KPIs That Measure Innovation, Not Activity
Most innovation dashboards count effort. Here is what to measure at each stage instead.
Ask an innovation team what they measured last quarter and you will usually get a list like this: workshops held, ideas submitted, startups scouted, pilots launched, people trained.
Every one of those is a real number. None of them tells you whether anything changed.
This is the core problem with innovation reporting. Activity is easy to count and easy to defend. Outcomes are slow, contested and often arrive after the person who started the work has moved to a different role. So teams report what they can prove, the dashboard fills up with green, and two years later somebody senior asks what the innovation function has produced and nobody has a good answer.
The fix is not more metrics. It is metrics matched to the stage the work is at, plus the honesty to report the ones that look bad.
Why activity metrics are seductive and useless
An activity metric measures effort. Ideas in the pipeline, hours of training delivered, number of startups screened, workshops run.
They are seductive because they are always available, always improvable, and always defensible. If your metric is workshops held, you can always hold another workshop.
They are useless because they have no ceiling and no direction. Doubling your idea intake does not double your output. Beyond a point it makes things worse, because a bigger funnel with the same conversion rate means more effort spent screening and more people whose submissions go nowhere.
The one honest use of an activity metric is as a leading indicator early in a programme, when you have nothing else yet and you need to show the machine is running. Report it for two quarters. Then retire it.
Metrics by stage
Innovation work moves through stages, and each stage answers a different question. Measuring all stages with the same yardstick is what produces meaningless dashboards.
Stage 1: Discovery
The question: are we looking in the right places?
- Problem quality, not idea quantity. How many validated problem statements do you have, meaning a problem a named internal owner agrees is real and costly? Ten of these beat five hundred ideas.
- Coverage of strategic priorities. What share of your discovery work maps to the two or three things the company has said matter most? If the answer is under half, your innovation function is running a hobby.
- Time from problem identified to problem owner assigned. Measured in days. This is the first place things die.
Stage 2: Validation
The question: are we killing the wrong things fast enough?
- Kill rate, and time to kill. A healthy portfolio kills most of what it starts. If your kill rate is under half, you are not testing hard enough. If your average time to kill is longer than your average time to launch, your process has no brakes.
- Cost per validated learning. Total validation spend divided by the number of assumptions you tested to a conclusion. It should fall over time as the team gets better at designing cheap tests.
- Assumption clarity. For each active project, is there a written statement of the one assumption that, if false, kills it? Percentage of projects with one. This is a process metric, but unlike activity metrics it directly predicts whether the next stage will work.
Stage 3: Pilot
The question: does it work in our actual environment?
- Pilot-to-decision rate. What share of pilots reach a clear yes or no within the planned window? Pilots that neither succeed nor stop are the most expensive thing in corporate innovation, because they consume budget and attention indefinitely.
- Time to first pilot. From signed intent to first real user or first real transaction. This is the single best proxy for whether your organisation can move. Track the median, and track the worst case, because the worst case is where the structural blockers show.
- Blocker taxonomy. Every delayed pilot gets a reason code: procurement, legal, IT security, data, budget, sponsor availability. After a year you have a ranked list of what is slowing your company down, and that list is more valuable to the board than any individual project.
Stage 4: Scale
The question: did it change the business?
- Revenue, cost or risk impact, stated in euros, attributed conservatively, with the method written down. Conservative attribution that survives scrutiny beats generous attribution that gets picked apart.
- Adoption depth. Not how many licences were bought, but what share of the intended users use it weekly three months after rollout.
- Survival rate at twelve months. What share of scaled initiatives are still running a year later. This is the metric that separates innovation from procurement.
Three metrics worth reporting that nobody wants to
Time to first pilot, by department. Uncomfortable, because it names which parts of the organisation are slow. Also the most actionable number you have.
Sponsor churn. How many active projects have lost their executive sponsor to a role change in the last twelve months. High sponsor churn explains more project deaths than any technical factor, and it is invisible unless you count it.
Ideas from outside the innovation function. If every project originates with the innovation team, you have an innovation department, not an innovative company. The share of initiatives originating in operating units is a direct measure of whether the capability is spreading.
How to present this to a board
Boards do not want a dashboard. They want to know three things: what did you learn, what did you stop, and what are you betting on next.
Structure the report that way.
Lead with the portfolio: how many initiatives at each stage, how many moved forward, how many were killed and why. Killing things is evidence of discipline, so report it as a positive rather than burying it.
Then one page on impact, stated conservatively, with the attribution method visible.
Then the blocker list, ranked, with what you need from the board to remove the top one. This turns the report from a status update into a request, which is what makes it useful.
Keep it to three pages. An innovation report longer than that is usually padding around a thin result.
The one rule
If a metric cannot go down, it is not a metric.
Ideas submitted only goes up. Workshops held only goes up. Startups scouted only goes up. A number that has no failure mode is a number that carries no information.
Kill rate can go wrong in both directions. Time to first pilot can get worse. Twelve-month survival can fall. Those are metrics, because they can tell you something you did not want to hear.
We have been running corporate innovation programmes from our hub in Vienna since 2017, with 45+ corporate partners and seven accelerator batches behind us. Most of what is above comes from watching which programmes survived their third year and which quietly did not.
