• UX Research
  • UX/UI Design
  • Design System

rwmrreviews from people you trust.

A social travel app was drowning in features. …

Client
Roam · Pre-seed Stage
Year
2024 – 2025
…

Overview

Roam let people log places they visited — …

The objective

Raise completion of the core posting flow …

01 Research

150 interviews • 320 Survey Response • 14 usability runs

What people came to do, versus what we asked them to do.

Interview Findings

I ran 150 moderated sessions…

Survey Findings

Extracted from the interviews…

Three findings during Usability Testing that set the direction

  • Finding 01

    Taxonomy was the product

    Users liked categorizing…

  • Finding 02

    Visuals are non-negotiable

    Contrary to our initial…

Three verbs, not nine fields.

The tree-test told us the category names had to do the explaining. I cut the taxonomy to Activity / Food / Stay and tested the labels until prediction accuracy crossed 90%.

The create-flow became: pick a verb, drop your photos, save. Type, price and companions moved into an optional "enrich" step — never blocking the save.

Verb-first type select screenverb-first type select2.1 · IA

The rebuilt create-flow

  1. 01

    Pick a Type

    Activity, Food, or Stay. Three tap-targets, no scrolling, plain language.

  2. 02

    Drop Visuals

    Camera roll opens first. Words are optional and come after the image.

  3. 03

    Share & Save

    One tap saves. Rating, price and tags become editable on the post itself.

03 Beyond the form

Profile • Explore • Create • Enrich • Post

The three verbs had to live everywhere, not just at "create".

Figuring out the create-flow was the lever, but the same taxonomy had to hold across the whole app. If “Activity, Food, Stay” only existed on one screen, the old confusion would creep straight back in everywhere else.

So the three verbs became the spine of the product: they tag every post on your profile, they filter the explore feed, and they carry through the location step before a photo is ever chosen.

“Enrich later” became a language, not a form.

The metadata we cut from the required flow did not disappear — it moved into an optional enrich step and changed shape entirely. Instead of dropdowns and price bands, people pick the words they would actually say out loud.

Six axes, plain language, up to six labels per roam. “Clean-ish” beats a hygiene score; “Worth the Hype” beats a 1–5 field for buzz. The taxonomy does the explaining here too.

Enrich step label pickerverb-first type select2.1 · IA

Rating and story come last.

Stars and a free-text story appear at the end of the post, following the type, photos, and optional labels. A roam can consist solely of a verb, an image, and three labels; all additional elements (such as the other 3 labels) serve as optional user-generated enrichment.

This placement strategy enhances user flow by allowing users to define the story type, add images, consider labels, and finally assign a rating. This sequence encourages ratings that reflect nuanced experiences rather than extremes, countering the common cognitive bias where users recall only the best or worst moments.

The story field is designed to be conversational with ample character limits and a placeholder that prompts authentic user feedback.

Rating and story step

The Label System, Six Axes with Human Words, and only 3 required labels

  • Axis 01 · Hygiene

    Dirty → Spotless

    Dirty · Clean-ish · Spotless

    Users noted that people rarely discuss this unless places are very dirty.

  • Axis 02 · Quality

    Disappointing → Top Notch

    Disappointing · Average · Top Notch

    An evaluation focused solely on quality, distinct from cost to prevent confusion between factors.

  • Axis 03 · Value

    Overpriced → Budget Friendly

    Overpriced · Fair Deal · Budget Friendly

    Monetary value is assessed clearly and independently, helping set expectations.

  • Axis 04 · Service

    Careless → Attentive

    Careless · Attentive

    A binary scale intentionally designed, service quality is a decisive factor users assess quickly.

  • Axis 05 · Hype

    Overrated → Worth the Hype

    Overrated · Underrated · Worth the Hype

    The most frequently utilized scale during testing, reflecting users’ primary feedback focus.

  • Axis 06 · Vibe

    Casual → Energetic

    Casual · Fancy · Romantic · Energetic

    The sole dimension guiding user choice of destination, beyond mere quality assessment.

The Impact

  • 43%

    Positive user feedback

  • 91%

    category prediction accuracy

  • -6

    required fields, 9 → 3

04 Design System

Tokens · Type · Icon set · Components

A small system, sized to the whole app.

None of this scales without a system underneath it. I built rwmr's design system as a tight set of reusable tokens, so the three-verb logic and the enrich vocabulary stay consistent as the app grows past the screens in this case study.

The rules are deliberately strict: a canvas that adapts to the user's preferred theming (light/dark), the key highlight color using the brand color (purple), only used to highlight certain elements and for primary actions, neutral foreground colors depending on hierarchy for text, and of course, a semantics palette for errors, warnings, and success. The semantic colors are also used to easily identify positive and negative reviews through label colors.

Text Styles, one font, reading with comfort, but with personality

Solving today's problems, ready for tomorrow's expansion

Developed a comprehensive design system comprising over 30 meticulously crafted text styles, leveraging the Poppins typeface for its balance of approachability, readability, and seamless integration within development workflows.

These text styles were engineered to be fully responsive and adaptable, supporting a range of devices including iOS and Android smartphones, iPads, tablets, and desktop environments. This forward-thinking approach ensures scalability and consistency across platforms, minimizing future maintenance by preemptively addressing potential typography adjustments within the system architecture.

Text style specimens: LargeTitle, Title 1–3, Headline and Subheading with weights and sizes

Email me to see the full design system!

nourhanghabour@gmail.com

Reflection

What I would test next.

Completion is up, but I want to watch whether the optional "enrich" step ever gets used — a save that never gets enriched might mean the metadata was never the point, and we could cut even further.

I would also localise the three verbs properly. "Stay" has no clean one-word Arabic equivalent that matches the tone, and the team shipped an English label as a stopgap.

Co-Founder & Lead Frontend Engineer

"Nourhan helped us launch our dream, professionally, and to perfection. We worked with various designers before, but no one was as easy to give feedback and give requirements to"

— Youssef Ammar, Co-Founder, rwmr

Next

e-stebdal

Returns & operations management SaaS solution

  • UX/UI
  • Branding
Read Case Study