Quick answer
A dashboard is a single screen that shows the few most important metrics for a goal side by side, so a person can tell at a glance whether things are on track. It is for noticing problems quickly, not for displaying all of your data.
What is a dashboard?
A dashboard is a single screen that puts the few most important numbers for a goal side by side, so a person can see at a glance whether things are on track. The name is borrowed from the car, where the instrument panel lets a driver check speed and fuel without stopping to read a manual. The data-visualization writer Stephen Few gave the definition most often quoted in business intelligence in 2004: a visual display of the most important information needed to achieve one or more objectives, consolidated and arranged on a single screen so the information can be monitored at a glance.
Two phrases in that sentence do the work. "Most important" means a dashboard is selective: it is not the place for every metric you can collect. "At a glance" means the reader should know whether to worry within seconds, before reading any detail.
Dashboard vs report vs scorecard
People use these words interchangeably, which is how dashboards become cluttered. A quick separation:
| Dashboard | Report | Scorecard | |
|---|---|---|---|
| Purpose | Notice that something changed | Explain in detail | Judge progress against targets |
| Reader time | Seconds | Minutes to hours | Minutes |
| Content | A handful of metrics, charts and alerts | Tables, many dimensions | Goals with target and status |
| Updated | Continuously or daily | On request or on schedule | Weekly or monthly |
| Question | "Is anything off?" | "Why did it happen?" | "Are we on plan?" |
If a screen needs scrolling, a legend and an explanation before it makes sense, it is a report wearing a dashboard's name.
Three types of dashboards
| Type | Audience | Typical contents | Refresh | Example question |
|---|---|---|---|---|
| Operational | The team running things day to day | Live or hourly counts, errors, queues, thresholds | Real time to hourly | "Is checkout failing right now?" |
| Analytical | Analysts and product or marketing managers | Trends, comparisons, breakdowns by channel or segment, with filters | Daily | "Which channel's conversion fell this week?" |
| Strategic | Leadership | A few company-level goals against targets and prior periods | Weekly or monthly | "Are we on track for the quarter?" |
The same number can appear on all three, but each type answers a different question and needs a different level of detail. Mixing them on one screen is the most common design failure.
Rules for designing one
- Start from a question, not from the data. Each card must answer one question that someone will act on. If you cannot name who acts, remove the card.
- Always show a comparison. A number alone is meaningless. Show the previous period, a target or a baseline next to it.
- Put the most important card top left. People read from there first.
- Define a threshold. Decide in advance what counts as a problem (for example, a relative drop of more than 10% against the previous period) and make it visible.
- Match the chart to the question. Trends over time need a line, comparisons between a few categories need bars, and a single current value needs just the number.
- Write the definition. Name the metric, the date range and the counting rule on the card. Two dashboards that both say "conversion rate" with different rules cause arguments, not decisions.
- Give it an owner and a review rhythm. A dashboard nobody opens on a schedule is decoration.
A worked example: reading a weekly dashboard
Sample data: a weekly dashboard for an online shop, this week against last week.
| Card | This week | Last week | Change |
|---|---|---|---|
| Sessions | 12,400 | 13,100 | −5.3% |
| Orders | 372 | 419 | −11.2% |
| Conversion rate (orders ÷ sessions) | 3.0% | 3.2% | −6.2% (−0.2 points) |
With a rule that flags relative drops larger than 10%, only the Orders card turns red. The dashboard did its job: it showed that something is off. It also contains the first clue about why. Orders equal sessions times conversion rate, and 0.947 (traffic) × 0.938 (conversion) = 0.888, which is the −11.2%. About half of the drop came from fewer visitors and half from a lower conversion rate. The next step is not on the dashboard; it is a breakdown by channel or landing page in an analytical view. That handoff, from dashboard to breakdown, is how a good dashboard is meant to work.
A starter layout
For a weekly website dashboard, five cards are enough: (1) sessions with the previous period, (2) conversions or orders with the previous period, (3) conversion rate, (4) the top three sources by conversions, and (5) an error or failure rate with a red threshold. Put conversions top left, add a one-line text card with the date range and counting rules, and leave out everything else until someone asks a question these cards cannot answer.
Common mistakes
- Too many cards. Every extra card makes every other one harder to see. If everything is important, nothing is.
- Vanity metrics at the top. Total visitors rarely drive a decision; a conversion rate or an error rate usually does.
- No comparison or target. "12,400 sessions" is neither good nor bad.
- Different date ranges per card without saying so. Readers assume one range.
- Chart decoration. 3-D, gauges that need a legend, and rainbow colours cost reading time and add nothing. See data visualization.
- Never reviewed. Metrics change meaning when products change; revisit the dashboard each quarter and delete what nobody used.
Where to go from here
This page defines the term. For which metrics belong on a dashboard, see analytics dashboard and KPI dashboard; for a version aimed at business owners, see business dashboard.
Building a dashboard in TapCub
TapCub's dashboards are made from analyses you have already saved. To rebuild the weekly example above:
- Save the analyses. Build the sessions, orders and conversion analyses once in the analyzer and save them. A saved analysis can then be used as a dashboard card.
- Create the board. Open All-platform analytics › Overview › My dashboards. A card can be a saved analysis, a text block or a divider; use text blocks to write the metric definitions from rule 6.
- Set the board's own range. A dashboard has its own date range, filters and identity (visitor, device, person or account), and if you do not change them each card keeps the setting of its own analysis.
- Make it the default, or pin it. A board can be the default, and it can be attached to the web, app or mini-program side menu; opened inside that menu, it shows only that platform.
- Add a threshold. For the 10% rule, create a monitor in All-platform analytics › Operations › Monitors, which checks a metric on a schedule and notifies you. It only notifies, it does not act.
- Share it. A dashboard can be shared as a read-only link, with an optional password, showing at most the last 90 days. See share a dashboard.
Industry packs also install ready-made dashboards together with events and goals, which gives a starting point you can then trim to the seven rules above.
Sources
- Stephen Few, Dashboard Confusion, Perceptual Edge (first published in Intelligent Enterprise, March 2004).
- Databricks, What are dashboards?
Was this explanation clear?
We use your answers to fix and expand this page.
Couldn’t send. Please try again in a minute.
Thanks for the feedback. We’ll use it to improve this page.You’ve already sent feedback on this page. Thanks.

Data Engineer
Writes about event design, data quality, and collection workflows, with verifiable implementation methods for reliable analytics.