Applying Themes and Conditional Formatting

This episode focuses on a practical PL-300 objective: Applying Themes and Conditional Formatting. Themes provide consistent report-wide styling, while conditional formatting uses data values to change presentation when particular conditions or exceptions need emphasis. The discu…

A sales report uses blue for strong performance on one page, green on another, and red on a third. Some ordinary values are bright orange, while one genuinely important exception is gray. Every visual renders successfully, and every number may be correct.

But is the report correctly designed? The evidence isn’t that Power BI accepted the formatting. The evidence is that a report consumer can interpret the same visual signal consistently, under normal and unusual data conditions, without having to relearn the color language on every page.

That distinction matters for the PL-300 course blueprint. The blueprint includes applying and customizing a theme and applying conditional formatting. These are related skills, but they solve different parts of the design problem. A theme establishes defaults across the report.

Conditional formatting changes presentation according to data values. Used together, they can create consistency while still drawing attention to meaningful conditions. Used without a clear rule, they can make a technically functional report harder to understand.

Imagine an analyst building a governed sales-and-operations report. The business wants a consistent appearance across pages, and it wants underperforming regions to be noticeable. Start by separating those requests.

Consistent typography, background choices, and a standard palette are report-wide design concerns. They belong to the report layer and are candidates for a theme. Whether a particular region is underperforming depends on data.

That emphasis is a candidate for conditional formatting. One requirement asks what the report normally looks like. The other asks when the presentation should respond to a value.

Neither requirement is primarily a Power Query transformation. Power Query shapes data as it’s prepared or refreshed. It could produce useful fields that later participate in analysis, but changing a report’s default visual appearance isn’t a query transformation. The semantic model and Data Analysis Expressions, or DAX, can provide the values and calculations that a report evaluates.

They don’t replace the report-layer decision about how those results should be presented. Publishing the report to the Power BI service also doesn’t turn inconsistent formatting into consistent meaning. The controlling behavior in this case is mainly in the report.

A useful way to locate the owning layer is to ask what must change when the business condition changes. If the requirement is that all report pages begin with the same visual defaults, the data itself doesn’t need to change. The report presentation does. If the requirement is that emphasis follows a calculated performance result, the model or DAX calculation may supply that result, but conditional formatting controls how the report presents it.

These parts cooperate without becoming interchangeable. Moving the presentation decision into data preparation wouldn’t create a report-wide visual default, and changing a theme wouldn’t alter the value produced by a calculation. The requirement determines which layer needs attention.

This layer distinction prevents a common kind of overengineering. Suppose the model already contains a valid measure for variance from target. The analyst doesn’t need to alter the source data merely to display an important negative variance differently. The measure provides the business result.

Conditional formatting uses a data value to alter its presentation. Conversely, a theme doesn’t repair an incorrect variance calculation. If the number is wrong, changing its color only makes the wrong result more prominent. Presentation should communicate meaning, not substitute for correct data or modeling.

The analyst can now restate the requirement more precisely. The theme should provide stable report-wide defaults. Conditional formatting should reserve exceptional emphasis for a defined business condition, such as a meaningful comparison with target.

That wording is much more useful than saying the report should look professional or make bad results stand out. It identifies the mechanism, the layer that owns it, and the evidence needed for acceptance. It also reveals the nearest bad alternative, which is manually coloring individual values or visuals without a reusable semantic rule.

A Power BI theme can be built in, organizational, or custom. A custom theme can be represented in a theme JSON file, where JSON is a structured text format used to describe settings. The important concept isn’t the file syntax.

It’s that the theme supplies defaults intended to operate across the report. This is what makes a theme different from opening every visual and making unrelated formatting decisions one at a time. A shared default reduces accidental variation and gives report pages a common starting point.

Defaults, however, aren’t the same as an unconditional command over every visual. Individually formatted visuals can interact with theme defaults. If someone has explicitly customized a visual, applying or changing the report theme may not produce the expected appearance in that visual.

That’s a useful troubleshooting clue. When most visuals follow the selected theme but one appears not to, don’t immediately assume that the theme failed. Inspect whether that visual has individual formatting that takes precedence over, or otherwise changes, the default you expected to see.

That interaction between defaults and individual choices is important when a report is customized. A shared theme can establish the normal appearance, while a particular visual may have an intentional reason to differ. The difference isn’t automatically a defect. It becomes a problem when nobody can connect it to a requirement or when it changes the meaning consumers take from the report.

