Fill empty CRM fields with data you already have

When a CRM is full of empty fields, the reflex is to buy an enrichment tool. Before you do, look at what your CRM already knows. Some of the gaps can be filled from data that is sitting in another field, on another record, or in an email address, at no cost and with less risk than any external source.

Five places the answer is already hiding

1. Another property of the same record

Over the years, a CRM collects look-alike properties: a "LinkedIn" field created by an import, a "direct phone" field from an old form, a "website URL" added by an integration. The standard property is empty, but the value is right there under another name.

The trick is to recognise those columns by what they contain, not by their name. A property whose values look like linkedin.com/in/… is a LinkedIn profile, whatever it's called. Forgeo classifies a text property when at least 80% of its filled values, and at least five of them, match a pattern: LinkedIn profile, LinkedIn company page, email, URL or phone number.

2. A twin record

When two contacts share an email, or two companies share a domain, one of them often has what the other lacks: the job title, the owner, the industry. Copying a value from a twin is safe when every twin agrees on it. Better still, merge the twins: see how to find and merge duplicate contacts in HubSpot.

3. Associated records

Associations carry information both ways:

• A company's domain can come from the work email addresses of its contacts, when they all use the same one, personal mailboxes excluded.
• A company's owner can come from its contacts when they share one owner.
• A contact's owner can come from the company it's associated with.

4. A sibling property

Some values can be derived directly. The most useful one in HubSpot: the company domain from the website. The domain is what HubSpot uses to deduplicate companies and to associate contacts by email (HubSpot: deduplication of records), so a website without a domain is a missed match.

5. The email address itself

An address like marie.dupont@example.com gives a first and last name. Done well, this needs a few rules: each part has at least two letters, compound names keep their hyphen (pierre-marie becomes "Pierre-Marie"), generic mailboxes like info@ or contact@ are skipped, and the result must agree with any name already entered.

Rules that keep internal fills safe

• Never overwrite a value that exists. Fill only what is empty. A value someone typed is a decision, even when it looks wrong.
• Require agreement. When two sources disagree, show both and let a person choose.
• Separate sure from probable. A domain read from the website is sure. A first name read from an email address is probable and deserves a look.
• Keep a log, and a way back. Record the previous value, the new value and who approved it, so a batch can be undone.

Then, and only then, enrich

Once the internal sources are exhausted, what's left are the fields only the outside world knows: current job titles, industries, company sizes, LinkedIn profiles. That's where enrichment from LinkedIn comes in, on fewer records and at a lower cost.

How Forgeo does it

After each sync, Forgeo's audit proposes values for the empty fields it flags, from exactly these sources: recognised look-alike properties, twin records, associated contacts and companies, the website for the domain, and the email address for first and last names. It never overwrites an existing value. Sure proposals are applied in one click after a confirmation; the others are reviewed record by record. Nothing reaches HubSpot until an owner or admin approves it, and every batch can be undone.