Introducing confidence labels: stable, low sample, real drop

Askority team 4 min read
Key takeaways
  • Run counts are shown next to every score.
  • Moves are labelled stable, low sample or real drop.
  • Low-sample moves never trigger alerts.

Why

Half of the drops customers asked us about were single bad samples. We would rather tell you that up front than let you write a content brief for a coin flip.

What changed

Charts, tables and the diagnostics panel carry the run count. Alerts fire only on real drops.

See what AI says about your brand
Six engines, a grade and the first three fixes. Free for 14 days.
Start for free

More from the blog

GEO basics

What is GEO, and how is it different from SEO?

Generative Engine Optimization is not SEO with a new name. Here is what actually changes when the result is a paragraph instead of a list of links.

9 min read
Research

We analysed 1,200 visibility drops. Most had one cause

Across six engines and three months of data, most significant drops traced back to a single competitor page that an engine started citing.

7 min read
Playbooks

The 403 problem: why AI crawlers cannot read a third of B2B sites

A checklist for finding out whether GPTBot, PerplexityBot and ClaudeBot can reach your pages, and what to change when they cannot.

5 min read