You have seen the dashboard.
Twelve accounts. Eight green dots. Three yellow. One red.
Six CSAT scores above 4.5. Task completion trending up. Fourteen new accounts onboarded this quarter, and the chart has a nice upward slope.
The dashboard looks healthy. The numbers are going in the right direction.
And then a customer you onboarded nine months ago – green dots all the way through – sends the cancellation email.
What did you miss?
Not the task list. They finished the task list. Not the satisfaction survey. They gave you a five.
What you missed is that nothing on that dashboard measures whether the customer actually reached value, whether the process was easy or just obediently followed, or whether the signal you were tracking was the one that predicts the outcome.
Most onboarding dashboards are full of metrics that feel productive to track and tell you nothing about whether a customer will still be here next year.
Let’s fix that.
The vanity metric problem in onboarding
Vanity metrics are numbers that move and tell you nothing. They are not wrong. They are just irrelevant to the outcome you care about.
In onboarding, the most common vanity metrics are:
Number of tasks completed. A customer can complete every task you assigned and still never reach value – because the tasks were about your process, not their outcome.
CSAT after kickoff. Everyone is happy at kickoff. The plan looks good. The product is full of promise. A CSAT collected at the moment of maximum optimism tells you nothing about whether the customer will be happy in month three, when the hard parts have happened and the promises either held or didn’t.
Number of accounts onboarded. Volume tells you the sales team is working. It says nothing about whether onboarding is.
Time to complete training. A customer who finishes training in three days and never logs in again is not a better outcome than one who takes two weeks and uses the product daily afterward. Speed without activation is a process metric dressed as a success metric.
Average time-to-value. Averages lie in onboarding. One complex enterprise deployment that takes six months can lift the average enough to hide fifty SMB customers onboarding in two weeks. You are not measuring the typical experience. You are measuring the one outlier.
The pattern is the same across all of these: they measure what is easy to count, not what predicts the outcome.
The fix is not to stop tracking them. It is to track them alongside metrics that answer the harder question: did this customer reach the outcome they paid for, and will they still be here next year?
The six metrics that actually predict outcomes
If you instrument nothing else, instrument these six.
| Metric | What it measures | Why it predicts the outcome |
|---|---|---|
| Time-to-value (TTV) | Calendar days from sign to value milestone | The strongest single predictor of retention. Every extra week of TTV raises churn probability. |
| Activation rate | % of customers who complete the first core action | If they don’t activate, nothing else matters. Low activation kills retention faster than any other factor. |
| Onboarding completion rate | % of customers who finish all milestones in the target window | A low completion rate is an early churn signal. Completion lags, which means you can see it forming. |
| Cycle time per phase | Days spent in each onboarding phase | Shows you exactly where the bottleneck lives. Most teams guess wrong. |
| Early churn rate | % of customers who cancel during or within 90 days of onboarding | The outcome you are trying to prevent. Track it monthly, by cohort, and watch the trend. |
| Customer effort score (CES) | “How easy was it to get started?” (1–7 scale) | Effort predicts churn more reliably than satisfaction. A customer who found it hard but enjoyable is more likely to leave than one who found it merely adequate and effortless. |
Let’s walk through each one.
Time-to-value (TTV)
TTV is the headline metric. It answers the question: how long does it take a customer to get what they paid for?
The definition is deceptively simple. Pick one value milestone – the exact moment the customer achieves the outcome they bought – and measure the median calendar days from contract signed to that milestone.
The hard part is picking the right milestone.
It is not “logged in.” It is not “completed training.” It is not “attended the kickoff call.” It is the specific action the customer paid for. If your product is a reporting tool, it is “ran their first real report with their own data.” If it is a collaboration platform, it is “created a project with three teammates and shared it externally.” If it is an onboarding platform, it is “sent their first customer-facing Space.”
Define it once. Instrument it once. Track it per account.
Then measure median, not average. The median tells you what the typical customer experiences. The average tells you what the average of your outliers is. Those are different numbers, and the median is the one that drives decisions.
Measure time-to-first-value (TTFV) separately. TTFV is the smaller, earlier milestone – the first thing the customer does that looks like value. First report exported, first integration connected, first teammate invited. TTFV predicts whether the customer will stay engaged long enough to reach full TTV. A customer with fast TTFV and slow full TTV has a process problem. A customer with slow TTFV has an activation problem. The fix is different in each case.
If TTV is consistently slow, measure time-to-first-value separately. The fix for slow TTFV (activation) is different from the fix for slow full TTV (process bottlenecks), and treating them as one number hides which one you actually have.
