Quick answer
Data enrichment adds fields to records you already have, such as company, industry or region, using derived, imported or third-party data. Each added field is a claim: store its source, date and confidence, measure usable coverage rather than match rate, and add only what you need and can justify.
Formula: usable coverage = records with a correct value ÷ all records
What data enrichment means
Data enrichment adds fields to records you already have, using data that did not come from the original collection. A contact record with only an email address gets a company name; a visitor with only an IP address gets a country; a user with a list of events gets a "plan" or "last purchase" field. The record gets wider, not longer: same people, more attributes. Microsoft's documentation for its customer data product frames it the same way, describing enrichment that fills in and standardises fields such as addresses, either before records from several sources are unified or after.
It helps to separate enrichment from three terms around it:
- Cleaning corrects fields that are wrong; enrichment adds fields that are missing. Standardising an address format sits in between and is often done first. See data preparation.
- Collection asks the person: a form field or a preference they chose. That is zero-party data. Enrichment gets the answer from somewhere else.
- Identity resolution decides which records belong to the same person. Enrichment usually needs that match first, because you can only add a field to the right record. See identity resolution.
The thesis of this page: an enriched field is a claim, not a fact. It has a source, a date and a chance of being wrong, and it should be stored with all three. And it should only be added when you need it and can justify where it came from.
Three kinds of enrichment, and how they differ
| Kind | Where the new field comes from | Example | Freshness | Main risk |
|---|---|---|---|---|
| Derived from your own data | Calculated from events and orders you already hold | "Active days in the last 30", "first purchase channel", "plan tier" | As fresh as your last refresh | Rule drift: the definition changes but old values stay |
| Imported from another system you own | A CRM, billing or support export matched on an ID | Company size from the sales system, subscription status from billing | As fresh as the export | Match keys disagree; the other system is wrong or stale |
| Appended from a third party | An outside provider matches your records to its own database | Job title, industry, household attributes | Unknown unless the provider says | Wrong matches, unclear origin, a duty to tell people (see below) |
Derived and imported fields are usually the safe place to start. They use data you already have a reason to hold, you can explain them to the person, and you control the definition. Third-party appends add a second organisation to the story. For what third-party data is and how it is sourced, see third-party data. The same logic that marketers apply when a platform collects data itself, rather than relying on bought data, is why customer data platforms put so much weight on first-party records.
Three numbers to measure any enrichment
An enrichment run produces a lot of rows with new values, and the first number everyone quotes, the match rate, is the least useful one. Track three:
- Match rate = records that received a value ÷ records you tried to enrich
- Accuracy = values that are correct ÷ values checked in a manual audit sample
- Usable coverage = records with a correct value ÷ all records
Usable coverage is the one that tells you what you can act on, and it is always lower than the match rate. The next section shows by how much.
Worked example: 1,000 contacts, a 70% match rate and 34% usable coverage
Sample data: a list of 1,000 sign-ups. 600 of them used a work email address, so you try to add "company size" to those 600 from an outside source. The source returns a value for 420 of them. You then audit 50 of the 420 by hand against a company's own website, and 41 are right.
| Step | Calculation | Result |
|---|---|---|
| Records attempted | given | 600 |
| Records matched | given | 420 |
| Match rate | 420 ÷ 600 | 70.0% |
| Share of all 1,000 records with a value | 420 ÷ 1,000 | 42.0% |
| Accuracy in the audit | 41 ÷ 50 | 82.0% |
| Estimated correct values | 420 × 0.82 | about 344 |
| Usable coverage | 344 ÷ 1,000 | about 34.4% |
So the headline "70% enriched" becomes 34.4% of the list with a correct value. Because the audit has only 50 records, the 82% has a wide range: a 95% Wilson interval runs from 69.2% to 90.2%, which puts the number of correct values between about 291 and 379, or 29% to 38% of the list. The honest sentence is "somewhere between 29% and 38% of records have a correct company size".
Two decisions follow. If you plan to route leads by company size, a quarter or more of the filled values may be wrong, so use the field to prioritise, not to exclude anyone. And a larger audit sample narrows the range far more cheaply than buying more data.
