HR is, itself, a non-core function in that sense. In an induction programme I used to lead, I used to call them the "profit-making" and "profit-spending" functions - in jest, I swear!

The most logical approach is to identify what you can measure and to measure what you can identify. Obvious candidates are absence (aim to reduce), turnover (identify a healthy rate) and recruitment expenditure (less is good). You can also measure the levels of employee relations activity (how many grievances, how quickly are they resolved; how many informal warnings are issued, how many of these go on to further action etc) and engagement with other business activities (has everyone completed their mandatory training? Does everyone have a development plan? Is the appraisal process being completed on time and in full?).

But be wary. The axiom has it that we do what we measure. In other words, if we're being judged on particular metrics then obviously we're going to focus on those metrics and that isn't always in the best interests of the business. For example, to draw on my own experience, it's common to judge the performance of a salesperson by the revenue they generate, but not all revenue translates into paid invoices. So a salesperson might be incentivized to make sales to parties knowing they won't pay their invoice but not caring because it all counts as "revenue". If, on the other hand, you only incentivize sales once invoices are paid you will promote different behaviours within the sales force.

It's the same with HR. If you only emphasize reducing recruitment expenditure, the HR team is motivated to gap roles for as long as possible and to use the cheapest possible sources of candidates even if the quality of candidates from those sources means that employees don't stay as long or do the same quality of work. So an intelligent balance of priorities needs to be maintained for metrics to be useful and productive.

In reply to Robey:

In addition to Robey's comments, you should invest time with each functional manager to get a shared understanding of what adds value to the business. As an example, in finance : reducing the level of unpaid bills, the size of the working capital pot that your company needs to keep going at any one time, the reduction of financial costs (interest rates and currrency swaps.....), etc. In my last company with 70bn turnover, these activities could improve results by tens or hundreds of millions each year - more than the revenue of smaller business lines!
Once you've had these discussions you can use you hr skills to advise on setting up objective setting and monitoring processes, which the functional manager will need to calibrate with his own specific expertise for his own context. Your other role will be to identify any transversal inconsistency or contradictions across the target setting of different departments who impact each other, such as buying and accounts payable....
Good luck

This is why analytics is set to become the next primary focus (post Brexit issues I'd say) in HR.

In reply to Andrew:

@Andrew, nothing new in what I've posted. Good successful companies have been identifying and monitoring relevant performance indicators since time immemorial.....

I agree with Robey and Ray's points. It is always challenging to design KPI's for the support functions as most of the goals are qualitative than quantitative. So other than having regular goals of the department every support functional department KPI must be also linked with the core organisational achievements. For example if the organisation's overall goal is to achieve a higher ROI then this must be a KPI target for all Area/ regional manager roles and above including support. Synergy needs to be built between operational and support departments. In order do so in my previous organisation we started paying half yearly incentives/bonuses to support functions, and this brought a lot of energy and zest in the support employees even they felt equal and important part of the organisation.