Activation rate
Activation rate is the percentage of new customers who complete a defined core action within a set timeframe.
This is the gate. If a customer does not activate, they do not onboard. They go through your process and never arrive.
The activation milestone has to be product-specific. “Logged in” is too early – it measures curiosity, not commitment. “Became a power user” is too late – by the time you measure it, the onboarding window is closed.
The right milestone sits in between: the first action that correlates with long-term retention. In a data analytics product, it might be “ran a query against their own database.” In a CRM, “imported their contacts and created a pipeline.” In an onboarding platform, “sent their first customer-facing Space.”
Find it by looking backward. Take the customers who are still with you after twelve months. What did they all do in their first week? That is your activation milestone.
Track activation rate per cohort and per segment. A blended activation rate of 70% that hides SMB activating at 85% and enterprise activating at 35% is not a 70% success story. It is two different problems, and one of them is urgent.
Onboarding completion rate
Completion rate is the share of customers who finish every milestone in the onboarding plan within the target window.
Simple to define. Harder to instrument well, because “complete” has to mean something. A customer who skipped the training module, never attended a Q&A, and marked the last task done themselves is not complete. They are done with your process.
Define completion as: every required milestone completed, plus product activation achieved, within the target timeframe. Set the timeframe from your own historical data. If your median customer takes 45 days to complete onboarding, set a target of 45 days, track how many hit it, and try to make that number go up.
A completion rate below 60% usually means the process is the problem, not the customer. The milestones are too many, too large, or too disconnected from the outcome the customer cares about.
Before you can measure completion you need a written list of what has to be true for an account to count as complete. Our customer onboarding checklist is a 38-item starting point, with an owner marked against each item.
Cycle time per phase
Total onboarding time is useful. Time broken down by phase is actionable.
Split your onboarding into phases – kickoff, technical setup, configuration, training, go-live – and measure the median days spent in each.
The phase where accounts consistently sit longest is your bottleneck. It is rarely the phase you think.
Most teams guess that technical setup is the long pole. In practice, it is often configuration or training – the phases that depend on customer decisions, not IT work. The customer has to choose how to configure something, or find time for training across a team, and the project stalls while they do.
Cycle time per phase tells you exactly where to focus. If kickoff-to-setup takes three days and setup-to-configuration takes eighteen, you know where the eighteen days are and you can design the delay out.
Early churn rate
The outcome metric. The percentage of customers who cancel during onboarding or within 90 days of completing it.
Track it monthly, by cohort. “Cohort” means the group of customers who started onboarding in the same month. This matters because a blended churn rate across all cohorts hides whether your process is improving or getting worse. If the January cohort had 12% early churn and the March cohort had 6%, something changed – and you want to know what it was.
Early churn is expensive in a way that later churn is not. A customer who cancels during onboarding cost you full acquisition cost and full onboarding cost with zero lifetime value. A customer who cancels in year three at least paid something back. The economics of early churn make it the single most expensive metric on this list, and the one worth the most effort to move.
Customer effort score (CES)
Most teams measure satisfaction. Smart teams measure effort.
CES is a single-question survey asked immediately after key onboarding tasks: “How easy was it to get started today?” on a 1–7 scale.
The insight is not the number. It is what the number predicts. High effort during onboarding is a stronger predictor of churn than low satisfaction. A customer who found your product enjoyable but difficult is more likely to leave than one who found it merely adequate and effortless.
Effort is also more actionable than satisfaction. “The customer is not satisfied” is ambiguous. “The customer rated the intake form a 2 out of 7 for ease” is a specific problem with a specific fix.
Ask CES after three moments: the first task completed, the halfway point of onboarding, and go-live. Track the trend per customer. A CES that starts at 6, drops to 4, and finishes at 3 is telling you something the task list cannot: the process is wearing the customer down.
Leading indicators versus lagging indicators
The single most important distinction in onboarding measurement:
Leading indicators move before the outcome is locked in. They tell you what is about to happen. Workspace engagement, response cadence, time-to-first-value, CES during onboarding – these all shift weeks or months before the renewal conversation.
Lagging indicators tell you what already happened. Churn rate, NPS at 90 days, expansion revenue, lifetime value – these are results, not signals. By the time a lagging indicator turns red, the customer is already gone or going.
A good onboarding dashboard is built around leading indicators. You track lagging indicators too – they tell you whether the whole system is working – but they are not where you spend your weekly attention. A weekly meeting that starts with the churn report is diagnosing a patient who has already left the building.
