Skip to main content
The Impact page shows what your completed actions changed. For each action you mark done, Bluemoon compares how often AI answers named you on the prompts that action targeted, before and after you shipped it. The headline figure also checks that change against your other prompts over the same period.

​
Where to find it

Open app.trybluemoon.com and select Actions → Impact in the sidebar. The line under the title shows how many actions you have shipped, since when, and how many are still to do. Use the selector at the top right to choose the span: Since first action, Last 30 days or Last 90 days.
The page stays empty until you mark at least one action done in Earned or Owned. Mark it done when the work is actually live: that date is what Bluemoon measures from.

​
Visibility gained on the prompts you targeted

This is the headline. It answers one question: across all your shipped actions, did the prompts you targeted gain more than your other prompts did over the same period?
  • A figure in points means the targeted prompts gained that much more than your other prompts. The line under it gives a range, “between X and Y pts”. A wide range means few answers behind the figure.
  • No proven change means the difference is too small, or rests on too few answers, to tell apart from chance.
  • A dash means one of the two periods holds too few answers. Nothing can be read yet.
The sentence below the figure spells out the counts: how many of your prompts your actions targeted, how often you were named on them before your first action and recently, and what your other prompts did. Why compare with your other prompts? AI answer engines shift for everyone. A model update or a fresh index can lift or sink your whole domain at once, whatever you shipped. Setting the targeted prompts against the rest of your domain takes that drift out, so what remains is the part more likely linked to your actions. The chart, Targeted prompts, before and after, draws the share of answers naming you on the targeted prompts and on your other prompts, day by day. A day appears once it holds enough answers on both sides. Below it, three key figures: Actions shipped, Prompts targeted out of all your tracked prompts, and Citations since first action, compared with the same number of days before it.

​
What each action moved

One row per shipped action. Each row shows whether it was Earned or Owned, when it shipped, and the share of answers naming you on that action’s prompts, before and after. The last column reads one of four ways: Click a row to see the exact counts (“named in X of Y answers”) before and after, and the range of the change. Two badges may also appear:
  • Your page is now cited: a page recorded with the action was cited in an AI answer after you shipped it. A coincidence in time, not proof.
  • First mention on [platform]: a platform that had never named you now has. This can be news for your whole domain, so do not credit it to this action alone.
A row measures the action’s own prompts. Unlike the headline, it does not subtract what your other prompts did. If a row shows a gain but the headline reads “no proven change”, your whole domain may have moved at the same time.
If several actions target the same prompts around the same time, their rows draw on the same answers and you cannot tell which one made the difference. Space them out when you want a clean read on one action.

​
How good our suggestions are

This panel grades Bluemoon’s own suggestions, before any lift is measured. It tells you whether the queue is worth working through, and it can tell you that it is not.

Backed by real evidence

The share of suggestions for this domain that rest on evidence from your answers, rather than aiming at questions you already win.

Adopted when reviewed

Of the suggestions you decided on, the share you kept rather than declined. Suggestions you never touched do not count.

Predictions track results

Whether the suggestions Bluemoon ranked higher really moved more once shipped. It reads yes, not yet, or how many shipped actions it still needs.
A dash is not a zero. “Nothing judged yet” means you have not reviewed any suggestion. A dash with counts beside it, such as “3 of 7, too few”, means there are not enough decisions yet to give a rate. What to do with it. High evidence and high adoption mean the queue is worth working top-down. Low adoption on one kind of suggestion is a signal about that kind, not about you. While predictions read “not yet”, treat the order of the queue as a reasonable guide, and use each action’s own row as the verdict.

​
What this page cannot tell you

It does not check that you shipped

Marking an action done is a declaration. Bluemoon does not check your site for the change; it takes the date and starts measuring.

It measures mentions, not causes

A gain means the targeted prompts moved more than the comparison. It is evidence, not proof that your action caused it.

Low-volume targets stay unmeasured

Prompts with very few answers will not become measurable by waiting. Track more questions on that topic to get a reading.

Quiet platforms add nothing

Everything here comes from recorded AI answers. A platform that has stopped answering for your domain adds nothing to either side. See Agent Analytics for the crawler side.