← All Articles

Outcomes, not outputs: how to measure whether a digital product is working

· agileminds

Product & Experience DesignMeasurement

Most digital programmes can tell you what they have shipped. Far fewer can tell you whether it is working. The release happened, the features are live, the steering group is satisfied, and then somebody asks a simple question: are customers better off?

That gap is the difference between outputs and outcomes. An output is something you deliver: a new app, a redesigned checkout, a dashboard. An outcome is a change in what people do or experience because of it. Outputs are easy to count. Outcomes are what the business case was actually promising.

Start with the customer, not the feature list

Before anyone writes a backlog, it helps to answer three plain questions. Who is this for? What are they trying to get done? What would be different for them if we got it right?

The answers rarely come from a workshop room. They come from talking to real customers, watching them use what exists today and mapping where they get stuck. Even a handful of conversations will usually surface a problem the internal team had not framed in quite that way.

Choose two or three measures, not twenty

A long list of metrics is a way of avoiding a decision. Pick two or three customer outcome measures that tell you whether the product is doing its job. For most digital products, the useful ones come from this short list:

  • Adoption: are the people you built it for actually using it, and are they coming back?
  • Task completion: can customers finish the thing they came to do, without help and without giving up?
  • Satisfaction: how do they feel about the experience, asked at the moment it matters rather than months later?
  • Time to value: how long does it take a new user to get the first real benefit?

You do not need all four. You need the ones that match the problem you set out to solve. A tool for internal staff may care most about task completion and time to value. A customer-facing service may care most about adoption and satisfaction.

Write them into the business case

This is the step that most often gets skipped, and it matters most. If the outcome measures are agreed in the business case, before funding is released, three useful things happen.

First, everyone shares one definition of success. Second, the team can make trade-offs against it: when scope has to give, you protect the thing that moves the measure. Third, you have a baseline. You cannot show an improvement in task completion if nobody measured it before you started.

Be honest about what you do not yet know. It is fine to say the target will be set after the first round of testing. It is not fine to leave the measure undefined and hope it becomes obvious later.

Test with real users early and often

The cheapest time to find out you are wrong is before you have built much. A prototype in front of real customers, early, will tell you more than months of internal debate. It does not need to be polished. It needs to be realistic enough that people react to it as they would to the real thing.

Then keep going. Each sprint demo can show customer evidence alongside the finished work: what people did, what they said, where they hesitated. A demo that shows only completed features invites a conversation about effort. A demo that shows customer behaviour invites a conversation about value.

Releasing early to a small pilot group helps too. You get real usage data on a small scale, and the chance to fix what you find before everyone sees it.

Keep measuring after launch

Launch is the point at which measurement becomes possible, not the point at which it stops. Set a review date, look at the same measures you agreed at the start and decide what to do next: improve, extend or stop. Stopping something that is not working is a perfectly good outcome if it frees money for something that is.

It also helps to give leaders a simple, shared view of those measures alongside delivery status, so the conversation is about results rather than activity.

A practical checklist

  • Name the customer and the problem in a sentence each.
  • Agree two or three outcome measures, and record the baseline.
  • Write them into the business case before funding is released.
  • Put a prototype in front of real customers early.
  • Show customer evidence in every demo.
  • Release to a pilot group before a full launch.
  • Review the measures after launch and decide what happens next.

This is how we approach every engagement: start with your customers, agree the outcome before the work begins and measure it after launch. You can read more about how we work, or see how our Product & Experience Design team puts it into practice.

If you have a product that has shipped but you are not sure it is working, or one that has not started and you want to get the measures right, Talk to us.