For more on the leading indicators that catch stalls early, see how to track customer engagement during onboarding.
Building the scorecard
A scorecard is not a dashboard with forty widgets. It is four to six metrics your team reviews every week and acts on.
Here is the shape of one that works:
| Metric | Frequency | Action when off-track |
|---|---|---|
| Time-to-value (median, per segment) | Weekly | Review the phase where accounts are slowest. Redesign that phase. |
| Activation rate (per cohort) | Weekly | If dropping, check whether the first post-kickoff tasks are too large. Make the first one smaller. |
| CES trend per account | After each key milestone | A dropping CES triggers a CSM conversation, not an automated survey. Ask what is hard. |
| Workspace engagement – accounts with zero activity in 7 days | Weekly | Flagged for human review. One small concrete next step, not a general check-in. |
| Onboarding completion rate (per cohort) | Monthly | Trend review. Is the target window realistic? Are the milestones the right ones? |
| Early churn rate (per cohort) | Monthly | If trending up, review the last three churned accounts for patterns. Look at CES, TTV, and engagement in the weeks before cancellation. |
The last row is where teams get stuck. They review completion against a target window nobody ever derived, so “off-track” means off a number somebody once liked the sound of. How long should customer onboarding take covers where that window should come from.
Rules for a scorecard that survives contact with reality
Define every metric in writing. “Active customer” means what, exactly? “Onboarding complete” means what? If two people on your team give different answers, the metric is noise. Write the definition down. Share it. Enforce it.
Segment everything. Track every metric by customer size, use case, and industry. An activation rate of 80% that hides enterprise activating at 40% is a dashboard that lies. Segmentation turns a comforting number into a useful one.
Tie every metric to an action. A metric without an owner and a next step is decoration. Before adding a number to the scorecard, answer: who acts on this, and what do they do when it moves?
Review weekly, not monthly. A month is too long. A stall that started on the 3rd of the month is a churn risk by the 17th. If you wait until the monthly review on the 30th, you are discovering it four weeks late. Fifteen minutes, once a week, looking at the leading indicators that move first.
Do not add metrics because they are easy to get. The number of metrics on your scorecard correlates negatively with how many you actually act on. Start with the six above. Only add a seventh when you have deleted one that was not earning its place.
Where the data comes from
The reason most teams track task completion instead of these six metrics is not ignorance. It is access.
Task completion is the only number that is easy to get when your onboarding runs across a spreadsheet, email, a project tool the customer cannot see, and a product analytics tool that nobody opens per account.
The data for the six metrics that matter lives in different places. TTV requires knowing when the value milestone happened, which means connecting process progress to product usage. Activation rate requires product telemetry. CES requires surveying at the right moments. Cycle time per phase requires knowing when each phase started and ended per account.
A shared onboarding workspace collapses most of this into one place. The customer opens their Space to complete tasks, view content, submit forms and follow the plan. The platform records timestamps, task states, form interactions and engagement per account. TTV, cycle time, completion rate and engagement data become a side effect of the customer doing their work in the tool. For the workspace-versus-project-tool decision, see our comparison of onboarding workspaces and project tools.
Valuecase, for instance, surfaces time-to-value, task progress, engagement and completion data per Space in a cross-customer dashboard that shows which accounts are on track, which are stalling, and where the bottlenecks are across the portfolio.
If you are running onboarding across a patchwork of tools, start by instrumenting the one metric that is easiest to get with what you have – probably activation rate or early churn – and add the others as you consolidate tooling. The scorecard does not need to be perfect on day one. It needs to exist.
What to do tomorrow morning
Open your current onboarding dashboard. Look at it honestly.
How many of the numbers on it are vanity metrics – things that move and tell you nothing about whether a customer reached value?
How many are lagging indicators – telling you what already happened, not what is about to?
How many have a written definition that everyone on the team agrees on?
If the answer to that last one is “none,” start there. Pick three metrics from the list above. Define them in writing. Instrument them, even if the instrumentation is a manual spreadsheet for the first month. Review them weekly.
The tooling question is real, and it matters, and it is not the thing that is stopping you. The thing that is stopping you is that measuring what predicts the outcome is harder than measuring what is easy to count, and nobody has made the hard choice yet.
Make it tomorrow morning.
Once time-to-value is defined and instrumented, the follow-on question is how to move it. How to reduce time to value in customer onboarding works through where the delay actually sits and how to design it out.