ACTIVITY ≠ OUTCOME

Recruitment has always been easy to measure badly.

Applications, interviews, offers, hires, time to hire, cost per hire, conversion rates. None of these measures are wrong, and most of them are useful. The trouble starts when we stop treating them as indicators and start treating them as proof that the work is good.

Most recruitment metrics are proxies for something else. Application volume stands in for attraction quality. Interview volume stands in for process effectiveness. Time to hire becomes shorthand for efficiency, and cost per hire becomes shorthand for value. That is fine until the proxy becomes the target.

Economists call this Goodhart's Law: when a measure becomes a target, it stops being a good measure. In recruitment it happens quickly. Ask for a lower time to hire and it will fall, perhaps because hiring managers are nudged into accepting the first shortlist. Ask for lower agency spend and it will fall, perhaps because the hardest roles are left open for longer. Ask for more applications and the top of the funnel widens, along with the pile of people nobody has time to consider properly.

The metric improves. Whether the work improves is another question entirely.

Take time to hire. If it drops from 52 days to 39, that may be good news: clearer decision making, better process design, stronger recruiter control. It could just as easily mean that easier roles made up more of the mix, that steps were removed without improving quality, or that people found ways around the process. Without context, all we know is that the process took fewer days. Everything else is interpretation.

It doesn't have to go that way. I have worked in a recruitment function where agency dependency had reached 88%. On paper, reducing that number could easily have become the objective. It wasn't. The real question was why the organisation needed agencies to that extent in the first place. Once we looked behind the percentage, the work became about rebuilding direct sourcing capability, creating internal control and giving the function the ability to deliver for itself. Agency dependency eventually fell below 15%, but that number was the consequence of fixing the system, not the target we designed the work around.

That is the difference between a metric used as a target and a metric used as a question. Organisations often respond to weak insight by collecting more data: more dashboards, more measures, more reporting. But more data does not automatically create more understanding. Sometimes it just gives people more confidence in the wrong conclusion.

The better question is not "What else can we measure?" It is "What are we trying to understand?" The measure has to serve the question, not the other way around.

So what should sit alongside the familiar numbers?

Metric	What it actually tells you	Read it alongside
Application volume	How many people applied	Quality at shortlist, source mix, drop-off rate
Interview volume	How much activity the process is generating	Interview to offer ratio, hiring manager feedback, candidate experience
Time to hire	How many days the process took	Role mix, offer acceptance, quality of hire
Cost per hire	What was spent	First-year retention, performance, ramp-up time, reliance on agencies

None of this means throwing out familiar recruitment metrics. It means using them properly. A useful metric should prompt a better question:

• Why did applications increase?

• What changed when time to hire fell?

• Why are candidates declining?

• Where is agency spend still adding value?

• Why does one business area convert better than another?

That is where the useful work starts. The strongest recruitment functions are not the ones with the busiest dashboards. They are the ones that understand what the numbers mean, what they do not mean, and where more evidence is needed before a conclusion is drawn.

Next time one of your numbers moves, pick it and ask what sits behind it. If you would like a second pair of eyes on what your data is really saying, get in touch.

The number tells you what happened. The work is understanding why.