check can be made, for example when each of the two weeks has at least 10 questions with a link. Week of appears, and not "Last week", when the weeks between measured nothing, and the figure is then from the nearest earlier week that did. What the five measures mean defines each measure.