MAKIN 
Start from the decision, not the data.
Ask what the viewer will do differently depending on what they see. If there is no answer, the chart is decoration. Most abandoned dashboards are abandoned because they answered questions nobody was going to act on.
One number, prominent, with context.
A figure alone is unreadable — is 84 good? Put the comparison next to it: last period, target, or trend. The comparison is what carries the meaning, and it is the part most dashboards leave out.
If it needs a legend, it needs redesigning.
Anything requiring decoding will be misread under time pressure. Label directly, cut the series count, and accept a less impressive chart that is understood at a glance over a dense one that is not.
Design for Monday, not for the demo.
Demos reward density; Monday morning rewards a screen that answers one question in five seconds. Build for the recurring glance and the dashboard survives past its first fortnight.
Write the decision list before opening the BI tool
Before any chart is drawn, list the recurring decisions the dashboard is meant to support, one line each. Who makes the decision, how often, what they would do if the number rose and what they would do if it fell. That list is the specification. Anything that cannot be traced to a line on it waits for a later version, however easy it would be to add now.
- Decision: reorder stock, chase an overdue account, rebalance a rep’s territory
- Owner: the one person who acts on it
- Rhythm: daily, weekly or at month-end
- Trigger: the level at which they act
- Source: the system the number comes from, such as Tally, Zoho or a Google Sheet
Agree the definition before the design
When people argue about a dashboard number, check first whether they are arguing about its definition. Does revenue mean invoiced, collected or booked? Is a customer active if they ordered this month or this quarter? Write each metric’s definition in one sentence, name the person who settled it, and keep that sentence one click from the number.
Before launch, reconcile every headline number against the source system for a period that is already closed. If it does not match, nobody will trust it on Monday, and a number people do not trust is worse than no number at all.
Choose the comparison that matches the decision
Target suits things the team controls, such as collections chased or calls made. Last period suits numbers that move with the calendar, but compare like with like: a part week set against a full week will look like a collapse. Trend suits noisy numbers, where a rolling average shows direction better than any single week does. Pick one comparison per number and keep it fixed, so the viewer learns what the colour and the arrow mean and stops having to check.
Rules that replace the legend
Each of these removes something the viewer would otherwise have to decode. If a chart still needs explaining after all of them, it is usually trying to show two ideas and should become two charts, or one chart and a table.
- Put the series name at the end of each line, in that line’s colour
- Use grey for context and a single colour for the thing that matters
- Title the chart with the finding, such as ‘Collections behind target in the north’, not ‘Collections by region’
- State the unit and the period in the title, so no axis needs reading
- Use bars to compare items and lines for change over time; avoid pies, dual axes and three-dimensional effects
- Sort bars by value rather than alphabetically
Laying out a screen for the Monday glance
Give each screen one job. The headline number and its comparison sit at the top left, where the eye lands first. Beneath it, only the charts that explain why that number moved, by region, product or rep, each tied to one decision on the list. Detail tables belong on a second screen that people visit when the first one raises a question, not on the screen they open every week.
Show when the data was last refreshed on every screen, and schedule the refresh to finish before people arrive on Monday. If it fails, say so on the screen rather than presenting last week’s figures as current. Check the layout on a phone as well, in case the first look of the week happens away from a desk.
Review whether it is still earning its place
A dashboard needs a review as well as a launch. Once it has been in use for a while, sit with each owner and ask what they did differently because of it. Charts that never changed a decision are candidates for removal. Tools such as Power BI can report which pages are opened, and a page nobody opens is an answer in itself.
If the numbers behind the screen are still scattered across spreadsheets and nobody agrees on them, that is the first job, and it is the problem our Business Intelligence Dashboard package is built around, starting at ₹45,000. For revenue by channel, rep and cohort, Sales & Revenue Analytics starts at ₹40,000.