KPI Examples That Change Decisions, Not Just Dashboards
A few years ago I sat through a quarterly business review that taught me more about KPI examples than any template ever did. Every chart went up and to the right. Signups were up, and page views were up as well. In fact, total registered accounts had crossed a round number, and someone had put a little trophy icon next to it. Forty minutes later the meeting ended, and yet nobody had changed a single plan. Then, two months after that, our renewal rate dropped hard enough that finance noticed before product did.
That meeting changed how I think about metrics. Now the question I ask is not “is this number healthy?” Instead, I ask “if this number moves, what will we do differently on Monday?” If nobody can answer that, then the metric is decoration.
I lead product analytics, and most of my week is spent helping product managers figure out which numbers deserve their attention. So what follows is the set of KPI examples I keep coming back to, the decision each one is meant to drive, and the ways each one can quietly mislead you. None of this is theory, though. Every one of these has either saved a roadmap or embarrassed me at some point.
The Decision Test I Run on All KPI Examples
Before a metric earns space on a dashboard, I ask the team to fill in one sentence:
“If this number goes below X for Y weeks, then [named person] will [specific action].”
That sentence does a lot of work. Specifically, it forces a threshold, a time window, an owner, and a response. As a result, a surprising number of popular metrics fall apart right there. For example, try it with “total registered users.” What number would make you act? And what would you do? Usually the honest answer is nothing, because a cumulative total almost never goes down.
The guidance from experienced product teams lines up with this. For instance, one useful framing from Uxcel suggests asking yourself, for every metric you track, whether a significant change would tell you exactly what action to take. Similarly, ProductPlan makes the point that good candidates measure something that changes, and those changes need to matter to the business.
I also add one more filter. In short, the metric has to be something a team can influence through its own work within a quarter. Otherwise, if it takes a company reorg to move it, it belongs in a board deck, not a product dashboard.
Why So Many KPI Examples Fail in Practice
Most lists of KPI examples you find online are not wrong. However, they are usually incomplete. While they tell you what to measure, they rarely say what to do when the number moves. Consequently, that gap is where dashboards turn into wallpaper.
There are three failure patterns I see over and over.
Cumulative counts. Think of total users, total downloads, or total revenue to date. Because these only grow, they always look good. Mailchimp, for example, describes vanity metrics as numbers that look good on paper but don’t give organizations the insight they need to make effective decisions.
Lagging outcomes without inputs. Revenue and churn matter enormously. Still, by the time they move, the decisions that caused them happened months ago. Amplitude puts it plainly: metrics like Monthly Recurring Revenue or ARPU tell you what happened in the past rather than predicting future results.
Surrogation. This one is subtle and dangerous. Essentially, it is what happens when a team forgets the strategy and starts serving the number. Harvard Business Review covered this in depth, describing it as the tendency to mentally replace strategy with metrics. Probably the best known case is Wells Fargo, where a cross selling target caused employees to focus on increasing that number rather than deepening the bank’s relationship with its customers.
I have seen smaller versions of this inside product teams, too. Once, a team was given a target for “features shipped per quarter” and hit it every time. Even so, customers did not notice. In other words, the number was real, but it measured motion, not progress.
12 KPI Examples That Actually Change Decisions
Here are the KPI examples I recommend most often. For each one, I explain what it measures, the decision it should trigger, and the trap to watch for. Admittedly, they lean toward B2B software and professional services businesses, because that is where I have spent most of my career. Nevertheless, the logic travels well.
1. Seven Day Activation Rate Tops My KPI Examples for Onboarding
What it measures: This is the share of new accounts that reach your product’s value moment within seven days of signing up. By value moment, I mean the first action that proves the product did something useful, such as sending a first invoice, publishing a first report, or inviting a teammate.
Decision it drives: Where to invest onboarding effort. So if activation drops, the next sprint goes to the first session experience instead of new features.
Watch out for: Defining activation as something easy, like “completed profile.” Instead, pick the action that correlates with retention three months later. After all, if you have not validated that correlation, your activation rate is just a guess.
2. Time to First Value
What it measures: Here you track the median time between signup and the value moment.
Decision it drives: Whether to cut steps from setup. For example, when we shortened this for one product by removing an optional integration step from the default path, retention improved more than any feature we shipped that half.
