Performance management ROI is the question most HR teams can't answer, and a new survey puts a number on it. Betterworks asked 934 HR and business leaders in August whether their performance process supports company priorities. Close to 85% of HR leaders said they were "very" or "extremely" confident it does. Only 44% said they could back that with business metrics.

Betterworks calls this "high confidence with low evidence." It is a polite way of saying that most HR teams believe something they can't show.

What the survey actually found

The numbers come from HR Dive's write-up of the Betterworks survey, published October 13. Beyond the headline gap, two details stand out. Just over 30% of HR leaders said they had "some business evidence," which is the survey's way of describing a handful of anecdotes and a slide. And fewer than one in four could point to engagement metrics at all.

That last figure is the strange one. Engagement is the thing HR is supposed to own. If fewer than a quarter of HR leaders can show engagement data connected to their performance process, then the process is running in its own bubble, disconnected from the one metric the department is most often judged on.

Doug Dennerline, the CEO of Betterworks, put the cost plainly: "If organizations can't show how performance improves execution, it becomes much harder to know what's working, where to invest or how to improve."

The timing matters. Most companies are about to start 2027 budget planning. When a CFO asks whether the performance program is worth the hours managers spend on it, "we're very confident" is not going to survive the meeting.

Why confidence and evidence drift apart

It is easy to read the survey as HR being complacent. That is probably not what is going on. The gap has more mundane causes.

The first is what gets measured. Most performance programs report on completion: 94% of reviews submitted on time, 100% of managers trained, calibration finished by December 15. Those are real numbers, and they feel like evidence. But they describe whether the process ran, not whether it changed anything. A company can hit 100% completion and still have the same mediocre execution it had last year.

The second is where the data lives. Goals are in one tool, reviews in another, engagement surveys in a third, and one-on-one notes in a manager's personal doc or nowhere at all. Each piece is fine on its own. Nobody can connect them, so nobody can ask the obvious question: did the teams with better review practices actually hit more of their goals?

The third is that the people running the process see it working up close. An HR business partner watches a struggling manager have a good, honest review conversation and sees a team turn around. That is real. It just doesn't show up in a spreadsheet, and so the confidence stays private.

How to build evidence before 2027

Betterworks' own recommendations are sensible and short. Pick three to five organizational outcomes that the performance process should affect. Stop reporting review completion and start reporting measurable progress on goals. And be honest about correlation versus causation. Here is what that looks like in a company of a few hundred people.

Choose outcomes you already track. Don't invent new metrics. Pick things leadership already looks at: voluntary attrition of strong performers, time for new hires to hit their first goals, percentage of quarterly goals met by team, internal promotion rate. Three is enough. Write them down before the next cycle starts, so nobody can pick favorable metrics afterwards.

Measure goal progress, not goal setting. "Everyone has goals in the system" is a completion metric. "Teams met 71% of key results this quarter, up from 58%" is an outcome. If goals are set and then not revisited until the review, there is nothing to measure. A monthly check-in on progress, even a quick one, gives you a trend line.

Connect engagement to teams. You don't need a 60-question annual survey. A short monthly pulse with the same four or five questions gives you a team-level trend you can lay next to goal progress and attrition. If the teams whose managers run regular one-on-ones also show higher engagement and lower attrition a year later, that is the beginning of an argument. We wrote about why engagement surveys fail when nobody acts on them, and the same logic applies here: data you never connect to anything is just a ritual.

Keep the narrative. The HR Dive piece also cites a 2026 Cornell University report suggesting that narrative feedback can work better than top-down, numbers-heavy approaches. That doesn't conflict with measuring outcomes. Numbers tell you whether something changed. The written feedback tells you why. A review process that keeps only ratings loses the part people actually learn from.

Be careful with causation. If engagement rose the same quarter you rolled out new review templates, it might also be because you hired a good VP or because bonuses landed. Compare teams that adopted a practice with teams that didn't yet. A rough comparison is still more honest than a before-and-after chart for the whole company.

A simple example. Say a 300-person company starts monthly one-on-ones in its engineering and support teams in January, while sales keeps the old quarterly rhythm until summer. By June you can compare three things across those groups: goal completion, pulse scores on "I know what's expected of me," and how many strong performers left. It is not a controlled experiment. It is still far better evidence than anything most HR teams bring to budget season.

The uncomfortable part

Once you start measuring, you might find that parts of your performance process don't do much. Maybe the annual calibration meeting takes 40 hours of senior time and doesn't change anyone's outcome. Maybe the self-assessment form gets copy-pasted every year. That is useful to know. It frees time for the parts that work, which in most companies are frequent manager conversations and clear goals.

It also changes the HR conversation with finance. "Teams that hold monthly one-on-ones had half the regretted attrition" is a budget argument. "We completed 98% of reviews" is a status update.

Where DTPulse fits

We built DTPulse so the pieces live in one place. Goals and key results are tracked by period, with progress updates rather than a single score at the end. 360 reviews collect competency ratings from the manager, peers and the employee alongside written comments, so you keep both the number and the story. Pulse surveys and eNPS give team-level trends over time, and one-on-one meetings keep agendas and agreements next to the employee's profile instead of in someone's notebook. Exit surveys add the last data point: why people actually left.

None of this proves your performance process works by itself. It does mean that when someone asks whether teams with regular one-on-ones hit more goals, the data to answer is already in one system. If you are looking at the review side specifically, our guide to 360 reviews people don't dread covers how to set them up.

Before the 2027 budget meeting, write down three outcomes your performance process is supposed to move, and find out whether you can pull the last four quarters of each. If you can't, that is the first project for next year, and it is a much smaller one than redesigning the review form.