Data Append
Data append is the practice of matching your existing contact records against external data sources to add specific missing fields, such as phone numbers, emails, or firmographics.
Key takeaways
- An append is a batch join: partial records in, requested fields and a match flag out.
- The uniqueness of the match key sets the ceiling on achievable match rate.
- Loosening the confidence threshold raises match rate and attaches wrong people to records.
- Clean and deduplicate the file before submitting, because you pay per row either way.
- An appended phone number provides reach, never consent or evidence of buying intent.
In depth
An append is a join between two tables. You hand the provider a file of partial records and nominate a match key, the provider joins that key against its reference database, and it returns the fields you asked for along with a match flag and often a confidence score. Keys are not equal: a corporate email or a domain identifies a record almost uniquely, while a name plus a city is ambiguous enough that several candidate rows compete. Unmatched rows come back empty rather than guessed.
Two things move the outcome. The first is file hygiene before submission, because duplicates, inconsistent company spellings and trailing legal suffixes all split what should be one match into several near-misses. The second is the confidence threshold you accept. Loosening it lifts the headline match rate and quietly attaches the wrong person to a record, which is worse than an empty field because nothing downstream flags it. Broader field requests also cost more per row and vary in reliability by field type.
The usual workflow is periodic rather than continuous: export, clean, submit, review, merge. Load the returned file into a staging table first, sample it, then merge only into fields that were empty. Old quiz leads are a common candidate, since a scorecard funnel captures an email and rich answers but rarely a phone number, and appending direct dials turns a dormant, well-qualified segment into something an outbound team can work without asking those questions again.
Append fills gaps in what you know; it says nothing about whether the person wants to hear from you. A phone number is reach, not permission, and cold calling appended numbers is restricted or prohibited for consumers in many markets. Appending also cannot rescue a list whose records are simply too old, because the underlying people have moved on and the correct new value belongs to a different company. And a match against a common name is a coin flip dressed up as data.
Example in practice
How to measure it
Start with match rate, calculated as matched rows divided by rows submitted, then look at fill rate per requested field, because a job can match the record and still return no phone number. Divide total cost by matched rows for cost per matched record, and by verified rows for cost per usable record. The gap between those two numbers is what a loose confidence threshold is costing you.
Then measure downstream. Appended emails should be judged by their bounce rate on the first send, and appended phone numbers by connect rate and wrong-number rate on the first calling block. Compare meetings booked from the appended segment against a comparable segment you did not append. If the appended cohort performs no better, the gap you filled was not the gap that mattered.
Common mistakes
The most expensive mistake is appending onto an unclean file. Duplicate rows are billed twice and come back with slightly different values, and inconsistent company names cause the same organisation to match three ways. Standardise the key columns, strip legal suffixes, deduplicate on email, and drop rows that lack any usable key at all before you submit. A short cleaning pass usually costs less than the wasted rows it prevents.
The second is merging the returned file straight into production. Load it into a staging table, check the match flags, and hand-verify a random sample against company websites before anything touches live records. Then write only into empty fields and keep the provider name and append date alongside each value. Without that provenance, a wrong value is impossible to trace and impossible to roll back a quarter later.
Frequently asked questions
What determines the match rate of a data append?
The uniqueness and quality of your match key matters most, such as a verified email or company domain versus a common name. Cleaner input data and a reputable provider both raise the percentage of records that return a match.
What is the difference between data append and contact enrichment?
Append is narrow and batch-oriented: you name the specific fields you want filled on an existing file and the provider returns those columns. Enrichment is usually continuous and broader, firing on a single record at the moment of capture and returning whatever profile the provider holds. Many teams use both, enrichment for new leads and append for periodic backfill of the database.
Which match key gives the highest match rate?
A corporate email address is strongest because it identifies both the person and the company. A company domain is next, though it resolves the organisation rather than the individual. Name plus company works reasonably; name plus postal address is workable for consumer files. Name alone is too ambiguous to be safe and should not be used as a key.
Can I cold call phone numbers added through an append?
Not automatically. Rules differ sharply by market and by whether the contact is a consumer or a business, and several countries require prior consent before any marketing call to an individual. Check national do-not-call registers, record where the number came from, and get a legal review before an appended list is loaded into a dialler.
Should appended data overwrite existing CRM values?
No. Fill empty fields and leave populated ones alone, especially anything the contact told you directly. Where the appended value conflicts with an existing one, write it to a shadow field and flag the record for review. Blind overwriting is how a correct, self-reported job title gets replaced by a stale entry from an old index.
How is data append usually priced?
Either per record submitted or per record successfully matched, and the difference matters enormously on a file that matches poorly. Ask which model applies before you commit, ask whether unmatched rows are billed, and check whether verified fields such as validated phone numbers cost more than unverified ones. Request a test file so you can compare providers on your own data.
How often should I run an append on my database?
Treat it as periodic hygiene rather than a standing process, typically alongside your other database cleaning cycles or ahead of a campaign that needs a specific field. Running it more often mostly pays for values that have not changed. A better trigger is a new use case, such as an outbound calling push that needs phone coverage the database does not yet have.