Every proposal you receive will quote the cost of building a performance management system. Almost none will tell you what it costs to run one, every year, forever — and that is the number that determines whether it survives. This is an honest account of the ongoing cost, expressed in hours rather than rupees, because hours are what you can verify against your own organisation and rupees depend on salaries we do not know.
A note on why there are no prices here. Consultancy fees, licence costs and salary levels vary enormously by sector, size and city, and any figure we published would be either meaninglessly wide or quietly self-serving. Hours you can convert yourself, using your own numbers, and check our arithmetic.
The four costs, and which one people forget
- Design. One-off. Writing result areas and measures for every role. Real, visible, and the only one most proposals quantify.
- System. Recurring, predictable, scales with headcount. The easiest to compare between vendors and usually the smallest of the four.
- Manager time, every cycle. Recurring, invisible, and almost always the largest. Nobody invoices you for it, which is exactly why it is underestimated.
- Administration and follow-through. Recurring. Chasing completion, consolidating, preparing the review pack, acting on what it shows.
The third is the one that decides whether the system lasts. It never appears on a proposal because no supplier bills for it, and it is the cost your managers actually experience. When a performance system is quietly abandoned in year two, this is almost always why.
Manager time, per cycle, per direct report
These are the hours a quarterly cycle genuinely takes when done properly. Multiply by your own manager count and average cost to get a figure that means something for your organisation.
| Activity | Per direct report | Notes |
|---|---|---|
| Goal setting (annual) | 45–60 min | Longer in cycle one; drops once roles are stable |
| Quarterly check-in | 20–30 min × 3 | The single most valuable item on this list |
| Gathering evidence | 15–20 min per cycle | Near zero if the system captures it as work happens; substantial if reconstructed at year end |
| Writing the review | 30–45 min (annual) | Doubles where no evidence was recorded during the year |
| The review conversation | 45–60 min (annual) | Rushing this destroys the value of everything above |
| Manager’s own review | 2–3 hours (annual) | Their own cycle, as a participant |
For a manager with eight direct reports, that is roughly fourteen to eighteen hours a year, or about two working days. Spread across a year it is modest. Compressed into the fortnight before an appraisal deadline — which is what happens without quarterly check-ins — it is two days lost in a busy period, and it is why appraisals get done badly.
The quarterly rhythm is not additional work. It is the same work, distributed, and it produces better evidence because nobody is reconstructing October from memory in March.
What drives the cost up or down
| Factor | Effect | Why |
|---|---|---|
| Evidence captured as work happens | Large reduction | Removes reconstruction, the biggest single waste in the cycle |
| Quarterly rather than annual only | Moderate reduction | Same hours, spread; better data; no deadline crush |
| Result areas stable year to year | Large reduction | Design cost is one-off only if roles stop being rewritten |
| Managers trained in evidence-based feedback | Moderate reduction | Untrained managers take far longer and produce less usable output |
| Ratings linked to pay from cycle one | Increase | Triggers negotiation, escalation and appeals |
| More than seven result areas per role | Increase | Every extra area costs time in every conversation, every cycle |
| No system, spreadsheets only | Increase | Version control and chasing consume administrative hours |
The cost of not running one
The comparison is never against zero. An organisation without a performance system is already paying, in ways that do not appear as a line item:
- Regretted attrition. Replacing a competent mid-level employee costs several months of their salary in recruitment, lost output and ramp time. Retaining one person you would otherwise have lost typically covers a year of running cost.
- Promoter time spent adjudicating. Every promotion argument that has no documented basis is resolved by the founder, repeatedly, at the most expensive hourly rate in the business.
- Misallocated increments. Where increments are decided by impression, some are paid to people whose contribution would not survive examination — and the people who notice first are your best performers.
- Decisions that cannot be defended. The cost is not usually legal. It is the credibility lost when a promotion cannot be explained to the person passed over.
We are not going to attach percentages to these. Any consultancy quoting a precise return on a performance system should be asked how the counterfactual was established, because there is rarely an honest answer.
A rough shape for a 150-person organisation
Not a quote — a structure to fill in with your own numbers.
- Year one: design (one-off, external or internal), plus roughly two working days per manager, plus system licence, plus manager training. The heaviest year by a clear margin.
- Year two: no design cost if roles held stable; manager time falls as the process becomes familiar; licence continues. Typically well under half of year one.
- Year three onward: a steady state of manager time and licence, plus a light annual review of whether the result areas still describe the work.
The pattern that matters: this is front-loaded, and organisations that judge it on year one alone consistently conclude it is not worth it — in the year when the cost is highest and the benefit has not yet arrived.
Frequently asked questions
Can we do the design ourselves and skip the consultant?
Yes, and some organisations do it well. It takes longer, and the common failure is that internal authors write result areas nobody disagrees with — which is comfortable and produces a framework that measures nothing contested. The value in the design workshops is largely in the arguments.
What is the minimum viable version?
Written result areas per role, one honest annual review with evidence, and a mid-year check-in. That is perhaps six to eight hours per manager per year and delivers most of the benefit. Everything beyond it is refinement.
How do we know it is worth continuing?
After two cycles, look at whether promotion decisions are now traceable to a record that predates them, whether managers can evidence their ratings, and whether regretted attrition has changed. If none of the three has moved, something is wrong with execution rather than with the concept — find out which before spending another year.
Does software reduce the cost or add to it?
It reduces administration and evidence-gathering substantially and adds a licence. Net, it usually reduces total cost above roughly eighty people and is not worth it much below that. The thresholds are in HRMS vs spreadsheets.
Where to go from here
If you are budgeting for this, the useful exercise is to take the manager-time table above, apply your own manager count and cost, and compare it with what one regretted resignation costs you. Most promoters find the comparison settles the question quickly.
For what the system does, see the performance management system; for the twelve weeks after sign-off, rolling out KRAs; and for whether you need outside help at all, consultant or software first. Our case study walks through a complete engagement.
BPro is based in Kochi, Kerala and works with organisations across India. Call +91 80869 08876, email care@bpropms.com, or use the contact form.