What to measure after launch: the few numbers that matter

Start with the action that means value

Every product has one action that shows a customer got what they came for: an order placed, a booking made, a report generated, a message sent. Count that action each week.

Visits and downloads show how many people looked. This number shows how many were served.

Measure the journey as steps

Break the route to that action into steps: arriving, signing up, doing the key action once, doing it again. Count the people at each step.

The step with the largest drop is where to work first. Improving a later step helps little if most people never reach it.

Do they come back?

Of the people who started in a given week, how many are still active a week later, and a month later? A product that people try once and leave grows only for as long as you keep paying to bring in new people.

Group users by the week they started. Then you can see whether a change you made helped the people who arrived after it.

Numbers that mislead

• Total registered users, which can only go up.
• Page views, without knowing what people did.
• Downloads, for an app nobody opens twice.
• Averages that hide two very different groups.

Setting it up

• Decide the events before launch: sign-up, the key action, payment.
• Name them consistently.
• Use one analytics tool, and check it is recording in the released version.
• Collect only what you need, and ask for consent where the law requires it.
• Compare the totals with a source you trust, such as the orders table.

Numbers and conversations

Numbers show where people leave. They do not show why. A few conversations with people who left will tell you more about the reason than another chart.

Review on a schedule

Look at the same short list every week. Decide one thing to change, write down what you expect to happen, and check the following week whether the number moved.