Adoption: Logins Are Not Value

Someone can be in your product every day and get nothing out of it.

That sentence takes new CSMs the longest to believe. The dashboard is green. Weekly active users climb. The champion answers emails within the hour. Then the renewal conversation arrives and the customer says, politely, that they never quite got what they were hoping for. Nobody lied to anyone. The team was measuring the wrong thing with real precision.

Adoption is the bridge between software being deployed and the business getting something out of it. Deployment means the product is installed, configured and switched on. An outcome is the result the customer bought it for: fewer security incidents, a faster monthly close, a shorter support queue. Adoption is the behavior in between. TSIA puts adoption at the center of the Customer Success mandate, and the practical reason is that it's the only link in the chain a supplier can watch happening in something close to real time. Good onboarding hands it a running start, because onboarding ends at first verified value, not at go-live.

Define adoption, or your analytics tool will define it for you

The word has no universal definition. Every tool that measures adoption will cheerfully supply one anyway.

The textbook formula is unglamorous: the eligible users or accounts that meet an explicit adopted criterion, divided by all eligible users or accounts, times a hundred. The trouble sits in two words. "Eligible" is a judgment somebody has to make and write down. "Adopted criterion" is where most teams quietly hand the decision to the product-analytics tool, which usually answers with a login or a session count. Pendo and the other product-analytics vendors define adoption through explicit activity-based criteria that change with context, and they're right to. There's no portable answer. The definition running in your instance is whatever someone configured, and the default was configured by a vendor who has never met your customers.

The denominator deserves the same scrutiny, which is why I keep telling teams to write the denominator down. Adoption can mean users, accounts, seats, workloads, transactions or features. Those are six different numbers from the same week of telemetry. A team that reports 80% adoption without naming the denominator is reporting a mood.

One cousin worth separating, because the terms get swapped in meetings. Activation is the early, narrow measure: the share of new users or accounts that reach one defined milestone. Adoption is the habitual measure, value-relevant use that persists. Both are worth tracking only if the event you chose has some logical or observed connection to value. An activation event nobody validated is a number with a progress bar attached.

Build the definition backwards

The useful chain runs in the opposite direction from the way most dashboards get built. Start from the outcome the customer named. Ask which workflow produces that outcome. Then ask which behaviors inside the product you can actually observe that tell you the workflow is running. Outcome, then workflow, then behaviors. Build it the other way and you end up with a usage report looking for a purpose.

What that looks like in practice. An analytics platform shouldn't celebrate logins when nobody is publishing analyses, connecting trusted data sources, or carrying outputs into actual decisions. A security platform shouldn't equate activated licenses with policy coverage; a license is permission to be protected, not protection. A collaboration tool has to separate messages sent from cross-team workflows that finished. In all three, the observable behavior is a proxy, and the only question worth asking about a proxy is whether it tracks the thing you cannot see.

Breadth, depth, frequency, quality

With a behavior list in hand, a CSM has four dimensions to work on. Breadth is how far adoption has spread across eligible users, teams or business units. Depth is how many of the valuable workflows are genuinely in use rather than merely available. Frequency is whether the rhythm of use fits the use case. Quality is what the behavior produced.

My opinion, and it costs me arguments: frequency is the most over-weighted of the four. Daily use is a property of some products and not others. A quarterly compliance workflow run once a quarter is fully adopted, and a team that marks it red because logins are sparse has confused its own engagement model with the customer's work. Depth and quality carry more diagnostic weight and get less attention, mostly because they are harder to pull from a dashboard.

The pattern to fear is the busy, empty account: wide breadth, high frequency, no depth, nothing produced. High activity can sit next to zero recognized business benefit for a long time. It is the account that looks healthiest in the portfolio review, because every metric anyone bothered to chart is pointing up.

The right adoption metric moves

Adoption is not one number held constant for the life of an account. Gainsight's adoption guidance separates product-level usage from adoption of critical features and makes the point that the relevant metric changes across the lifecycle, which matches how the work actually feels. Early on, the question is whether the first workflow ran end to end. A few months in, whether it spread to the people who were supposed to have it. Before a renewal, whether any of that behavior produced the outcome written in the success plan.

The metric should also follow the revenue model. Snowflake runs a consumption business and reported 125% net revenue retention at January 31, 2026; once a customer has adopted, its stated focus is additional workloads and more consumption. Seat utilization would tell you almost nothing there. In a per-seat product, workload growth is not a number you have. Adoption measurement that ignores how the customer pays, and how value accumulates, ends up precise and irrelevant.

Inside a health score, adoption usually carries one of the heavier weights, something like 30% in a starting model, and those weights should shift by lifecycle and segment. Worth remembering when you watch a score move: adoption is one input among several, and the score itself is a decision model, not a fact.

Low adoption is a symptom

When the number is bad, resist the urge to prescribe. Low adoption has at least four common causes and they look identical on a chart. The workflow may not fit how the team actually works. The people who would run it may never have been enabled. The data it depends on may not be flowing. The administrator who knew how to operate the thing may have left in March. Each of those has a different fix, and generic "let's get you back on track" outreach addresses none of them. Red is a smoke alarm, not a diagnosis.

Diagnosis here means talking to the people who are supposed to be doing the work, not only the sponsor who signed. Barriers live at the workflow level and they are usually mundane: a step that takes eleven clicks, a permission nobody granted, a report the team does not trust yet.

Where adoption sits

Adoption sits in the middle of the chain this series keeps returning to: promise, adoption, outcome, value, retention, growth. It is not the finish line, and treating it as one is how usage reports end up standing in for value. What makes it worth this much attention is timing. Adoption is the earliest honest evidence available about whether the promise is on its way to becoming an outcome, which means a team that defines it carelessly has traded away its only early warning. Measuring the wrong behavior looks exactly like measuring the right one, right up until the renewal.