Store the source, the date and the confidence
Every enriched field should travel with three pieces of metadata, in the same table or a sibling table:
- Source:
derived,crm_export, or the name of the provider. - Date: when the value was written or last confirmed. Job titles and company sizes go out of date in months, not years.
- Confidence or method: exact match on a verified ID, or a fuzzy match on a name. A fuzzy match should never overwrite a field the person gave you directly.
A simple rule keeps the data honest: a value the person gave you outranks anything you appended. If someone states their company size, it replaces the appended one, and the appended one is not written back later. Without the source and date columns you cannot apply that rule, cannot expire stale values and cannot answer a person who asks where a field came from.
The privacy boundary: only what you need, and tell people
Enrichment is where analytics data quietly turns into profiles. Three questions keep it inside reasonable limits. This is general information, not legal advice; check your own situation with a qualified person.
- Do you need this field? The EU General Data Protection Regulation (GDPR, Article 5(1)(c)) asks that personal data be adequate, relevant and limited to what is necessary for the purposes. A field you "might use later" is a field you may not be able to justify. See data minimization.
- Is the new use compatible with why people gave you the data? Article 5(1)(b) requires that data be collected for specified purposes and not further processed in a way incompatible with them. A newsletter sign-up is not obviously a reason to add household attributes.
- Have you told them? When personal data is obtained from somewhere other than the person, Article 14 requires the controller to provide information including the source of the data, within a reasonable period and at the latest within one month.
Rules differ by country and by data type, so the practical habit is the same everywhere: write down the purpose of each enriched field, keep the list short, and delete fields when the purpose ends.
Common mistakes
- Enriching before matching. If the record key is wrong, the new field is attached to the wrong person. Check identity first.
- Reporting the match rate as the quality. A 70% match with 82% accuracy is 34% usable. Report all three numbers.
- Overwriting what people told you. Appended values should fill gaps only.
- No expiry. A job title appended two years ago is a guess about the present. Keep a date and a re-check period.
- Enriching everything. Adding fields for the sake of completeness raises risk and cost and rarely changes a decision. Name the decision first.
- Putting contact details into event properties. Email addresses and phone numbers belong on a user profile, not in the properties of each event.
How to enrich users and events in TapCub
Take the 1,000 sign-ups above. In TapCub you can add fields from your own data, and you can see who got them. TapCub does not provide third-party data enrichment; it neither buys nor appends outside data, so everything below uses data you already hold.
- Send the field with the event or the login. Event properties describe one occurrence, such as the plan chosen on a particular sign-up. User properties describe the person and are written when you call
tapcub.login,setPropsorsetPropsOnce.setPropsOncewrites a value only if there is none, which matches the rule that a value the person gave you outranks anything appended. User profiles require identification mode; anonymous visitors have no profile. - Import a CSV of user attributes from your own CRM or billing system. On All-platform analytics › Data › Data import, choose CSV user attributes. Each row needs at least one of a user identifier, an email or a phone number, plus the attributes you want to add. Emails and phone numbers used as identity are hashed before they are bound. Keep each file under 100,000 rows and map the columns in the preview.
- Turn a calculation into a tag. On All-platform analytics › Users › Segments & tags, the Tags tab turns a condition into a reusable field. A tag can write a fixed value when a rule matches, take the first or last value of an event property, or record the event name a person triggers most often in a window. This is how a derived field such as "heavy user" gets one definition used everywhere.
- Check the profile and the dictionary. Open a person in All-platform analytics › Users › People to see the current value of each property. The property dictionary under All-platform analytics › Data › Property dictionary classifies fields as direct identifiers, indirect identifiers or non-personal, and masks direct identifiers in lists and exports by default.
- Avoid contact details in event properties. By default TapCub rewrites strings that look like emails or phone numbers in event properties into placeholders instead of storing them as written, which is one more reason to keep contact data on the profile.
Sources
Was this explanation clear?
We use your answers to fix and expand this page.
Couldn’t send. Please try again in a minute.
Thanks for the feedback. We’ll use it to improve this page.You’ve already sent feedback on this page. Thanks.

Data Analyst
Explores traffic sources, conversion funnels, and retention, using clear metric definitions and concrete examples to explain what the numbers mean.