Watch out for: Averages. Just a handful of accounts that take three months will drag the mean into nonsense. Therefore, use the median and look at the distribution.
3. Week Four Cohort Retention, Core to Retention KPI Examples
What it measures: Of the accounts that started in a given week, this shows how many are still actively using the product four weeks later.
Decision it drives: Whether a recent change helped or hurt. Specifically, cohort views let you compare people who joined before a release against those who joined after. Blended retention, by contrast, hides this completely.
Watch out for: Defining “active” too loosely. After all, a login is not usage. So tie activity to the core action.
4. Feature Adoption Depth
What it measures: Rather than how many accounts tried a feature, this counts how many used it three or more times in a month.
Decision it drives: Whether to keep investing in a feature, fix it, or sunset it. Breadth tells you the launch email worked. Depth, on the other hand, tells you the feature works.
Watch out for: Celebrating launch week numbers. Naturally, every new feature gets a curiosity spike. For that reason, wait a full month before drawing conclusions.
5. Task Success Rate in Usability KPI Examples
What it measures: This is the share of attempts at a key workflow that end in completion without errors or abandonment.
Decision it drives: Where to focus usability work. Notably, this comes from Google’s HEART framework, which covers Happiness, Engagement, Adoption, Retention, and Task Success. In addition, the framework gives a good rule for picking metrics: ratios, percentages, and averages per user are more informative than raw numbers.
Watch out for: Measuring only the happy path. Instead, track where people quit, not just where they finish.
6. Net Revenue Retention Among Revenue KPI Examples
What it measures: Take revenue this period from customers you had a year ago, including expansion and minus churn and downgrades, then divide it by what those same customers paid a year ago.
Decision it drives: Whether to prioritize expansion features (more seats, higher tiers, add ons) or new customer acquisition. If NRR is strong, then your existing base is your growth engine.
Watch out for: Its lag. Because it moves slowly, pair it with leading signals like the next example.
7. License Utilization
What it measures: Simply put, seats in active use divided by seats paid for.
Decision it drives: Which accounts need customer success attention before renewal. For instance, when utilization sits low for two consecutive months, that account is a churn risk, no matter how friendly the champion sounds on calls.
Watch out for: Treating high utilization as the goal by itself. Once it reaches roughly full use, it becomes an expansion signal, and so sales should hear about it.
8. Support Contacts per Active Account
What it measures: Monthly support tickets divided by active accounts.
Decision it drives: Whether a release created confusion, and which areas of the product need clearer design or documentation. In fact, rising tickets after a launch are the fastest feedback loop most teams have.
Watch out for: Pushing it down by making support harder to reach. If the number falls because people gave up, then you have made things worse.
9. Cycle Time in Delivery KPI Examples
What it measures: Days from the moment a team commits to an item until a customer can use it.
Decision it drives: Whether to shrink scope, split work, or remove handoffs. Long cycle times mean slow learning, and unfortunately slow learning compounds.
Watch out for: Gaming it by committing late. To prevent that, define the start point clearly and keep it consistent.
10. Share of Roadmap Items With a Measured Outcome
What it measures: Of the items shipped last quarter, this counts how many had a success metric defined before work started, and how many were checked after launch.
Decision it drives: Essentially, this is a KPI for the product team itself. So when it is low, we stop approving new work without a stated hypothesis.
Watch out for: Turning it into paperwork. Honestly, a one line hypothesis is enough. Ultimately the point is closing the loop, not writing documents.
11. CAC Payback Period
What it measures: Months of gross margin from a new customer needed to recover what it cost to acquire them.
Decision it drives: How hard to push acquisition spend, and which channels or segments to cut. Furthermore, if payback stretches beyond what the business can fund, growth becomes a cash problem.
Watch out for: Blending all channels together. Otherwise, a healthy average can hide one channel that loses money on every deal.
12. Realization Rate in Services KPI Examples
What it measures: For consulting and professional services firms, this is the share of billable hours worked that actually get invoiced and collected at standard rates.
Decision it drives: Pricing, scoping, and which clients or engagement types to pursue. Utilization tells you people are busy. Realization, however, tells you whether that busyness turns into revenue. Surprisingly, many of the services businesses I have advised track the first and ignore the second.
