Multi location | New portfolio

How a migration constraint shaped a better design decision.

Wix Restaurant Orders is an online ordering app for food businesses on Wix. When we rebuilt the app, multi-location support was the most requested missing feature — and migrating an existing user base from the legacy app meant any change to the ordering flow had to be handled carefully. That constraint became the lens for every design decision.

One thing worth noting: Wix is a DIY platform, which means every screen I designed functions as a skeleton — the layout, interactions, and logic are mine, but each business owner customizes the visual style to match their own site. The design had to work across thousands of different aesthetics.

Role

Lead Product designer

Platform

Desktop, mobile & Wix editor

Team

UX writer, PM, Dev

Contributions

UX, Product Strategy, Visual design

Year

2025

Selecting an AI powered designer

Challenge

We had a large user base on the legacy app waiting to be migrated. The data made the priority clear:

Of legacy app users had multiple locations

Of total sales income came from multi-location sites

Of the app's top users were multi-location businesses

Migrating them came with a condition: the experience for their customers had to either improve or stay largely the same. A dramatic change in the UX risked churning the very users we were trying to bring over.

The redesign did indeed move both KPIs we set out to improve.

Screens & user flow in the legacy app, a "menu first" approach

Research

Our research was anchored around one central question: how far into the ordering journey should a customer get before being asked which location they're ordering from? We approached it from two angles.

Competitor analysis

Every direct competitor used a "location first" approach: customers must select a location before browsing the menu. Clean and unambiguous, but it adds friction for undecided customers who want to browse before committing. 92% of competitors accepted that tradeoff.

Example of competitor user flow "location first" approach

User interviews
I ran interviews with 5 users across the old and new apps. 4 out of 5 believed the location picker should appear first — but 3 of those 4 wanted it to be dismissible.

Additional insights: no concept of a "main" location, consistent menus across locations, and a strong need for scheduling clarity given that locations often have different opening hours.

Tradeoffs

The research pointed us toward a decision: location first (the industry standard) added too much friction, and menu first (the old app) risked confusing customers about which location they were ordering from.

I proposed a middle-ground approach that I hadn't seen any competitor use: the location picker appears first but can be dismissed, letting customers browse anyway.

Example of competitor user flow "location first" approach

This way customers actively decide whether they're ready to commit or just browsing — and either way, they've seen the picker and know a decision is needed to place an order.

Decision 1

Menu first approach— Shipping lean to protect the migration

New screens and flow in a Wix site, according to users site theme & typography. Menu first approach

We shipped menu first as a deliberate MVP to minimize migration risk. It wasn't the ideal end state — but it was the right first step. The middle-ground approach was fully designed and ready to roll out in V2 once the migration was complete.

Full flow in context, new design

Decision 2

Making the next step clearer

New banner in a Wix site (using the sites theme and color)

We made the "Order now" banner sticky and by default use the sites CTA's color. It's sticky and persistent, serving as a constant but non-aggressive reminder to select a location. A second entry point to the picker without getting in the way. The old app's equivalent had a 7% click rate — almost nobody noticed it.

Banner in the legacy app

New design

Decision 3

Bringing clarity to location discovery

Legacy app (dropdown)

New app (lobby)

A dropdown in the old app made locations easy to miss and hard to compare. A visual lobby gives each location equal weight and equal visibility, helping customers make a more informed choice with one less click — which aligned directly with the user interview insight that no location should feel like a "main branch."

Decision 4

The Wildcard designer: designing for the feeling of control

We added a sixth option: the Wildcard. Technically, selecting it means the system picks a designer on the user's behalf. But the user doesn't experience it that way — they've made a choice, and that choice carries a sense of anticipation going into the generation. Agency — even symbolic agency — changes how people feel about a result.

Tradeoffs

Improving discoverability was the north star of this project — which meant the entry point needed attention too.

Enhancing the entry point (shipped)

Previous entry point

New entry point

A subtle but impactful update: a new icon, revised phrasing, and improved visual hierarchy that made the feature more visible without disrupting the flow.

Post-send prompt (dropped)

Since designer selection is optional, it's highly prone to being skipped. We explored surfacing it as a dismissible prompt after hitting send — a last chance to choose before generation kicks off. We dropped it when we realized interrupting the flow at that moment carried too high a risk of abandonment.

Smart prompt analysis (scoped out)

We looked at using AI to auto-select a designer based on the user's prompt. The logic was sound, but getting it right would require significant development work. We scoped it out for this version and flagged it for a future iteration.

Impact

Discoverability & usage:

+16%

Users completing designer selection

+7%

Users engaging with the entry point

-18%

Drop-off rate

Distribution is more equal:

The redesign did indeed move both KPIs we set out to improve.

Designer selection distribution

Takeaways

1. Optimizing one step can quietly damage the ones that follow.

It's tempting to push for maximum engagement with any single feature. But in an onboarding funnel, every step exists in context. The real measure of success here wasn't designer selection rate — it was whether users went on to generate, publish, and upgrade. We kept that lens on every decision, including the ones that led us to drop the mandatory-step concept entirely.

2. Simpler visuals communicate more.
The more specific the imagery, the more it narrowed users' expectations — and the more confused they became. Pulling back to mood-based thumbnails gave users just enough to make a confident call without over-promising a specific output. Less really was more, and the data confirmed it.