How to design a SaaS dashboard people actually use
Start a SaaS dashboard from the one question people open it to answer. What to put first, what to move a click away, and which widgets to cut.
On this page 5 min read
Key takeaways
- [object Object]
- [object Object]
- [object Object]
- [object Object]
- [object Object]
- [object Object]
You shipped a dashboard with nine charts. It looked great in the demo. Now people log in, scroll past all of it, click the same menu item they always click, and get on with their day.
That's the usual sign that the dashboard was built to fill a screen. People open it to answer one question. So start from that question, put the answer first, and move everything else a click away. Borrow patterns people already know, like folders, and don't fill space with charts nobody reads.
What should go at the top of a SaaS dashboard?
The answer to the one question people log in to ask, in words or one number, with the next action right beside it.
Every product has that question. For a booking tool it's "who is coming in today?" For a project tool it's "what's late?" For an invoicing tool it's "who do I need to bill?"
Write the question down before you draw anything. Then design the top of the page as its answer.
Before: a project tool opens on a donut chart of tasks by status, a line chart of hours logged and a bar chart of tasks per person. To find out what's late, you read three charts and work it out yourself.
After: the page opens with "3 tasks are overdue". Under it are those three tasks, each with the owner and the due date. One click opens the task. The question is answered before you scroll.
If you can't agree on the one question, ask the people who use it. Five short calls with real customers usually settle it faster than a week of internal debate.
How many charts does a dashboard need?
As many as answer a question someone actually asks. For most small products that's one or two, and sometimes none.
A chart earns its place when the shape matters. Is revenue going up or down? Is this week unusual? A line can show that faster than a number.
A chart is filler when a sentence says the same thing. A donut with six slices that tells you "most tasks are in progress" could be one short line of text.
A quick test for each chart: what would someone do differently after looking at it? If the answer is nothing, it's decoration.
Which dashboard widgets can I remove?
Start with the five below. Each one takes space near the top and answers no question.
- The big welcome banner. "Welcome back, Sam" with an illustration, taking up a third of the screen. A small greeting is fine. A banner pushes the real content down the page.
- The all-time total. A number that only goes up, like total invoices ever sent. It feels good for a week. After that it tells you nothing about today.
- The quick-links grid. Six tiles that repeat what's already in the sidebar. Now there are two menus to learn.
- The "recent activity" feed of everything. Every edit, login and comment in one stream. It's too long to read. If some events matter, show only those and call them what they are, like "needs your reply".
- The donut chart with many slices. Hard to compare, easy to replace with a short list sorted from biggest to smallest.
The one to keep at the top is the answer to the main question, with its next action.
You don't have to delete the rest. Move it to a Reports page. The people who want the numbers will find them there, and everyone else gets a faster start.
How should I organise everything else?
Use a structure people already know. Folders are the safest one, because almost everyone has used a file explorer.
Most of a product is lists of things: clients, projects, orders, files. The temptation is to invent a clever way to browse them. Tags, boards, a custom view with its own rules.
Each new pattern is something the user has to learn before they can do any work. Jakob's Law puts it simply: people spend most of their time in other products, so they expect yours to work the same way. A folder needs no lesson. You open it and the related things are inside.
Before: one long table of every invoice for every client, with five filters above it. To see one client's history, you set two filters and hope you picked the right ones.
After: a list of clients. Open a client and everything for that client is there.
The same goes for smaller patterns. A search box at the top. Settings behind a gear or your profile picture. The newest items first. People bring those habits from other apps they use, so you get them for free. There's more on this in the 5 Laws of UX.
What should I leave out of the product?
Anything that would bring its own screens and statuses without serving the main question. Here's a made-up example.
Say you're building a small tool for freelance photographers to send invoices. The main question is "I need to bill this client." Every client is a folder. Pick one, and the business details, bank details and client address are already filled in because they were saved once.
Then come the feature requests: drafts, estimates, payment reminders, expense tracking. Each one is reasonable. Each one also adds a list, a status and a reason to claim a spot on the first screen. Add all four and your clean dashboard turns back into the nine-chart wall.
Saying no is a trade. Some people will want those features. But a smaller product with a clear first screen is easier to use than a bigger one where you have to hunt. If you do add a feature later, give it its own page and keep the top of the dashboard for the main question.
How do I check if my dashboard works?
Show it to someone for a few seconds, hide it, and ask what they would do next. If they can't say, the top of the page is wrong.
You can run this with a teammate or a customer on a call. Don't explain anything first. If they describe the charts ("there was a blue graph") but can't name an action, the page is showing data without an answer.
Then watch where people click first after logging in. If most of them leave the dashboard right away for the same page, that page is your real dashboard. Bring its key content to the front.
If your dashboard has grown into a wall of charts, book a call. Show me the screen and tell me what your users log in to do. I'll tell you what I'd put first and what I'd move out of the way.
Questions
-
It answers the main question a user has when they log in, and it puts the next action beside that answer. Everything else is a click away.
-
There's no fixed number. Keep the ones that change what someone does next, and move the rest to a reports page. For a small product that often means one main item and a short list.
-
No. If people always go straight to one list, such as their clients or their projects, make that list the first screen.
Read next
View all posts
Best text for a call to action button (not Submit)
The best text for a call to action button names what happens after the click. See rewrites of Submit, Click here and Learn more.
Consistent icon style in UI: why mixed sets look cheap
Mixed icon sets make a UI look stitched together. How to spot it, why emoji fail as icons, and how to pick one family and weight.
How to design a website without content (no lorem ipsum)
Copy not written yet? You can still start designing. Use rough draft copy instead of lorem ipsum. Here's how, step by step.
Have a project in mind?
Book a short call. We'll talk through what you're building, and I'll tell you how I can help.
- Length
- 15 minutes
- Where
- Cal Video
- Time zone
- Metro Manila