Designing dashboards around decisions

A dashboard should help a team understand something well enough to decide what to do next.

It is not just a place to display metrics. A useful dashboard connects product questions, workflows, metrics, context, and decisions in a way that helps people interpret evidence.

Dashboards are outputs of measurement design. They should not be the starting point.

Before / after dashboard structure

Practical dashboard prompts

Before adding a chart, ask:

• What question does this chart help answer?
• Which workflow does it describe?
• Which decision could it support?
• What context does the reader need?
• What would someone do differently after seeing it?

Why decision-led dashboards matter

Many dashboards are built around available data.

They show metrics because those metrics exist, not because the team knows what decision they support. Over time, dashboards can become collections of charts that are reviewed out of habit but rarely used to change anything.

A decision-led dashboard starts from a different question:

What does this team need to decide, monitor, investigate, or improve?

That question shapes which metrics belong on the dashboard, how they are grouped, and what context is needed to interpret them.

Start with the decision

Before designing a dashboard, define the decision or set of decisions it should support.

For account registration, the team might need to decide:

• should we simplify the registration form?
• should we improve validation messages?
• should we change the email verification step?
• should we investigate email deliverability?
• should we compare registration performance by device, channel, or entry point?

These decisions point to different metrics and views. A dashboard for improving form completion may need different evidence from a dashboard for investigating email verification.

If the decision is unclear, the dashboard will usually become too broad.

Organise around workflows

A dashboard is easier to use when it reflects the workflow the team is trying to understand.

For registration, the workflow might be:

View registration form
↓
Enter details
↓
Submit form
↓
Verify email
↓
Access product

The dashboard could then show metrics that describe progress through the workflow:

• registration start count
• form submission rate
• form error rate
• email verification rate
• registration completion rate
• time to complete registration

This structure helps the team see where users move forward, slow down, fail, or leave. It also makes the dashboard easier to discuss because the metrics follow the shape of the user behaviour.

Provide enough context to interpret the numbers

A dashboard should not make people guess what a metric means.

For important metrics, include enough context to explain:

• what the metric measures
• which workflow it relates to
• how it is calculated
• which users or sessions are included
• what filters or exclusions apply
• when the metric was last updated
• who owns the definition

This does not mean putting long definitions beside every chart. It means making the meaning accessible where the dashboard is used.

A metric without context may look precise while still being misunderstood.

Show useful comparisons

A single number rarely tells the team enough.

Useful dashboards usually help people compare behaviour across time, workflow steps, user groups, or product changes.

For registration, useful comparisons might include:

• this week compared with last week
• before and after a release
• desktop compared with mobile
• invited users compared with self-serve users
• one entry point compared with another
• form submission compared with email verification

Comparisons help the team see whether a number is normal, changing, concentrated in one area, or linked to a product change.

Make investigation easier

A dashboard should help the team move from signal to investigation.

If registration completion falls, the dashboard should help the team see whether the change is linked to form submission, validation errors, email verification, first product access, or a specific segment.

That does not mean the dashboard needs to answer every question. It should give the team enough structure to know where to look next.

A useful dashboard helps people ask better follow-up questions.

Keep the dashboard focused

A dashboard should not include every available metric.

Include metrics that help the team:

• monitor an important workflow
• diagnose a problem
• evaluate a change
• compare meaningful segments
• understand progress towards an outcome
• decide where to investigate or act

Remove or separate metrics that do not support the dashboard’s purpose.

A focused dashboard is easier to trust, easier to maintain, and easier to use in product discussions.

Common mistakes

Common dashboard mistakes include:

• starting with available data instead of decisions
• mixing unrelated workflows in one view
• showing metrics without definitions
• using charts that do not support action or investigation
• adding more metrics instead of clarifying the question
• keeping unused charts because they already exist
• treating the dashboard as the measurement strategy

These problems are usually signs of weak measurement design, not weak visual design.

How to apply this

When designing or reviewing a dashboard, ask:

• what decision should this dashboard support?
• which workflow does it help the team understand?
• which metrics answer the most important questions?
• what context is needed to interpret those metrics?
• which comparisons or segments matter?
• what should the team investigate if a metric changes?
• which charts can be removed because they do not support use?

Start with the dashboard’s job. Then choose the metrics, layout, and context that support that job.

Related articles

• Designing a measurement framework
• Building a metric tree
• Measurement review checklist

Key takeaway

Design dashboards around decisions, not available data.

A useful dashboard helps teams understand behaviour, interpret evidence, and decide what to investigate, improve, or maintain.