Almost every data conversation we have starts with the wrong word.
Someone asks for their list to be “cleaned” when what they need is fields that were never there. Someone asks for “enrichment” when the actual problem is four records for the same company. The words get used interchangeably, the brief gets written loosely, and the work comes back solving a problem nobody had.
They are three different jobs. Here is what each one does to a row.
Cleaning fixes what is already there
Cleaning does not add information. It takes the data you have and makes it consistent and trustworthy: standardising Ltd, Ltd. and Limited into one form; fixing capitalisation; splitting a single name field into first and last; normalising phone numbers to a dialling format; removing records that bounce.
The important half of cleaning is deduplication, and it is harder than it looks. Two rows with the same email are obviously the same person. Two rows reading Priya Sharma, Head of Ops, Acme and P. Sharma, Operations Director, Acme Industries probably are, and no rule catches that without also merging people who are genuinely different.
That is why we write the match rule down before we run it and show it to you. A merge you did not agree to is not a fix.
Appending adds fields you never collected
Appending fills gaps. You have a name and a company; you need the job title, the company size, the industry, the location, a direct line. The record exists — half of it is missing.
This is usually what people mean when they say enrichment, and it is the one that decides whether a list is segmentable. A list of two thousand contacts with no job titles cannot be split by seniority, so every message in it has to be written for everybody. That is why it performs badly, and no amount of better copywriting fixes it.
Enrichment adds context that was never in the record at all
Enrichment goes outside the row: what technology a company runs, whether it is hiring, whether it has raised money, opened a site, changed leadership. Not attributes of the contact — signals about the account.
This is the one that changes timing rather than targeting. Appending tells you who to write to. Enrichment tells you when it is worth writing.
Why a high fill rate can be a bad outcome
Here is the part that matters commercially, and the part most vendors will not put in writing.
Fill rate — the share of records where a field was successfully added — is the number everyone reports. It is easy to make it look good. Loosen the matching rule and the fill rate climbs, because more rows find a match. Some of those matches are the wrong company with a similar name, or the right company and the wrong branch, or a person who left two years ago.
Wrong data costs more than missing data. A blank job title makes a record less useful. A wrong job title makes it actively misleading — it gets segmented into the wrong campaign, addressed with the wrong message, and the failure shows up as a poor response rate rather than as a data problem, so nobody traces it back.
So the fill rate is only meaningful next to two other things: the match rule that produced it, and a sample you can check yourself. We give you both, on every job. If a strict rule gets you 61% and a loose one gets you 88%, the 61% is usually the better file, and we will tell you so even though the bigger number would look better on our side.
Which one do you need?
A rough test, and it is usually enough:
- Records look messy, duplicated, inconsistent → cleaning
- Records are fine but half the columns are empty → appending
- Records are complete and you still do not know who to contact this month → enrichment
Most files need two of the three. Almost none needs all three at once, and anyone who tells you otherwise is quoting for the biggest job rather than the right one.
If your CRM is the thing slowing you down, our Data Ops service runs cleaning, appending and monthly hygiene as one job, and data appending on its own is a fixed piece of work with a documented match rule and a measured fill rate.