Performance Management Software: Who Decides?

You bought the performance management software. It is configured, the logins work, and the dashboards are empty, because it is still waiting for someone to decide what goes in it.

That is the most common way performance management software fails, and it has nothing to do with the software. A system can store goals, KPIs, check-ins and reviews. It cannot decide what those should be. The order that works is: decide the KPIs from your strategy first, then choose and configure the tool to hold them, and judge any tool on whether it keeps work, evidence and measures connected.

What performance management software does, and what it cannot do

Strip away the feature lists and most performance management software does four jobs.

It stores goals, KPIs, review forms, feedback and notes in one place.
It reminds people when check-ins, reviews and updates are due.
It routes work between people: a self-review to a manager, a calibration to HR, a sign-off to a director.
It reports, showing completion rates, ratings and goal progress across teams.

All four are useful. None of them answers the questions that decide whether performance management works:

What should the company, each department and each person be measured on this year?
Which of those measures tie to the strategy, and which are just habit?
Who owns each one?
What counts as evidence that someone did well?

Those are leadership decisions. Software can hold the answers and make them visible, but if nobody has made the decisions, the system faithfully stores whatever each manager typed in on the first day: vague goals, inconsistent across teams, disconnected from the plan.

Why a new system so often changes nothing

Performance management has a confidence problem that predates any particular tool. In Gallup's 2016 report "More Harm Than Good: The Truth About Performance Reviews", only 2% of CHROs strongly agreed that their performance management system inspired employees to improve.

It is tempting to read that as a software problem and buy something newer. But a system that inspires nobody is usually measuring the wrong things, or measuring the right things without evidence. Moving the same goals and the same annual ratings into a better interface gives you a nicer version of the system nobody believed in.

The software you choose matters. It matters second.

The order of operations

Here is the sequence that gives software something worth holding.

Decide the company KPIs from the strategy. Three to five measures that show whether the year's priorities are being met. How to create KPIs walks through turning the strategy you already have into these.
Connect a North Star. Decide the one company number the others serve, so every department knows what it is ultimately contributing to. See how to pick a North Star metric.
Set department KPIs, with one owner each. Each department owns the part of the company measures it can actually move. Not the company number repeated; its own input to it.
Decide the evidence. What record of work will reviews draw on: completed work linked to priorities, commitments met, peer feedback captured near the moment.
Decide the cadence. How often goals are checked, how often people talk about progress, and when formal reviews happen.
Then evaluate and configure the software against the decisions above.
Pilot with one or two departments for a cycle before rolling out, and fix what the pilot reveals.

Steps one to five can be done in a few working sessions, without any software at all. Doing them first also makes the buying decision easier, because you know exactly what the tool has to hold.

A buyer's checklist for performance management software

Once the decisions are made, this is what to evaluate. Ask to see each one working in a demo account with real-looking data, not described on a slide.

Area	What to ask	A good answer	A warning sign
KPI structure	Can company, department and individual measures be linked, so I can trace one to the next?	The link is visible and each measure shows what it serves	Goals are free text with no connection between levels
Ownership	Does every measure have exactly one owner?	One named person per measure, visible to everyone	Measures owned by a team, or by nobody
Evidence	What do reviews draw on besides the manager's rating?	A record of completed work, commitments and feedback collected through the year	The review is a blank form filled in from memory
Cadence	Does it support regular check-ins, not just the annual review?	Short, frequent check-ins that feed the review	Everything happens in review season
Connection to daily work	Can I see how the work people do each week relates to the goals?	Work is linked to priorities as it happens	Goals live in one place and work lives somewhere else
AI features	Which features propose and which act on their own? Is every AI output logged?	AI drafts, a person decides, and there is an exportable record	"AI-powered" with no answer on what it did or who approved it
Data ownership	Can we export everything, in a usable format?	A full export with a stable structure	Export only by request, or only as screenshots
Admin effort	Who keeps it up to date, and how much of their week does it take?	Most updates come from work people already do	A full-time administrator just to keep it current
Manager adoption	What does a manager do in it each week?	A few minutes, clearly useful to them	A chore done for HR
Employee trust	What can each person see about their own record?	People see their own evidence and can add to it	Records people cannot see until review day

Two rows predict success more than the others. Connection to daily work, because goals that never touch the work are decoration. And employee trust, because a system people do not trust gets filled in defensively, which makes every record in it less useful.

On AI features in particular, how to audit what your AI assistant did lists twelve questions that will tell you quickly whether a vendor can show you what its AI actually did.

Evidence over ratings

The biggest design choice in any performance management system is what a review is based on. Most systems still center on a rating, a number from one to five, entered by a manager once or twice a year. The rating is the output everyone sees, and it is usually produced from memory.

A better system centers on evidence: a record of what someone delivered, against which priority, collected through the year rather than reconstructed in review week. The rating still exists, but it becomes a judgment about a record both people can see, not an impression neither can check. Evidence-based performance reviews covers what to collect and how to run the conversation.

Evidence also depends on cadence. A record that is only updated at review time has the same recency problem as memory. Continuous performance management covers the rhythm of check-ins that keeps the record current.

How Alygn approaches this. Alygn builds the evidence from the work itself. Every task is scored against the company goal, the team goal, the person's role and the person's own KPIs, using the four-pillar method, and drift shows up weekly rather than at review time. The KPIs still have to be decided first; the scoring is only as good as the measures it is scored against.

Questions to settle before your first demo

Bring the answers to these to any demo, and the conversation will be about your company rather than the vendor's feature list.

What are our three to five company KPIs this year, and who owns each?
Does every department have its own measures with one owner, or is that still to do?
What evidence do we want reviews to rest on?
How often do we want people to check progress: weekly, fortnightly, monthly?
What will a manager do in the system each week, and why would they want to?

If questions one and two do not have answers yet, start there. A demo will go much better with them, and so will the rollout.

Frequently asked questions

What is performance management software?

It is software that stores and organizes goals, KPIs, check-ins, feedback and reviews, reminds people when they are due, and reports on progress across the company. It holds the decisions about what people are measured on; it does not make those decisions.

What is the difference between a performance management system and performance management software?

The system is how your company manages performance: what people are measured on, how often progress is discussed, what evidence reviews use, and who decides. The software is the tool that holds and runs that system. A good tool cannot rescue a system nobody has designed.

Should we decide our KPIs before buying performance management software?

Yes. Deciding company and department KPIs first tells you what the software has to hold and makes vendor demos far more useful. If you buy first, the system fills up with whatever each manager types on the first day, and you end up redesigning goals inside a tool that was configured for the old ones.

Can AI decide our KPIs for us?

AI can draft candidate KPIs from your strategy and check them for gaps and conflicts, which saves real time. It should not decide them. Choosing priorities, setting the ambition of each target and naming one owner per measure are leadership decisions, and they need a person's name next to them.

How do we know if our current software is the problem?

Look at what is in it. If the goals are vague, inconsistent between teams and unconnected to the strategy, the problem is the decisions, not the tool, and a new tool will inherit the same goals. If the goals are clear and the tool cannot link them, show the evidence, or fit how your managers work, then the software is worth replacing.

Performance Management Software: Who Decides What Goes In It?