Snapshots

Snapshots is your brand's run history: one row per tracking run. Each run asks every tracked prompt on the engines set for that prompt, shown in its Platforms column on Prompts, and the results of each run stay here, so you can confirm a run completed and open exactly what the engines said that week. Open it from Snapshots in the Monitor section of the sidebar.

How runs are scheduled#

Runs are automatic. There is no button to start one: Ranqo checks every brand once a day, and runs a brand when its last run is more than 7 days old. Each brand therefore runs on its own weekly anniversary.

A brand is skipped when:

• no prompt is tracked. Track at least one on the Prompts page;
• tracking is paused for the brand, with Pause on its Home page;
• the account's trial has ended without a paid plan. See Billing and trials.

The Home page header shows when the next run is due, for example Refreshes Sep 30. When the date has passed, it reads Refreshes tomorrow, because the next daily check picks the brand up.

Pausing and resuming does not reset the schedule. A brand resumed within the week waits for its original anniversary. One resumed after that runs at the next daily check.

The very first run starts as soon as setup finishes, or as soon as you add a brand's first generated prompts to tracking, without waiting for the schedule.

The run list#

The All Tracking Runs card lists runs newest first, a page at a time. Each row shows:

• The status: a green check for completed, a red cross for failed, a spinner while pending or running.
• The date and time the run completed, or started if it has not finished. The top row is marked Latest once it has completed.
• Visibility for the run, with its change from the run below it in green or red.
• The engines recorded for the run, and the number of results: answers that count toward the run's numbers. Each prompt is asked on its own engines, so the engines in the results can differ from the icons.
• Pos and Sent: the run's average position and an older sentiment measure. Sent averages +0.7 for each answer that named you positively, 0 for each neutral one and -0.7 for each negative one, so it runs from -0.7 to +0.7. It is not the 0 to 100 Sentiment score shown on the run's own page and everywhere else.

A run in which no answer named you shows Pos 0.0 and Sent +0.00: there was no position or sentiment to measure, not a neutral reading. The run's own page shows -- for Sentiment in that case.

Click a row to open that run.

A run's page#

The page is titled Tracking Run with your brand's name, the date and the run's status. Back to Snapshots returns to the list.

• Visibility, Position and Sentiment, each compared with the previous run, and Total Sources: the distinct domains cited, with the share of answers that cited your own site. Position reads 0.0 for a run in which no answer named you.
• All Platforms, All Categories and All Themes filters that narrow the run to part of its answers. They apply to Sentiment, Total Sources, the cited count under Position and every card below them. Visibility and Position always show the whole run, and Sentiment's change compares the filtered answers with the whole previous run. Platform Breakdown and All Results ignore the theme filter.
• Brand Ranking: how you rank against the competitors named in this run.
• Citation Sources: the domains cited in this run, with % Answers, Total Pages and % Citations. See Sources for what each column means.
• Share of Voice, Category Visibility, Position Distribution and Platform Breakdown, as on the Visibility page, for this run alone.
• All Results: every prompt on every engine in the run. Search it, filter by engine, category and outcome (Mentioned, Not Mentioned, and Failed or Imputed when the run has any), and click a row to read the answer. The dialog shows the reformatted answer when there is one, which for a long answer covers only its first 8,000 characters. See Why numbers change.

When prompts or engines changed since the previous run, a banner says Tracking configuration changed since last run and lists what changed. The comparisons with the previous run may not be like for like.

Partial runs#

An engine can fail to answer a query: it times out or returns an error. A run with some failed queries still completes. For each failed query, Ranqo looks for the most recent successful answer to the same prompt on the same engine from no more than 2 tracking intervals back (the limit is counted in days at your tracking cadence):

• Found: the earlier answer's result is reused for this run and counts toward its numbers. The row is marked Imputed in All Results, and a banner reads, for example, 3 of 60 queries used historical data.
• Not found: the query is left out of the run's numbers entirely. It shows as Failed in All Results, and it is not counted as an answer that failed to mention you. A Partial coverage banner on the run's page says how many queries could not be answered, and that the scores describe only the ones that did.

Failed runs#

A run is marked failed when it could not finish, including a run that stalls for hours. A failed run does not count as the week's run: the next daily check tries the brand again.

If a brand's very first run fails, the Visibility page says We couldn't finish your first analysis and offers Retry analysis, or Retry setup when the brand has no prompts yet.

Related#