I have to design a prototype where the user has to choose from a lot of selection, multi-selection and dropdown widgets. All choices are dependent on each others and there will be a lot of complex logic behind it (e.g. if the user chooses option A, they have some possibilities A1, A2, A3, if they choose B some other possibilites will be available, but if they choose A and C some other possibilities like A1, A2, B3, C1 will be available).

I need to keep things simple and I'm looking for some kind of pattern that I can use to design and implement a not-too-complicated UI while also satisfying the logic and the complexity needs.

  • Will the label for the second choice always be the same title? Ex. Question 1: Pick Car Question 2: Pick Car Color Commented Sep 9, 2015 at 16:32
  • 2
    Can't you use more than one select box and populate each successive one based on the prior inputs? If the complex logic will take a lot of time to process, then you can separate it into multiple screens with multiple questions (piping). Commented Sep 9, 2015 at 16:54
  • Hello @nkint, could it be possible to add an example of what you have for data? Some real case could help to understand the problem better :-)
    – Xabre
    Commented Sep 10, 2015 at 16:27
  • How many levels of nesting occur during this "filtering process" in your estimation? Commented Sep 16, 2015 at 5:49
  • hi, I think you are looking for something like this ux.stackexchange.com/questions/92157/…
    – LNubiola
    Commented May 12, 2016 at 16:16

1 Answer 1


Sending post within the UK and abroad requires lots of decisions and different bits of information depending on previous selections. It's probably a similarly complex process in many other countries too, but the UK Royal Mail Price Finder actually does a pretty good job of this, as it's otherwise a bit of a nightmare.

The overall impression is reasonably simple - it doesn't look complicated at the start - just one main box for 'where are you sending to'. Further potentially unnecessary steps don't appear until needed, and you only get to see what's relevant after any decision point.

Some of the decision trees are probably graphs - joining up with alternative routes, and in cases where the user may not know enough to make a decision a 'Not sure' option asks a different question to re-route the user's path and get them back on track.

At the end when there's a number of shipping options (as opposed to decisions), you get a sortable comparison table.

What this definitely does require is a good analysis of all the possible paths, dependencies and questions to be answered. And lots of testing.









way :)

enter image description here

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.