Training ROI

What training ROI is

Training ROI (عائد الاستثمار في التدريب), also written as the return on training investment, is the net result of a training programme set against what the programme cost, expressed as a percentage. The formula is the benefit minus the cost, divided by the cost, multiplied by 100. This is the general return on investment calculation, which divides the net gain from an investment by its cost, applied to spending on training.

The figure is built rather than measured. Every element in it rests on an earlier decision: what was counted as cost, what was counted as benefit, over what period the benefit was added up, and how much of the result was attributed to the programme. Changing any one of these four can turn the result from negative to positive.

Training ROI and the benefit to cost ratio

A second figure appears alongside the return: the benefit to cost ratio (نسبة المنفعة إلى الكلفة), which is the benefit divided directly by the cost. The relationship between the two figures is fixed. The return as a percentage equals the benefit to cost ratio minus one, multiplied by 100.

A ratio of 1.432 therefore corresponds to a return of 43.2%, and a ratio of 0.955 corresponds to a negative return of 4.5%. The two can be confused when results are presented. Anyone who reports a ratio of 1.432 and has it taken as 143.2% has presented a figure more than three times the 43.2% it stands for. Every figure should therefore say which of the two it is.

The cost in training ROI goes beyond the invoice

In the example that follows, 8 employees attend a programme for 3 days, and each earns a monthly wage of SAR 14,000:

• Direct cost: SAR 3,500 per participant, or SAR 28,000 in all.
• Cost of time: dividing the monthly wage by 21 working days gives a daily rate of SAR 666.67, so three days for eight people cost SAR 16,000.
• Total: SAR 44,000.

The divisor here puts an internal price on working time. It is not a calculation of any wage owed to the employee; that is a separate subject, with its basis in the Saudi Labor Law (نظام العمل). Even so, the choice of divisor matters. On a month of 30 days, the daily rate would be SAR 466.67, the cost of time SAR 11,200 and the total SAR 39,200. The programme and the attendance are unchanged, and the difference in cost between the two calculations is SAR 4,800. A cost presented without the basis used to derive the daily rate cannot be checked.

The divisor of 21 days in the example is assumed, and the effect of changing it is shown above. In the sources we reviewed, we found nothing that prescribes a divisor for this purpose.

Beyond both calculations lie costs that appear in neither: the time of whoever organised the programme, the time managers spent nominating participants and following their progress, and the work that was held up during those days. These can be estimated but not measured. The honest course is to say that they are outside the calculation, not to assume that they are zero.

The benefit in training ROI has to be measured

The benefit must be a change in a named quantity that was measured both before the programme and after it. Without an earlier baseline there is no return, only an impression.

Suppose the hours of rework caused by errors in the team fell from 62 a month to 41, a reduction of 21 hours. Spreading the same daily rate over an assumed day of 8 hours gives an hourly rate of SAR 83.33. Valued at that rate before rounding, the 21 hours are worth SAR 1,750 a month, or SAR 21,000 a year.

This benefit is valued at the same wage that was used to value the cost of time, and that is deliberate. Valuing the cost with one yardstick and the benefit with another is a mistake this calculation is open to, and its effect is that the return can be moved by changing the yardstick alone. Whoever uses the hourly wage in the denominator has to use it in the numerator as well, and anyone who wants to include costs other than the wage when valuing an hour has to include them on both sides.

How the horizon sets the sign of training ROI

The cost is incurred once, while the benefit recurs, so the period over which the benefit is added up is a decision that can change the result:

• One year: a benefit of SAR 21,000 against a cost of SAR 44,000 gives a return of minus 52.27%, with a benefit to cost ratio of 0.477.
• Two years: a benefit of SAR 42,000 gives a return of minus 4.55%, with a ratio of 0.955.
• Three years: a benefit of SAR 63,000 gives a return of plus 43.18%, with a ratio of 1.432.

The figures in these examples are assumed in order to show how the calculation is structured. In the sources we reviewed, we found no published reference for the Saudi market on which to base a benchmark for training ROI.

The same programme can be presented as a loss or as a gain depending on the period alone. The horizon is therefore stated with the figure, and justified: what gives reason to think the effect lasts three years? If half the team changes within two years, a horizon of three years assumes a persistence that will not occur. The team’s retention rate, the share of the people employed at the start of a period who are still employed at its end, bears on that assumption.

