In a discussion we were having about how to measure the value of learning, Julie Wedgwood recommended me to Alan Bellinger’s ‘Magic Numbers’. As Alan is a long-standing colleague of mine from the Institute of IT Training and he had packaged up his advice so intriguingly, I had to take a look. I found a description on the site of Parity Training.
When you get beneath the ‘magic’, Alan’s ideias are really no more than a simple guide to making ROI calculations for learning interventions, but when you consider the relative innumeracy of so many in the l&d profession, any simplification has to be a good idea.
Alan maintains that, when you get down to it, there are essentially two reasons for any learning intervention: (1) to increase performance or (2) to reduce risk (as with a compliance programme). I couldn’t help wondering about those situations where you wished to take advantage of a window of opportunity (the flip side of risk reduction), to prepare employees for a future promotion, or simply to increase employee loyalty, but let’s not get picky.
When the aim is performance improvement, the numbers you need are as follows:
- The number of participants (N).
- The average salary and all related benefits of these participants (AS).
- Your estimate of the increase in performance that the intervention will lead to (P%).
- The total cost of the intervention, from development through to delivery – the investment (I). This might seem like a single number, but it’s actually a total of all sorts of different costs. To be credible, it has to include estimates for all the internal labour put in, not just the cheques signed for outside services; it must also include all participant expenses and, most importantly, an allowance for participant time away from the job.
The return (R) is calculated as (N x AS) x P%. The ROI is, naturally enough, R / I.
I do have a technical problem with the return being measured as a proportion of participant salary and benefits, as there is no reason to believe that the economic benefit to an organisation of increased performance is directly proportional to cost. As an example, a salesperson’s performance might increase by 10%, but the benefit of the additional sales is likely to be many times greater than 10% of the salesperson’s cost to the organisation. Alan’s formula is a gross simplification, but it is a lot better than nothing.
When the aim is to reduce risk, the numbers you need are as follows:
- Your estimate of the exposure the organisation is facing (EX). In other words, if the risk – a fire, a law suit, a loss of reputation, a loss of vital data, or whatever – was to occur, what would it cost the organisation?
- Your estimate of the current risk of this occurring (CR%).
- Your estimate of the reduced risk following the intervention (RR%).
- The total cost of the intervention, from development through to delivery – the investment (I). As before.
The return (r) is then calculated as EX x (CR% – RR%). Again, the ROI is R / I.
Of course, both calculations depends on the estimates that you provide. The more cynical would ask why anyone should believe you? After all, you’re not going to estimate a figure that shows your programme won’t deliver a return, now are you? At some point, you’ll have to back these estimates up with hard figures based on actual measurements from completed interventions. Once a climate of trust has developed, these ‘back of an envelope’ calculations should be more than enough to make your point.