Watch out for: Pushing realization up by refusing reasonable scope changes. That approach protects the margin on this project and then loses you the next one.
Pair KPI Examples So None Can Lie Alone
Any single metric can be gamed, usually without anyone meaning to. Fortunately, the fix is pairing.
For structure, I borrow from the North Star framework. First, you choose one metric that reflects the value customers get. Next, you map the inputs that drive it. Amplitude’s playbook describes a good North Star as one that represents the value users get from your product, sits within product and marketing’s sphere of influence, and leads revenue. Around it, you then set key inputs that teams can directly influence with their day to day work.
One line from that playbook stuck with me. As John Cutler, co author of The North Star Playbook, notes, “If you can move your North Star directly, it’s probably not a good North Star.” Put differently, the North Star tells you where you are heading, while the inputs are what your teams actually push on.
After that, add guardrails. Basically, these are metrics that must not get worse while you chase the main one. Below are some pairings that have worked for us.
- Activation rate, guarded by week four retention. For example, if activation goes up but retention does not follow, we probably made onboarding easier without making the product more valuable.
- Feature adoption depth, guarded by support contacts per account. Clearly, adoption that comes with a wave of tickets is not a win.
- CAC payback, guarded by net revenue retention. Likewise, cheaper customers who leave quickly are not actually cheaper.
- Cycle time, guarded by defect rate after release. Similarly, shipping faster only counts if what ships works.
HBR’s research on surrogation backs this up. Indeed, one of its central recommendations is to loosen the link between metrics and incentives and use multiple metrics. As a result, a pair of numbers is much harder to fool than one.
Write the Decision Down for Your KPI Examples
This is the habit that made the biggest difference for my teams, and moreover it costs almost nothing.
For every KPI on the main dashboard, we keep a short decision log right next to it. Each entry has four parts:
- A threshold that triggers attention.
- Some time window, so one bad week does not set off alarms.
- One named owner who decides.
- A default response, agreed in advance.
For example: “If seven day activation falls below our baseline by more than five points for two straight weeks, the onboarding PM pauses roadmap work and runs a session review with design within five business days.”
Writing it down in advance does two things. First, it removes the debate in the moment, when people are most likely to explain the number away. Second, it quickly shows you which metrics nobody actually cares about. So if a team cannot agree on a default response, the metric probably should not be on the main dashboard.
Google’s HEART authors suggested a similar discipline years ago with their Goals, Signals, Metrics process. In that model, goals are broad objectives, and signals are indicators that your team is making progress toward them. Therefore the metric is the last step, not the first. Sadly, most teams do it backwards.
KPI Examples I Took Off Our Dashboard
When I first applied the decision test to our main product dashboard, it had twenty seven tiles. After the review, however, it had nine. Here is some of what came off, and why.
Total registered users. Since it never goes down, it never triggers anything. Instead, we replaced it with weekly new activated accounts.
Average session duration. Longer sessions sometimes meant people were stuck. In other words, the number could not tell good engagement from frustration.
Page views. These are useful for the marketing site. Inside the product, though, they mean little.
NPS as a headline number. We still survey customers. Even so, the quarterly score moved too slowly and too vaguely to steer anything. In contrast, the written comments were far more useful than the score, so we kept those in a separate review.
Features shipped. As described above, it measured effort, not results.
To be clear, none of these metrics were deleted from our data warehouse. They remain available for investigations. Meanwhile, they just stopped competing for attention in the places where decisions get made. Basedash frames this well, suggesting that every dashboard element should map to an action, and anything that does not is either context or clutter.
Rolling Out New KPI Examples Without a Metrics War
People get attached to their numbers. For instance, a marketing lead who has reported total signups for three years will not love hearing that it is a vanity metric. Here is how I have handled that.
Start with one team, not the whole company. Pick a team that already feels the pain of unclear priorities. Then run the decision test on their metrics, build the decision log, and let results speak.
Do not delete, demote. Move metrics to a secondary view instead of removing them. As a result, resistance drops and the data stays available.
Review thresholds every quarter. Baselines change over time. Consequently, a threshold that made sense at launch might be meaningless a year later. Amplitude’s guidance on North Star work makes the same point: the goal is to build a North Star that’s directionally accurate and then adapt as you learn.
