The database question
If every search starts on LinkedIn, what is your database for?
A new job lands. You open LinkedIn. Somewhere in your CRM are people you've interviewed, placed or nearly placed for work like this. If finding them feels harder than starting again, the record count isn't telling you much.
Starting elsewhere may be a perfectly rational choice
A recruiter has a job to fill. If the last few database searches produced weak results or profiles that took too long to understand, going straight to an external source makes sense from their side of the desk.
Telling them to 'use the database more' doesn't address that experience. Neither does celebrating another import. You need to know what happens between entering a search and deciding whether someone is worth a call.
Sometimes the answer is simple: the agency hasn't worked that market before. Sometimes the right people are there, but their experience is buried in old CVs, notes or application history. Those situations need different treatment.
Can you find someone you already know is right?
Take a live brief and name three people you'd expect a sensible search to find. Choose candidates because you know their relevant experience, not because their profiles happen to be tidy.
Now search without using their names. Start with the work the client needs done, then try the titles and skills a recruiter would normally use. See which of the three appear and which disappear when you add filters.
Open the record of someone who didn't appear. Is the experience in the CV but absent from the skills field? Is their title unusual? Have you filtered them out on an availability field nobody has touched for a year?
You've now got something to fix. 'Our database is rubbish' could mean almost anything. 'Our location filter excludes people with no postcode, including candidates we know are local' is a problem you can test again after changing it.
Finding the person is only half the job
A result appears. You open it. There are four CVs, a row of activity entries and a note saying 'not interested'. Interested in what? At what salary? Was that last month or three years ago?
You can call and ask again. But if every promising result needs that much reconstruction, the database is making the recruiter repeat work the agency has already done.
Compare that with a note explaining that someone liked the work, turned down the commute and asked to hear about similar roles nearer Leeds. You can judge the new opportunity and open a much more useful conversation.
That's the difference between retaining a profile and retaining something the agency learned. Both take up space in the CRM. Only one saves the next recruiter from starting the relationship again.
Give the cleanup a job to do
A full database cleanup can be worthwhile. Give it a purpose before deciding the scope. If the desk can't find known contractors, start with the records and fields that explain those missed searches.
Separate information that's missing from information that's wrong. A blank availability field means you need to ask. A field marked 'available' from last spring needs a different kind of scepticism.
Choose a small group of relevant records and run them through the actual search-and-call process. Fix what prevents that work. Then check whether another recruiter can use the records more easily.
An import total won't tell you whether that happened. A known candidate appearing in the right search, with enough history to make the call, will.
Make the next search worth trying
The point isn't to keep recruiters inside the CRM at all costs. External sourcing is still part of the job. The question is whether the agency gets any advantage from the people it has already spent time getting to know.
Pick a brief you would normally take straight to LinkedIn. Give the database a fair attempt. Record why you stayed or left. Repeat with a few jobs and you'll have a better brief for improving the system than 'we need better data'.