Recruitment Agency KPIs That Actually Predict Revenue

The recruitment KPIs that predict revenue are the ones that sit just upstream of a placement: live exclusive or retained roles, CVs sent per role, CV-to-interview conversion, and how long it takes to get a shortlist to the client. Calls made and CVs sourced are activity. They only matter if they turn into those numbers.

There is no universal benchmark to copy. The ranges published by software vendors vary a lot, and I haven't found an independent source that proves any single ratio predicts revenue. The reliable method is to track a handful of ratios on your own data and see which ones move before your billings do.

What is the difference between an activity metric and a predictive one?

An activity metric counts effort: calls, emails, CVs sourced, meetings booked. A predictive metric measures something that has to happen before a fee can be invoiced, and that you can still change.

Activity is easy to track, so agencies track it. The trouble is that two consultants can make the same number of calls and bring in very different revenue. Common guides on recruitment KPIs tend to say the same: pair pipeline activity with the economics, not one or the other.

Which KPIs should a small agency track?

Start with six. Each one can be worked out from a spreadsheet or your CRM.

1. Fees per fee earner. Total placement fees divided by the number of consultants who bill. Accountancy firms that work with agencies use this as a basic effectiveness measure (see the Menzies recruitment sector KPI sheet). It is a result, not a predictor, but it tells you whether the others are working.
2. Live roles per consultant, split by type. Count roles you are actively working, and separate exclusive or retained from open contingent roles. Roles you share with five other agencies convert differently from roles only you hold. If you've read retained vs contingent, you'll know why the split matters.
3. CVs sent per role. Too few and you're not giving the client choice. Too many and you're usually guessing. Watch for the number that tends to precede an interview.
4. CV-to-interview ratio. Of the CVs you send, how many get an interview? This is one of the earliest signals of whether you understand the brief. If it falls, the problem is usually screening or role quality, not effort.
5. Interview-to-offer and offer-to-acceptance. These show where deals are lost late. A low offer-to-acceptance rate often points to candidate expectations not being managed, or a client that is slow to move.
6. Time to shortlist. Days from taking the brief to the client seeing CVs. Speed is often what separates the agency that gets the interview from the one that doesn't, though I'd treat that as a hypothesis to test in your own data rather than a rule.

What about new-business metrics?

If you win clients through outreach, add two upstream numbers: new qualified conversations per week and conversation-to-signed-terms rate. The word "qualified" matters. A reply saying "not now" is not a lead. Count a conversation only if someone with hiring authority engaged with a real vacancy or hiring plan.

Also track signed terms of business as a share of active clients. Unsigned terms are a revenue risk waiting to happen, and they are the first thing a backdoor hire exposes. If you want to see what a structured outreach process looks like, DCA Recruit covers the finding, writing and follow-up side of it. The point here is the measurement, not the tool.

How do I find which KPI predicts my revenue?

Use your own history, not a blog's.

1. Export the last 6 to 12 months of placements, with the role, consultant and invoice date.
2. For each, note the date the role was taken, first CVs sent, first interview, offer and start.
3. Work out each ratio per consultant per month.
4. Compare to billings one to three months later. Which ratio moved first?
5. Keep the one or two that lead your billings and drop the rest from your weekly review.

With a small team, this will be noisy. A few placements a month is a small sample, so look for patterns across quarters rather than reacting to one bad week.

What does a good weekly scorecard look like?

One page, same layout every week:

• Live roles (by type) and roles added or lost
• CVs sent and interviews booked
• Offers made and accepted
• Fees invoiced and fees forecast for the next 60 days
• New qualified client conversations and terms signed

Review it with the team for fifteen minutes, then act on one thing. If the page takes an hour to build, you'll stop doing it. Several of these numbers can be pulled automatically from a CRM, and it's reasonable to automate the reporting once you know which numbers you trust.

Limits: when this doesn't apply

• Very small desks. If you place two or three people a quarter, ratios swing wildly. Track the pipeline by name instead.
• Retained and executive search. Volume ratios matter less than stage progress and client engagement. Track search stages and time between them.
• Temp and contract. Revenue is driven by hours and margin per contractor, so headcount on assignment and gross margin per week matter more than CV-to-interview.
• Gaming. Any KPI tied to pay gets gamed. Use them to find problems, not to rank people.

If you'd like help setting up a scorecard or the reporting behind it, get in touch.