Show one decision that changed. Nothing convinces leadership faster than a story where a metric triggered a response and the response worked. In my experience, one good example beats any slide about measurement philosophy.
Keep humans in the loop. Metrics should start conversations, not end them. So when a number moves, the first step is understanding why. Similarly, MIT Sloan Management Review argued that the problem is rarely measurement itself, and that strong organizations relentlessly focus on defining metrics and KPIs that inspire strategic success. Ultimately, the goal is better metrics, not fewer opinions.
The Short Version
Good KPI examples share a few traits. Above all, they move when something real changes. Each one also has an owner, plus a threshold and a time window. Most importantly, every one comes with a decision that someone agreed to make before the number ever moved.
If you take one thing from this, take the sentence: “If this number goes below X for Y weeks, then [person] will [action].” This week, run every metric on your dashboard through it. Whatever survives is your real set of KPIs. Everything else is context, and that is fine, as long as you stop pretending it steers anything.
In the end, the dashboards I am proudest of are not the prettiest ones. Rather, they are the ones where, every Monday, at least one tile changes what somebody works on.
Frequently Asked Questions About KPI Examples
What are good KPI examples for product managers?
Strong KPI examples for product managers include seven day activation rate, time to first value, cohort retention at week four, feature adoption depth, task success rate, and net revenue retention. However, each one should be paired with a decision and an owner. For a helpful starting point, see Amplitude’s North Star Playbook.
What is the difference between a KPI and a metric?
Every KPI is a metric, but not every metric is a KPI. Specifically, a KPI is a metric tied directly to a business objective and used to make decisions. ProductPlan’s glossary explains how to align each KPI with a larger objective.
How many KPI examples should a product team track?
Most teams do well with one North Star metric, three to five input metrics, and two or three guardrails. Beyond that, attention usually splits. Likewise, the HEART framework guide from Product Compass recommends focusing only on the metrics that match your immediate goals.
What are vanity metrics, and why should I avoid them?
Vanity metrics look impressive but do not guide decisions. Total signups and page views, for example, are common ones. Mailchimp’s guide to vanity metrics explains how to spot them and what to track instead.
How do I stop teams from gaming KPIs?
Use paired metrics and guardrails, and also avoid tying any single number too tightly to bonuses. Harvard Business Review’s article Don’t Let Metrics Undermine Your Business covers this problem, known as surrogation, in detail.
Should revenue be a product KPI?
Revenue matters, but it lags. For that reason, product teams usually do better tracking leading indicators of customer value that predict revenue. Amplitude’s article on good and bad North Star metrics explains why lagging metrics make weak primary targets.
References
- Harris, M. and Tayler, B. Don’t Let Metrics Undermine Your Business. Harvard Business Review, September to October 2019. https://hbr.org/2019/09/dont-let-metrics-undermine-your-business
- Schrage, M. Don’t Let Metrics Critics Undermine Your Business. MIT Sloan Management Review. https://sloanreview.mit.edu
- Amplitude. About the North Star Framework. The North Star Playbook. https://amplitude.com/books/north-star
- Amplitude. Defining Your North Star. The North Star Playbook. https://amplitude.com/books/north-star
- Sholtz, J. What Makes a Good vs Bad North Star Metric. Amplitude Blog. https://amplitude.com/blog
- ProductPlan. Key Performance Indicator (KPI). ProductPlan Glossary. https://www.productplan.com/glossary/
- ProductPlan. HEART Framework. ProductPlan Glossary. https://www.productplan.com/glossary/
- Product Compass. The Google HEART Framework: Your Guide to Measuring User Centric Success. https://www.productcompass.pm
- CleverTap. Using the Google HEART Framework to Set UX Goals. https://clevertap.com
- Mailchimp. Vanity Metrics. Mailchimp Resources. https://mailchimp.com/resources/vanity-metrics/
- Uxcel. Vanity vs Actionable Metrics. Product Analytics Course. https://uxcel.com
- Basedash. Vanity Metrics vs Actionable Metrics: How to Tell the Difference. https://www.basedash.com
- Markovitz Consulting. The Road to Hell Is Paved With Metrics. https://www.markovitzconsulting.com