The first version of a custom dashboard almost always wants twelve widgets. By the time it is in daily use, it is closer to six. The same six, in roughly the same order, regardless of the industry. The work of a good build is finding which six belong to you.
Why six
Six is not magic. It is the count that fits above the fold on a laptop screen without sacrificing readability. More than six and the eye starts to skim. Skimming is where dashboards quietly die. Day one, the user scans for the number they came for. Day thirty, they only open the dashboard if scanning takes less than four seconds.
Six tiles, well chosen, end the scan in two seconds. That is the whole point.
The shape of the right six
Across the builds we have shipped, a pattern emerges. Most teams need one of each of these six.
1. The deadline tile
A countdown, a date, or a target with a date attached. Almost no generic BI tool puts this on the dashboard. Every team we have worked with should have one. Without a deadline, the rest of the dashboard is just trivia.
2. The pace tile
How are we doing against the deadline, right now. Not last month. Not last quarter. Today, scaled. A pace tile turns the deadline tile from anxiety into action.
3. The blocker tile
The one thing that, if it stayed broken for a week, would make the deadline impossible. Vendor lead time. Open tickets above severity two. Cash buffer in days. There is always exactly one, and the team usually knows it before you ask.
4. The trend tile
Direction over time. A small line chart, not a number. Trend tiles are the only place we use sparklines, because their compactness is the entire point. The eye reads slope faster than it reads a number.
5. The breakdown tile
The same number, split by the dimension that drives the conversation in your team. Region. Channel. Owner. Product. Whichever cut your weekly meeting actually argues about.
6. The change tile
What is different this week from last week, in plain language. Not a chart. A sentence the AI generates. Three sentences max. This is the tile that turns dashboards from passive reports into things people actually open.
The cull
If your draft has more than six tiles, the cull is mechanical. For every candidate tile, ask:
- Has anyone made a decision because of this in the last month? If no, cut it.
- Could the next tile over answer the same question with two more clicks? If yes, cut it.
- Does it tell you the company is fine when the company is not fine? Vanity metric. Cut it.
Most candidate tiles fail at least one of these. The dashboard that survives is rarely the one anyone drew on the whiteboard.
What goes in the second row
Everything else. Below the fold is fine for tiles that exist because someone asks for them four times a year. The dashboard is not a graveyard for old reports. It is a working surface, and a working surface has a top six.
A worked example
A logistics client landed on these six in week one, and has not changed them in eight months.
- End-of-month volume target, with a countdown.
- Pace against that target, scaled by working days remaining.
- Vendor with the slowest current lead time.
- Weekly volume trend, four weeks rolling.
- Volume split by region.
- AI-generated note: what moved this week, in two sentences.
Twelve tiles to six. The dashboard is opened nine times a day, across the operations team. That is the metric that matters.
