No data yet, or numbers that look wrong

The dashboard is empty

Most likely one of these:

• Onboarding is not finished. Without prompts there is nothing to poll. The dashboard says so and links you back.
• The first poll has not run yet. Polling starts within 24 hours of finishing onboarding, at 06:00 UTC. Use Run now on a prompt if you do not want to wait, or Preview with sample data to see what the dashboard will look like.
• The schedule is set to Off, or the free trial has ended so automatic polling has stopped. Check the Next data refresh card.
• You are looking at a different brand. Check the brand switcher.

A number changed and I did nothing

That is the product working. AI answers are not deterministic: the same question can produce a different shortlist tomorrow, because the model was updated, because its retrieved sources changed, or simply because of sampling. That is why we poll daily and read trends rather than single days.

Rules of thumb:

• A one-day move on one engine is noise. Look at the 7-day and 30-day line.
• A sustained drop on one engine while the others hold usually means that engine’s sources changed. Check Citations for that engine.
• A drop across all engines at once usually means something changed on your side: a site migration, a robots.txt change, pages returning errors. Run a Site Scan.

I asked ChatGPT myself and got a different answer

Expected, and not a bug. Your own chat has your history, your account’s personalisation, your location and often a different model version. We poll with a clean context, the same way every day, precisely so that the numbers are comparable over time. A measurement you can compare is worth more than one that matches your personal chat window.

The engines never mention me

Uncomfortable but useful information. In order of how often it is the cause:

1. The prompts are too broad. “Best CRM” is a question about Salesforce. Try the specific question your actual buyer asks.
2. No third-party coverage. Engines repeat what sources say. If no directory, review site or article has covered you, there is nothing to repeat. Start with Listings and section 3 of the Playbook.
3. The site is hard to read. Run a Site Scan and fix the crawlability and answerability failures.

A poll shows Partial or Failed

Partial means one engine was unavailable or rate limited during that run; the prompts that completed are counted normally. One partial is not a problem. Several failed runs in a row is, so contact support.

Related: Crawl schedule, Metrics explained, Contacting support.