Text size

A little more comfortable

Make room
to read.

A little more room. A little less effort. Find the size that feels right for you.

Default

David Zerbe (Raiken), founder of RAIKEN.io 2 min read UX, Prototyping

Building complex low- and high-fidelity prototypes in Axure can test your patience fast. When the choice is purely mine, I usually prefer writing richer prototypes directly in HTML, CSS, and JavaScript. Even so, I have run into plenty of project situations where a data-driven Axure prototype bridged the exact gap between static screens and real code.

Where a static click-dummy falls apart for me in user tests

As long as I only need to check whether someone understands a linear path from A to B, linking static screens together works fine. The moment the product depends on the participant's own input, though, such as filterable data tables, shopping carts, or multi-step forms, a static click-dummy gives itself away within thirty seconds.

I have watched it happen in user tests many times: the participant types in a search term they actually care about, hits enter, and lands on a results screen showing hardcoded dummy rows that have nothing to do with what they typed. Right then, the illusion snaps. Instead of observing how they solve the task, I spend the next five minutes explaining that the prototype cannot really filter.

How I used Axure repeaters to fake a working database

In projects where spinning up a coded prototype was overkill, Axure's repeater widgets and global variables saved me a lot of duplicate screens. Instead of copying ten artboards for every filter combination, you drop your records into a repeater dataset once and attach rules for what happens on user input:

  • Live search and sorting: The participant types a query or clicks a column header, and the table filters or sorts the matching rows on the spot.
  • Carrying input across screens: Global variables remember what someone entered in step one and display their actual values on the review screen at the end.
  • One row template instead of twenty copies: When we notice between test sessions that a card needs one more field, I change the repeater item in one place instead of repairing twenty separate screens.

Where I draw the line and switch to HTML, CSS, and JavaScript

As capable as repeaters and conditional logic in Axure are, they have a very clear tipping point. As soon as I catch myself clicking through three layers of nested if-else dialogs and spending an hour debugging why a variable did not reset, I stop.

Past that threshold, moving straight to a lean HTML, CSS, and JavaScript prototype inside the design process is almost always faster and far less fragile, especially when we want to pull in real data or test responsive behavior genuinely.

What I go by today

I pick the fidelity based on what would break the participant's focus in the test: for a straightforward flow check, linked screens are enough. When the test hinges on realistic filtering and form state without a code setup, repeater tables do the heavy lifting. And the moment clicking rules together gets harder than writing code, I open my code editor instead.

Related Articles

Continue reading about design, UX, and creative tools
FRI
28
AI: your daily work is feeding the machine
Mar 28, 2025

AI: your daily work is feeding the machine

What used to be common practice still holds true today: a few years ago, proving you were human meant clicking on traffic lights, deciphering warped โ€ฆ