AnswrForge

Read your results

What each AnswrForge metric means, how scores are averaged, and how to compare brands and time periods.

The four core metrics

Click any number to see its counts, then select View matching answers to read them.

  • Mention rate: the share of analyzed answers that name your brand.
  • Recommendation rate: the share of analyzed answers that recommend your brand. It counts all analyzed answers, not only those that mention you.
  • Average position: your average rank in answers that recommend options in a clear order. Lower is better.
  • Own-domain citation rate: the share of answers that cite your website, out of answers with citation information.

How scores are averaged

AnswrForge calculates a rate for each prompt on each provider. It averages prompts equally within each provider, then averages providers equally. So 80% on one provider and 20% on another gives 50%, whatever their answer counts.

That's why the raw counts under a score can differ from the headline. Answers still awaiting analysis, or whose analysis failed, are left out, never counted as negative. Limited sample means fewer than 10 answers.

Filters and chart controls

Filters apply to every number and chart on the page, and they stay in the page link.

  • Period: Last 7 days, Last 30 days (the default), Last 90 days or All time.
  • Provider and Group, plus Country when your project tracks more than one country.
  • Chart interval: Daily, Weekly, Monthly or 7-day rolling. Calendar periods use UTC and weeks start Monday.
  • Group by: Prompt groups, Provider or Country, always with a dashed Overall average line.

Prompts and Competitors

  1. 1Tick at least two brands under Brands to compare.
  2. 2Select Compare selected brands. Only answers checked for every selected brand are used.
  3. 3Select Clear comparison to go back.

Prompts lists each prompt with its answer count and metrics. Open one to see its trend, Latest checks by provider and Competitors on this prompt.

On Competitors, each brand uses the answers checked since its own tracking started. A brand added later can rest on different answers, so compare like with like:

Reading changes over time

  • Changes compare your period with the same length of time just before it, in percentage points or positions.
  • Only answers collected with the same saved settings in both periods are compared. Not enough comparable data means there was no overlap.
  • Partial period marks a chart period cut short by your dates. A gap means no data, while a measured zero shows as zero.
  • Measured prompts/providers changed from the previous period is a coverage warning, not proof of improvement.

Still need help?

Contact us