# Duplicates & data hygiene

> How Kepler avoids duplicate records, what the merge dialogs do, and habits that keep your database clean.

_Collection: Companies, Contacts & Candidates (records) — Kepler Help Center. Canonical: https://keplercrm.com/support/articles/duplicates-and-data-hygiene/_

## How Kepler prevents duplicates

**Kepler checks several identity signals.** For candidates, a matching **email or LinkedIn URL** blocks creation and opens merge review; a matching phone number or an exact name warns you but can be overridden. Contact creation checks for matching contacts and also warns you when the person already exists as a candidate. Wherever Kepler creates records automatically, it matches on those signals first:

  - Accepting an inbound applicant finds an existing candidate with the same email instead of creating a twin — the application and CV attach to the existing record.

  - The [CV import](/support/articles/import-candidates-from-cvs/) flags drafts that match existing candidates and asks whether to attach the CV only, update the existing profile, or create a new candidate.

  - The [Chrome extension](/support/articles/linkedin-chrome-extension/) checks by LinkedIn URL and email before saving.

  - Contact auto-creation from synced mail checks for an existing contact before creating one.

  - Portal applications from the same email to the same job within 24 hours are deduplicated automatically.

## Merge dialogs

If you set an email on a contact or candidate that another record already uses — or a domain another company uses — Kepler opens a **merge dialog** instead of silently creating a conflict, so you can combine the records deliberately.

You can also start a merge yourself: on the **Contacts** or **Companies** list, tick exactly two rows and choose **Merge** from the **More actions** menu in the selection bar.

## Editing at the same time as a teammate

Kepler protects concurrent edits: if two people change the same record at once, the second save shows **"Conflict Detected — this record was updated by another user"** and refreshes the data. Nothing is silently overwritten — just re-apply your change on the fresh values.

## Habits that keep data clean

  - **Always capture an email address.** Records with emails self-link to conversations and resist duplication; records without them are orphans.

  - **Import in order** — companies, then contacts, then candidates — so links resolve as records land.

  - **Use selects over free text** for anything you'll filter by. "London", "london" and "Greater London" are three filters; one select option is one.

  - **Review Notifications regularly** — AI-proposed updates (new phone numbers, title changes) keep records current with one click instead of drifting stale.

## Spotting duplicates

Global search is the quickest duplicate check — search a name before creating a record and Kepler's matching surfaces near-identical spellings. On list pages, sort by name or email to eyeball clusters.

> **NOTE:** If you find duplicate candidates, contacts or companies, open a merge dialog to compare their values and choose what to keep. Kepler consolidates the selected records and their linked data into the surviving record. Ask support for help with a large historic duplicate set.