The historical analysis shows which senders would have received banners in the recent past. Use it to spot false positives and add reliable senders to your safe list before the rollout. It adds no banner and changes nothing in the mailboxes.
- Do
- Expected result
- Verify it worked
- Revoke or undo
- Limit the scope
- What does it analyse?
- What does the report contain?
- How do I use the report?
Do
- Open Run historical analysis and select up to 100 users in Users. Pick users from different departments, so that the analysis covers varied traffic.
- Click Analyze. Analyses may take several minutes to complete.
- When the analysis appears in the table (Date, Begin analyse, End analyse, Users), click the download icon on its row.
Run historical analysis: users selected, Analyze, and the finished analyses with their download icon
Expected result
After a few minutes the analysis appears in the table, and its download icon gives you the report (see What does the report contain?, below).
Verify it worked
The analysis is listed in the table with its download icon, and the step is ticked.
Revoke or undo
Nothing to undo, the analysis changes nothing in the mailboxes.
Limit the scope
Only the users you select are analysed, 100 at most per analysis.
What does it analyse?
The emails received by the users you select, up to 100 users per analysis, over the last 12 weeks. It reads at most 100 emails per user per day. The users do not need to be active in Smart Banners; users who Can't be activated are not listed.
The analysis runs fewer checks than live Smart Banners. It does not flag first-time senders, recently created domains, file-transfer links, or the impersonation of business partners and key people, and it does not run the AI brand check. Once users are active, banners are therefore more frequent than the report shows, grey ones especially.
What does the report contain?
A CSV file (banners_analysis- followed by the date) with one row per sender whose emails would have received at least one banner. The senders with the most banners come first. Senders that would not have triggered any banner are not listed.
| Column | What it contains |
|---|---|
| email_address | Address of the sender |
| processed | Emails from this sender that were analysed |
| banners | Emails from this sender that would have received a banner, all levels together |
| suspicious | Emails that would have received an orange banner |
| very_suspicious | Emails that would have received a red banner |
| info | Emails that would have received a grey banner |
How do I use the report?
Start from the top. A sender you know and trust, flagged on most of its emails, is a candidate for your safe list (Step 5: Define your safe lists). Add the address, or the whole domain if every address of that domain is reliable. The file does not say which reason triggered each banner.
Details: Safe list