Pretty UI is a rounding error.
If your product underperforms, it is rarely because the interface is not pretty enough.
Yes, good design matters. But in most SMEs I work with, the real conversion killer is much less glamorous:
People land in your app/portal/SaaS and they do not know what to do next.
So they hesitate. They click around. They bounce. Or worse, they complete the wrong thing and then your ops team gets the angry email.
The other common culprit is even more predictable:
You built the journey on internal opinions instead of evidence.
It happens with the best intentions. A senior person says, "Our customers will want X." A stakeholder says, "Competitor Y has this." Someone else says, "Make it cleaner and more modern." Then you end up polishing the wrong thing very nicely.
If you want a practical fix (that does not require a 12-week discovery phase and a small novel of documentation), do this before you open Figma.
Step 1: Write the one-sentence job-to-be-done.
Not a feature list. Not a mission statement. A single sentence that finishes this:
"When I am in situation X, I want to do Y, so I can achieve Z."
Examples:
- "When I need to reorder stock, I want to do it in under 2 minutes, so I can get back to serving customers."
- "When I need to submit a claim, I want to know exactly what is required, so I can get it approved first time."
If you cannot write this sentence, you are not ready to design. You are about to design vibes.
Step 2: Speak to five real users.
Five. Not fifty. You are not trying to produce statistically significant research. You are trying to stop guessing.
And here is the part most teams get wrong: do not ask what they would do.
Ask about the last time they tried to do that job.
"Tell me about the last time you tried to [do the job]. Where were you? What triggered it? What did you do first? What got in the way? What did you do when you got stuck?"
People are brilliant at describing what happened.
They are terrible at predicting what they will do in a hypothetical future.
Step 3: Map the next step.
Once you have those five conversations, you will spot patterns fast:
- The moment they get confused
- The words they actually use (which will not match your internal terminology)
- The hidden steps you assumed were obvious
- The workaround they already do (usually in email and spreadsheets, because of course they do)
Then, and only then, open Figma.
At that point, the interface decisions get easier because you are not designing "a dashboard". You are designing a clear next action for a specific person in a specific moment.
The design will almost draw itself.
And if you still want it to look lovely at the end, great. Just earn the right to polish by making it obvious what to do next.