Dashboards go wrong in a small number of ways, over and over, in products built by good teams who were not being careless. The reason is structural: nobody designs a dashboard badly on day one. It gets one more chart per quarter, one more sidebar link per customer request, and after three years the screen is a record of every conversation the team has had rather than a tool for the job somebody opens it to do.
These are the problems that turn up in almost every review, and what we did about each one when designing the dashboards and web apps in our own portfolio.
It opens on everything, so it answers nothing
The default dashboard shows all the data the team was proud of shipping. A user opens it with one question in mind and has to search the screen for the answer, every single time, several times a day.
The fix is to decide what the first question is and put it at the top alone. The investment dashboard we designed opens on four numbers and no more: index level, amount invested, current value and return, each in its own tinted card across the top, so the eye lands on the one it came for without reading labels. Everything below answers a follow-up question, in the order those questions get asked.
The test is easy to run on your own product. Ask three users what they open it to find out. If you get three answers and none of them is on the first screen, you have this problem.
The sidebar has become a list of everything the product does
Navigation grows one link at a time and nobody removes any, so the rail ends up as eighteen destinations of equal weight. Users stop reading it and rely on muscle memory, which works until anything moves.
Short rails work. The investment dashboard runs on five destinations with one active. The compliance dashboard holds five in a dark navy rail: clients, users, audits, projects, questions.
When the product genuinely is that large, collapse rather than list. A records platform we designed carries incident reports, analytics, reporting, learning, communication, audits and history in one application, so its sidebar reduces every module to a single line and expands only the one in use. A page three levels down still shows where it sits in the whole.
Three things are competing to be the brightest
Every team that adds a feature adds a coloured button for it. Six quarters later, four controls on the screen are shouting and the user’s actual next step is one of them.
The rule we hold to is one bright element per screen, and it is the action you came to perform. In the compliance dashboard that is the single create button in the accent colour. In an order operations web app, the only strongly coloured control on the page is the one that creates something new. Everything else is quiet.
This is not minimalism for its own sake. Emphasis only works by contrast, so a screen where four things are emphasised has no emphasis at all.
Every status is a filled badge
Status columns start with two states and end with nine, each as a filled pill in its own colour. The table becomes a pattern of coloured blocks and the rows underneath stop being readable.
In the order operations app, status is a coloured dot and a word rather than a filled badge, which lets many states share one column without the table turning into wallpaper. The colour still carries the state at a glance; the word carries it for anyone who cannot separate two of the colours.
The related rule: colour for state, never for decoration. In a compliance tool where hundreds of rows have to stay legible, every decorative use of colour spends attention that a real signal then cannot borrow.
Checking one row costs you your place
An operator scrolls to row forty, clicks it to check something, lands on a detail page, goes back, and the list has reset to the top. Repeat that thirty times an afternoon and you have designed a product people describe as slow when nothing about it is slow.
Expand in place instead. In the order operations app a row opens where it sits, because someone hunting one order among hundreds should never lose the list to look at it. The same principle applies to filtering: on the tutor search we designed, the filter panel changes the result list in place, with no separate results page to travel to and come back from.
The numbers people compare do not line up
Comparison is most of what dashboards are for, and comparison happens down a column, not across a row. When the figure a user is weighing sits in a different horizontal position in every card, they have to read each entry in full to compare two of them.
The tutor search puts hourly rate and star rating in a fixed right-hand column, so they line up down the page and can be compared without reading a single row properly. Whatever your users are actually choosing between, that value belongs in the same place on every card.
Search does not accept what people paste into it
Users arrive at a dashboard holding an identifier from somewhere else: an order number from an email, a reference from a phone call, a client name from a spreadsheet. If the search field only matches titles, they end up filtering by hand.
The order operations app takes an order number directly in the search field, and every column header sorts, because the table carries only what identifies an order in the first place: sales channel, order id, status, shipping name and address, country and date. Deciding what a row is for is what makes both of those possible.
The thing every task starts with is buried
In multi-tenant tools, almost every task begins by choosing whose records you are in. That control is often a dropdown inside the sidebar, three clicks deep, styled like a filter.
In the compliance dashboard the client switcher sits in the top bar instead, above the navigation, because it governs everything below it. Find the one choice that scopes every other action in your product and give it the position that says so.
Every list screen was designed separately
Clients, users, audits and invoices each get their own layout, because each was built in a different sprint by a different person. Users then learn four interfaces to do one kind of task.
One pattern, repeated, is the cheaper answer: title, search field, a single create button, a table with sortable headers and generous row height, then a result count with pagination underneath. In the compliance dashboard, clients, users and audits all share that layout without any of them being redrawn, which also means the next list screen costs almost nothing to add.
The check worth running this week
Open your main dashboard, set a timer for five seconds, and write down what you can answer without scrolling or clicking. Then ask two people who use it daily to do the same. The gap between those lists is the first thing to fix, and it usually costs a layout change rather than a rebuild.
If you want two designers to run the full pass instead, the free 15-point audit takes a live URL or a test account and comes back within one working day, ranked, with a screenshot and a suggested fix for each finding.