The definition does not settle the right horizon for a particular programme. That depends on the nature of the skill and on how quickly the team changes, and the horizon is declared rather than assumed. The return also shows the size of the result at the chosen horizon, not the point at which the benefit catches up with the cost; that point is the payback period, the time an investment takes to recover its initial cost.

Attribution in training ROI

The calculation above attributes the whole fall in rework hours to the programme, and that assumption may not hold. Over the same period there may have been a change of tool, a change in how work was allocated, or a lighter season.

If 60% of the effect is attributed to the programme, the benefit over three years becomes SAR 37,800 and the return becomes minus 14.09%. The sign has flipped a second time, and nothing in the programme or in the measurement has changed.

Two things make an attribution defensible. The first is a group that did not attend the programme, measured on the same indicator over the same period. The second is staggered attendance, which allows those who attended early to be compared with those who attended later. Where neither is possible, the attribution share is stated as a declared estimate, and the return is presented at two shares rather than at one.

Training ROI measurement is designed before the programme

A figure cannot be discussed when measurement began only after the programme had ended. Four decisions that are not taken before the programme cannot be taken after it:

• The quantity. Name it before the trainer is chosen, because picking the quantity after seeing the results amounts to picking whatever improved.
• The baseline. Measure it far enough ahead of the programme to know how much it normally fluctuates. An indicator that moves 15% from one month to the next with no intervention cannot be judged on a difference of 10%.
• The comparison group. Define it before nomination, because nomination is already a selection: the employees put forward for the programme may be the ones more ready to improve in any case.
• The attribution rule. Write it down before the result is known; otherwise the share chosen will be the one that gives the desired sign.

Writing these four down on one sheet before the programme gives a figure that can be defended afterwards. Leaving them out gives a figure that can be believed or rejected, but not examined.

Training ROI across the training portfolio

An organisation spent SAR 1,200,000 on training in a year across 14 programmes. The organisation measured 4 of those programmes, which accounted for SAR 264,000 of the spending, and their average return was 38%. The accurate statement is “the measured return covers 22% of training spending”, not “the organisation’s training ROI is 38%”.

The difference is not a matter of wording. The four programmes may have been measured because they were the ones that could be linked to a figure, and those can be of one kind: operational skills whose effect shows soon. The other ten have effects that are further off and harder to measure, and nothing in the figure of 38% says anything about them in either direction.

Three figures are therefore presented together: the share of spending that was measured, the average return within that share, and what remains unmeasured and why. The three are presented as part of the organisation’s HR analytics rather than on their own, because the question of how much is measured is a question of method before it is a question of training.

What distorts a training ROI figure

• Measurement taken only afterwards. A figure taken after the programme, with no baseline, is compared with nothing.
• Benefit drawn from satisfaction. A participant’s satisfaction with the day spent in training is not a change in a quantity, and putting it into the numerator creates a return out of an impression.
• Benefit counted after departure. A benefit calculated over three years for a participant who left in the seventh month is a benefit that never occurred.
• Programmes chosen for ease of measurement. Measuring only the programmes that are easy to link to a figure produces a high average that does not describe the training portfolio.
• Avoided cost counted as revenue. Hours not spent on rework are not money received unless they were put to productive use, and that has to be stated plainly.

What training ROI is not

• Training needs analysis comes before the programme and decides whether training is the right response in the first place, while training ROI follows the programme and measures what resulted. A return calculated for a programme that addressed a problem other than a skill gap measures, with precision, the effect of the wrong remedy.
• The ADDIE model sets out the stages of building a programme, whereas training ROI is a figure derived after the programme has run.
• Activity indicators, such as training hours per employee and the attendance rate, describe what took place rather than what resulted. Presenting them in place of the return changes the question being asked.

Before a training ROI figure is presented

Four lines accompany every return: which quantity was measured and what its baseline was; what the horizon is and what justifies it; how much was attributed to the programme and on what grounds; and what went into the cost and what was left out.

A return presented with these four lines can be discussed. A bare percentage looks like information, yet it gives a decision maker nothing to agree or disagree with and nothing on which to base a decision.

Whether a return is a sound basis for deciding to continue a programme is a question the definition does not settle. Some programmes are run for reasons that do not translate into a quantity, and a negative return on such a programme is not a verdict on it. Likewise, a positive return on another programme is not sufficient reason to repeat that programme if the reason for which it was created no longer applies.

This is an explanation of the concept, not legal advice.

Related terms

• Return on Investment (ROI)
• Retention Rate
• Payback Period
• Hawthorne Effect
• Cost per Hire