Building reports — quiz

What makes a scorecard genuinely useful?

1. Compact numbers
2. A comparison date range showing the change
3. A border
4. A long title

Answer: A comparison date range showing the change — A number alone gives no sense of whether it is good or bad.

Which setting makes a dense table scannable?

1. Wrap text
2. Bars or heatmap on a metric column
3. More rows per page
4. A summary row

Answer: Bars or heatmap on a metric column — Visual encoding in cells beats reading digits.

A time series with a 30-value breakdown dimension gives:

1. A clear chart
2. 30 unreadable lines
3. An error
4. A pivot table

Answer: 30 unreadable lines — Filter to the top few or use a table.

Optional metrics allow the viewer to:

1. Change the date
2. Add metrics from a menu themselves
3. Edit the report
4. Export data

Answer: Add metrics from a menu themselves — It keeps the default view clean.

Where should a reusable calculation be created?

1. In each chart
2. At the data source level
3. In Theme
4. In a filter

Answer: At the data source level — Chart-level fields cannot be reused and tend to diverge.

SUM(Revenue)/SUM(Sessions) differs from SUM(Revenue/Sessions) because:

1. It is faster
2. The second averages per-row ratios, weighting rows equally
3. The first fails on nulls
4. They are identical

Answer: The second averages per-row ratios, weighting rows equally — Ratio of sums is the correct form for a rate.

Every metric in your blend is null. Likely cause?

1. Wrong chart type
2. The join keys do not match in format
3. Too many sources
4. Caching

Answer: The join keys do not match in format — Normalise the key with a calculated field before blending.

Join keys in a blend must be:

1. Metrics
2. Dimensions
3. Dates only
4. Calculated fields

Answer: Dimensions — Blends join on dimensions.