Too Many Metrics Is the Same as None

A healthcare technology leadership team brainstormed the measures it needed and covered a board. Five plain questions per metric showed how few it could actually produce. A second organization shows what the alternative looks like once it is running: a few measures, read in the room where decisions get made.

07 SIGNAL
The condition

Asked what to measure, a leadership team fills a whiteboard with metrics, many of which it has never measured and cannot say how it would.

A destination, not a reading

The executive leadership team of a healthcare technology company sat in a loose semicircle in a half-empty conference room. Some of them met our eyes. Several kept their gaze on the floor. Many had come from companies where a measurement system already existed. Here, none of it did yet, and the frustration was visible. They had agreed they wanted to improve how work moved through the company, and building the measurement underneath that ambition was starting to feel overwhelming.

We had just walked them through four categories of metrics we typically start with, and asked what outcomes they wanted for the year. The finance leader answered without hesitation: a recurring revenue figure, two retention rates, and a profitability milestone, all by year end.

Those were achievable goals, we said. But they described where the company needed to be at the end of the year. They would not tell anyone, halfway through, whether it was on track to get there. What the team needed were measures that gave continual insight into progress toward those outcomes. Without them, a team that can see only its year-end destination cannot tell whether it is heading toward it. And because what you measure drives behavior, measuring nothing in between left everyone steering by instinct.

A board that filled in minutes

We handed out sticky notes and asked the leaders to brainstorm the measures they thought would help them reach those annual targets, one category at a time.

Value and outcome measures came first, the ones that answer whether the work is making the expected impact: revenue by product, gross margin, client count, cost of goods sold by product line. Flow and capacity measures came next, the ones that show how quickly and predictably demand turns into finished outcomes. One leader wrote as fast as he could, as if afraid of forgetting something: throughput, work in progress, cycle time. Others added plan to done, sales pipeline, operating expenses. Quality and risk measures followed: defect rate, days receivables outstanding, a risk register, a client health score. Last came sustainability and growth, the measures that would show whether the company could keep its pace over time: employee turnover, engagement, happiness.

When the leaders stepped back, the board was overflowing. One of the executives said it plainly. This was far too much, and some of these they had never even measured before.

Figure 1
What went on the board
Question it asksExamples
Value and outcomeIs the work making an impact?Revenue by product, gross margin, client count
Flow and capacityHow fast and predictable?Throughput, cycle time, plan to done
Quality and riskCan delivery be trusted?Defect rate, risk register, client health
SustainabilityCan the pace last?Turnover, engagement, happiness
Every category filled quickly, and most notes were reasonable on their own. Together they were a list no one could produce, let alone read.

Five questions per metric

We opened a blank spreadsheet, put it on the screen, and added one row per metric. For each row we asked the room the same five questions. Do we have it? Have we ever measured it before? Where would the data come from? How exactly do we calculate it? Who owns it?

The room grew restless within a few rows. The spreadsheet turned out to be extremely hard to fill in. The leaders agreed that many of the metrics were genuinely needed. For most of them, though, nobody could say how the data would be gathered or how the number would be calculated.

That was the turning point. The brainstorm had produced a list of things worth knowing. The five questions showed how much of it the company could not yet know, and they made that gap visible to the people who would have to close it. A metric with no data source, no agreed calculation, and no owner is not a measure. It is an intention. Putting it on a dashboard would only add one more number nobody trusted to the ones they already ignored.

Figure 2
The five questions every metric had to pass
  1. Do we have it?
  2. Ever measured it?
  3. Where is the data?
  4. How is it calculated?
  5. Then it counts Who owns it?
Many metrics on the board could not yet get through these five questions. The ones that pass all five are the ones worth putting in front of leaders.
Two people working at a wall densely covered in notes.

Taken on as a quarterly goal

One of the executives ended the exercise by taking it on herself. She would make it one of her goals for the quarter and set up a small team to build the measurement process properly. She formed the team and asked us to take part.

The order of work that session established is the part worth carrying forward. First decide what the company needs to know to judge progress toward its outcomes. Then prove, metric by metric, that it can be sourced, calculated, and owned, and give each survivor a baseline, a target, a timeframe, a threshold for concern, and a named owner. Only then does a number belong in front of leaders. The list gets shorter at every step, and that is the point.

What a short list looks like running

At a different organization, a global consumer goods enterprise working across six regions, the same discipline showed up in a more mature form. Once its regions and global function were deciding together in a shared forum, leaders needed to know whether the new way of working was holding. They deliberately chose a few lightweight measures, balancing the need for oversight against the risk of burdensome reporting.

There were three. Cycle time, meaning how quickly initiatives moved from idea to pilot or release. Cross-region collaboration, seen in initiatives that were co-owned or workstreams that had been merged. And early signs of return on investment, measured against both planned and actual outcomes. The forum reviewed them regularly, so it could spot a trend such as rising cycle times or falling collaboration and act before it became critical.

Each of the three answered a question the forum was actually responsible for, and each was read by the people who could act on it. That is what a measure is for. A wall of numbers gives attention nowhere to land. Three numbers reviewed in the right room give it a place to start.

Figure 3
A few measures, each with a job
TracksWatch for
Cycle timeIdea to pilot or releaseRising cycle times
CollaborationCo-owned or merged workDeclining collaboration
ReturnEarly return indicatorsPlanned vs. actual outcomes
At the six-region enterprise, three lightweight measures were reviewed in the forum that owned the decisions they informed. This is a separate organization from the board above.

A metric nobody can source, calculate, or own is not a measure yet. A short list that someone reviews and acts on is worth more than a long one nobody can produce.

Share With Your Team

Choose who it is for. You send it yourself from your own email; Agileon never writes to your colleagues.

Not sure where to start?

Answer five questions about where work stalls. We send back a starting plan: where to begin and what to look at first.