NEON DEER DATA LABS Search documentation...

neondeerdata.com

Scans and schedules

Check your Attio People and Companies for duplicates with a full scan, run one on a schedule, and let automatic detection check new and updated records between scans. You can also read a scan's status, find its results, and pick up a scan that did not finish.

Three ways duplicates are found

MethodWhat it checksWhen it runs
Full scan Every record of one object, People or Companies, against every switched-on rule, including domain redirect checks when they are on. When you press Run full scan.
Scheduled scan The same full scan of one object. Daily, weekly or monthly at a time you choose. Not included on every plan; see Settings → Billing in Neon Deer.
Automatic detection A record that was just created or updated in Attio, compared with records already known. As Attio reports each change, between scans.

People and Companies are scanned separately, each with its own rules, results and schedule. None of these methods changes Attio or merges anything: merges run only after someone approves them (see Merging records).

Running a full scan

  1. Open Scans. The page lists previous scans. Press Run full scan; the Run a full scan panel shows one row per object.
  2. Choose Scan People or Scan Companies. One scan per object runs at a time on each Attio connection, so a running object has nothing to start until its scan ends.
  3. On a first scan, pick its depth. The first scan that would check websites offers Fast scan (recommended), which finds most duplicates quickly and leaves website checks to the next scan, or Complete scan, which also checks where each website redirects (through checkredirects.io; see Domains and redirects) and takes longer.

Your plan includes a number of scans each month. The Scans page shows how many you have used. Each scan is labeled with how it started: Manual, Scheduled or Re-scan.

Following a running scan

You can leave a running scan's page and come back later; progress updates on its own. The page shows:

  • Stages: Queued, Reading records, Checking domains (when the scan checks websites), Running rules, Building sets and Finished.
  • Now and Progress: records read, domain checks finished and rules complete, as counts rather than percentages.
  • The Rules table: each rule's state, one of Waiting to run, Running, Finished, Skipped, Stopped or Failed.
  • The activity log: Live activity while the scan runs, Scan activity afterwards. Use Show to filter it to Milestones, Rules or Warnings.

To cancel, press Stop this scan.

Reading scan status

StatusWhat it meansWhat to do
RunningThe scan is in progress.Wait, or follow it on its page.
CompleteEvery record and check finished.Review the results.
Partial The scan finished but did not cover everything. The reason is shown beside the status, for example unfinished website checks. See When a scan does not finish.
StoppedSomeone stopped the scan.Start a new scan when you are ready.
FailedThe scan could not finish.Use the action the scan's page offers.
InterruptedThe scan was stopped on Neon Deer's side, for example by a service interruption.Run the scan again.

The Duplicates overview summarizes each object in its Duplicate coverage section with the words Fully checked, Partially checked, Running, Stopped, Failed or Interrupted.

When a scan does not finish

  • Partial, with unfinished checks. Press Retry unfinished checks. It resumes the same scan and retries only the checks that did not finish, so nothing already done runs again. If the scan can no longer be resumed, the page says so and offers Run a new scan.
  • Partial, at your plan's record limit. The scan stopped at the record limit: the rest of the workspace was not checked and the scan proposed no merges.
  • Stopped, Failed or Interrupted. Nothing was written to Attio. The scan's page offers the next step for the cause: run the scan again, Reconnect in Settings when the Attio connection lapsed, Adjust the rules when the scan ran out of time, or Email support when only Neon Deer can help.

Where the results are

A finished scan's page sums up what it found:

ItemWhat it shows
Records scannedHow many records the scan read.
Sets found and New setsAll duplicate sets the scan found, and those not found before.
DurationHow long the scan took.
Matches by ruleSets per rule. A set can match more than one rule, so these can add up to more than Sets found.
Scan issuesAnything that limited the scan, such as unfinished checks.
ExportDownload CSV saves the sets this scan found.

The sets themselves are reviewed on the Sets tab, sorted into Not reviewed, Ready to merge and Needs attention. Reviewing duplicates explains what each set shows and the decisions you can make.

What a new scan keeps

A completed full scan replaces the older results for that object. A partial scan only adds what it found.

Kept after a full scanCleared after a full scan
Sets it finds again, with their flags and assignments; records you marked Not duplicate; the record you chose to stay; duplicates automatic detection found since the last scan. Sets it no longer finds, with their flags, assignments and Ready to merge status.

Replaced scans stay on the Scans page for reference.

Scheduled scans

A schedule runs a full scan of one object for you. Each object has its own schedule. Whether scheduled scans are included depends on your Neon Deer plan (see Settings → Billing in Neon Deer); when they are not, the Schedules page links to Billing.

FrequencyYou choose
DailyThe days of the week it runs, and the hour.
WeeklyOne day of the week, and the hour. New schedules start as weekly.
MonthlyA day from 1 to 28, so no month is skipped, and the hour.
  • Times use the workspace's time zone, which you change on the Schedules page or in General settings. The change applies to every schedule at once.
  • Scheduled scans use the same monthly scan allowance as scans you start yourself.
  • A schedule can pause on its own. Its row says why: the monthly scan limit was reached (it resumes when allowance is available), payment was failing, or the plan no longer includes scheduled scans.

Automatic detection

Automatic detection receives record changes from Attio, so new and updated People and Companies are checked for duplicates without waiting for the next scan. Matches it finds join your results alongside scan findings, and a later full scan keeps them.

  • It runs through the Duplicate Detection app in Attio. It is a step in setup, and its status appears in Settings » Automatic detection.
  • If it stops working, Settings and setup say that new and updated records are not being checked. Reconnect it in Attio to resume, then run a full scan to catch the changes made while it was off.
  • It checks only records that change. A record nobody has edited since it was created is checked by a full scan.
  • It uses the record data in Attio. Domain redirect checks run only in full scans.
  • It never merges records or writes to Attio.

Example: a weekly routine with an import

A team keeps about 8,000 Companies and 20,000 People in Attio (illustrative figures). Companies change steadily through the week; People arrive in bulk from event lists.

WhenWhat runsWhy
Every Monday, 07:00 A Weekly schedule for Companies. Rechecks every company, including domain redirects, which automatic detection does not check.
Tuesday, after importing 1,500 event contacts Run full scan, then Scan People. Compares the whole import with every existing person at once, instead of waiting for a People schedule.
All week Automatic detection. A rep adds a company on Wednesday that is already in Attio under another name; it is checked as soon as Attio reports it.

On Tuesday the People scan ends as Complete with, say, Records scanned 21,500 and New sets 140. The team opens the Sets tab and works through Not reviewed. If the scan had ended Partial because some domain checks did not finish, Retry unfinished checks would finish the same scan.