Write explanations that match the selected evidence

Understand the method

A dashboard explanation should help the reader decide what the numbers support and what to investigate next. It is not a transcript of every card. Start with the material observation, add the context that changes its interpretation, state the relevant limitation and propose a proportionate next action. This sequence makes a short note useful without making it sound more certain than the data allows.

Make the observation concrete

For May/All, completed-order value is €3,420 versus €2,900 in April, an increase of €520 or about 17.9%. Each month contains 20 completed orders, so overall AOV rises from €145 to €171. State the period and population. Calling this simply “performance up 18%” would leave the measure and denominator unclear, while rounding to whole percentages can hide differences when readers compare a separate table.

Add the context that matters

Both channel AOVs fall: web from €100 to €90 and store from €550 to €180. Store’s order share moves from 10% to 90%. April’s store baseline is only two orders and includes verified A020 at €1,000. These facts qualify the headline. Keep the legitimate large order in baseline values, and label any exclusion scenario rather than silently substituting a more comfortable result.

Use precise change language. A share moving from 10% to 90% rises by 80 percentage points. It rises by 800% relative to its starting share. These statements are numerically different. For the dashboard explanation, percentage points are often the clearer expression of a share difference. The completed-value change uses a different denominator, April completed value, and should not be conflated with the mix change.

Keep the narrative tied to the view

If the reader selects May/Web, the observation changes to €180 from two orders, AOV €90, and completed value down 90% against April/Web. A fixed headline saying value rose 17.9% would be wrong for that view. Separate dynamically selected facts from wider-fixture context. It is acceptable to retain a static note saying “Across April and May, store’s share changes from 10% to 90%” because its wider scope is explicit.

Say what remains unknown

Observed delivery means are about 3.64 days for 14/20 completed April orders and 3.53 for 19/20 May orders. Changing coverage limits an all-order service claim. Costs, margins, experimental allocation and longer history are absent, so the extract does not establish profit or causation. This is a material analytical boundary, not a generic warning appended to every sentence.

Propose a useful next action

A suitable next step is to investigate the channel-mix shift, gather comparable periods and obtain cost evidence before recommending expansion. Do not turn a descriptive association into a claim that the initiative caused the increase. An honest dashboard can support a request for more evidence. Its value is that the reader can see why that evidence is needed and which decision it would inform.

Read the note aloud and remove repeated numbers that add no context. Keep the essential figures, denominator and limitation. A title such as “May’s larger total needs channel context” is more informative than either an exaggerated success claim or an empty label such as “Key insights.” The tone should make the reasoning easy to inspect.