The analyst therefore needs to distinguish an explained exception from an accidental override. If an exception is intentional, its purpose should remain clear alongside the shared defaults. If it’s accidental, leaving it in place weakens the consistency the theme was meant to provide. The theme is the baseline, not proof that every visual now follows that baseline.

This also explains why merely importing a theme isn’t sufficient verification. The import can succeed while the finished report remains inconsistent. Some visuals may retain deliberate or accidental individual choices.

Other design choices may technically follow the theme but communicate the wrong business meaning. A successful configuration action proves that Power BI accepted the configuration. It doesn’t prove that every relevant report page now gives the consumer a coherent experience.

Conditional formatting has a different job. It changes presentation based on data values. In the sales report, the analyst might use the variance-from-target result to determine whether a value receives additional emphasis.

The business rule has to come first. If a tiny negative variance is routine and requires no action, coloring every negative number as an urgent exception exaggerates ordinary movement. If crossing an agreed threshold requires attention, that threshold provides a meaningful basis for data-driven emphasis.

Notice that color has no universal business meaning by itself. A high value could be desirable for revenue and undesirable for product returns. A negative number could indicate a problem, or it could represent a favorable reduction in cost.

The analyst can’t select a conditional formatting rule solely because green feels positive or red feels negative. The metric, comparison, and audience determine what deserves emphasis. Consistency means the formatting communicates the same intended idea whenever it appears, not that every metric uses an identical treatment regardless of meaning.

A tempting approach is to make every value visually expressive. Strong colors appear to make a dashboard more informative, but routine values then compete with genuine exceptions. The consumer has to decode decoration rather than follow a clear signal.

A better decision rule is to reserve data-driven emphasis for meaningful thresholds or comparisons. Most values can use the stable defaults supplied by the report theme. Conditional formatting can then identify the smaller set of values that call for interpretation or action.

Now consider what happens when a consumer filters the report to one region. Filter context is the set of filters affecting the data evaluated for a report visual or DAX expression. The value used by conditional formatting may therefore differ after a filter or interaction changes the data being evaluated.

If the rule responds to the resulting value, its presentation may change too. That isn’t inconsistency when the business condition truly changed. It’s the intended consequence of data-driven formatting.

Suppose the unfiltered sales page shows total variance slightly above target, so the summary receives no exception treatment. A consumer selects the Northwest region, where variance is below the agreed threshold. The measure now produces the result for that filter context, and the conditional presentation should reflect the Northwest condition. The theme hasn’t changed.

The report’s defaults remain stable. What changed is the data result that controls the conditional formatting. Keeping those mechanisms separate makes the behavior predictable.

The analyst can test that distinction by comparing the value and its presentation in both states. With no regional filter, the summary value is above the threshold and receives the routine treatment. With Northwest selected, the displayed result moves below the threshold and receives the defined emphasis. Removing the filter should return both the value and its presentation to the original state.

Those observations connect the formatting change to the evaluated business result rather than to the act of selecting something. If the emphasis appeared whenever any region was selected, even when the resulting value stayed within the accepted range, the report would be signaling selection rather than underperformance. That would be a different meaning from the stated requirement.

This gives the analyst a concrete validation method. First, identify a normal case that should not receive exceptional emphasis. Record the relevant value and its expected presentation. Then reproduce a case that crosses the meaningful threshold, perhaps by selecting a region or period with a known result.

Compare the observed value and formatting with the expected business rule. The goal isn’t to prove that some color can appear. It’s to show that the right presentation appears for the right data condition.

Edge cases matter because a normal case can hide a faulty rule. Imagine that the initial report state happens to produce a value far above target. Several incorrect threshold definitions might all make that value look acceptable.

The defect appears only near the boundary or after a filter changes the result. Testing a value on each meaningful side of the threshold gives better evidence. It distinguishes a rule that merely works for the starting data from one that implements the requirement.

The analyst should also test that routine values remain routine. This sounds obvious, but it’s an important part of correctness. A rule that identifies every value as exceptional hasn’t implemented useful emphasis, even if it technically evaluates each data point. Check both the values that should receive attention and those that shouldn’t. The absence of emphasis can be evidence too, provided it matches the stated rule.

