Cross-highlight or cross-filter? Choosing the right interaction for each Power BI visual

Filter, highlight or none. Each has a job. Choosing deliberately, visual by visual, is what makes a report answer questions instead of raising them.

Use cross-filter when the reader wants to focus on a selection, cross-highlight when they need to see the selection against the whole, and no interaction when a visual should stay fixed as a reference point. The right choice depends on the question each visual answers, not on a rule applied to the whole report.

Power BI gives every visual three ways to respond when a reader clicks something elsewhere on the page. Most reports leave them on the default. The better reports choose.

Cross-filter: focus on the selection

Cross-filtering removes everything unrelated to the selection. Click the North region and the other visuals show North's figures only, rescaled to fit.

It is the right choice when the reader's next question is about the selection itself. What is driving North's sales? Which products sell there? Which customers are behind the number? Filtering answers those questions cleanly, because nothing else is competing for attention.

It is also the only option for line charts, scatter charts and maps, which cannot be cross-highlighted.

Cross-highlight: the part against the whole

Cross-highlighting keeps all the data visible and emphasises the part that relates to the selection. Click the North region and a product chart still shows total sales for every product, with North's share of each bar highlighted.

It is the right choice when proportion is the point. Is North's mix of products the same as everyone else's? Is this customer segment over-represented in one category? Highlighting lets the reader see the answer without holding two views in their head.

The cost is noise. Faded bars still draw the eye, and a small highlighted segment at the bottom of a tall bar can be hard to read. Use it where the context of the whole genuinely helps, not by habit.

None: the fixed reference point

Setting an interaction to none means the visual ignores selections elsewhere on the page. It is the most underused of the three.

It suits visuals that act as an anchor:

• Headline totals the reader should always be able to see, such as company revenue next to a regional breakdown.
• A target or benchmark visual that makes no sense when filtered.
• Navigation elements and text that should not change.
• A visual that is expensive to query and adds nothing when filtered.

A fixed reference point is often what makes a selection meaningful. North's revenue means more when the company total is sitting beside it, unchanged.

How to decide, visual by visual

For each visual, ask what the reader is trying to find out when they click something else on the page.

• If they want to know more about the selection, filter.
• If they want to know how the selection compares to the whole, highlight.
• If the visual is there as a reference, set none.

Then be consistent. Readers learn how a report behaves from the first few clicks. If similar visuals behave differently for no clear reason, they stop trusting what they see. That matters more than which option you choose.

A sensible starting point for most reports is to make cross-filter the default, then change individual visuals to highlight or none where the question calls for it.

Interactions are part of the design

Interactions are easy to treat as a technical setting. They are really a design decision about what question each visual answers and how readers move from one question to the next. The same thinking applies to layout and purpose, which we covered in three jobs, one dashboard.

If your reports look right but readers still struggle to use them, our Power BI team can review how they behave, and a free analytics audit is an easy way to start.

Choose the interaction the question needs. Then keep it consistent.