When the observed result differs from the requirement, isolate the layer before changing anything. If the displayed number itself is wrong, investigate the data preparation, relationships, calculation, and relevant filter context. Conditional formatting can’t correct that underlying result.

If the number is right but its presentation is wrong, inspect the conditional formatting rule and the value or comparison that controls it. If many visuals use unexpected defaults, inspect the theme. If only one visual differs, check its individual formatting before rebuilding the report.

Consider the analyst’s report after a theme change. The page background, text styling, and most visual colors follow the new defaults, but a single chart keeps an old accent color. The fact that the other visuals changed is observable evidence that the theme is active.

The isolated exception points toward a visual-level choice rather than a report-wide failure. Removing or revising that individual choice may allow the visual to use the intended default. Randomly editing the semantic model would target the wrong layer and wouldn’t explain the symptom.

Now take the opposite symptom. The chart follows the theme, but a region below target isn’t emphasized. The analyst confirms that the variance measure displays the expected negative result. That narrows the investigation.

The data result appears correct, and the general report appearance appears correct. Attention should move to the conditional formatting configuration and the business comparison it represents. Perhaps the condition uses the wrong boundary, or perhaps it evaluates a value different from the one the consumer is reading. The exact cause must be demonstrated rather than guessed.

A controlled comparison keeps that investigation focused. Observe one condition in which the displayed variance is within the accepted range and another in which it crosses the defined boundary. If the values are correct but the presentation doesn’t distinguish the conditions as required, the failure is in the report behavior that maps those values to formatting.

If the displayed values themselves don’t match the expected conditions, changing the colors would hide the symptom without correcting its cause. This is why expected-versus-actual evidence matters. It separates a calculation or filter-context problem from a presentation problem and prevents the analyst from changing several unrelated parts of the solution at once.

After correcting the controlling setting, the analyst retests the original failure and the normal case. That second check protects against an overcorrection. A revised rule might successfully emphasize the underperforming region while also emphasizing regions that are within the accepted range. Fixing one screenshot isn’t enough. The observable result should match the requirement across the conditions that matter to consumers.

The same discipline applies after data changes. A report can look correct today because current values happen to fall into convenient ranges. When refreshed data introduces a new high, low, or boundary value, conditional formatting may reveal an assumption that wasn’t visible before.

The appropriate response isn’t to distrust all conditional formatting. It’s to validate the behavior with representative data conditions and compare the observed presentation with the agreed meaning. Configuration success alone can’t provide that assurance.

Consumer interpretation is the final test. A theme can create polished consistency while the report still uses emphasis incoherently. Conditional formatting can respond accurately to values while making routine results appear alarming. Ask what a consumer should conclude from the visual treatment. If the same treatment sometimes means below target, sometimes means merely selected, and sometimes has no defined meaning, the report has a semantic consistency problem even though each choice is available in the interface.

This is why there’s no universally correct formatting option. Power BI provides multiple valid choices because reports have different metrics, audiences, governance needs, and lifecycles. Familiarity with a feature doesn’t make it appropriate for every requirement. The analyst has to connect the business outcome to the controlling report feature, preserve correct model results, and verify what the consumer actually experiences.

For the sales-and-operations report, the finished evidence is a chain rather than a single successful click. The theme supplies the expected defaults across relevant pages and visuals. Any intentional exceptions can be explained, while accidental visual-level overrides are corrected.

Conditional formatting responds to the defined performance comparison. Normal values remain visually routine, meaningful exceptions receive consistent emphasis, and filters produce changes only when they change the evaluated business result. That chain shows not only that the report works, but why it works at the correct layer.

Themes and conditional formatting support the same goal from different directions. A theme gives the report stable defaults, while conditional formatting changes presentation when a meaningful data condition calls for emphasis. Neither mechanism proves its own correctness merely because Power BI accepts the configuration. You still need to check the relevant values, filters, visual-level exceptions, and the consumer’s interpretation. When routine results remain routine, true exceptions receive consistent emphasis, and the behavior continues to match the business rule under different conditions, you have testable evidence that the report communicates meaning rather than simply displaying color.

Separate what a theme controls from what conditional formatting communicates. Which value change would prove that the formatting responds to the intended rule? A focused checklist and the lesson's exact labels are available in the companion episode reference. This lesson was produced by Bare Metal Cyber dot com.

Applying Themes and Conditional Formatting

headphones Listen Anywhere

More Options »
Broadcast by