Year: 2026
Position: Product Designer
Tasks: UX & product audit, onboarding flow redesign, map interaction & filter UX, information architecture for saved searches and hex bookmarks, product strategy
Tools used: Figma

Mapinamo helps entrepreneurs choose locations for new businesses. The user defines a few parameters — business type, factors that matter to them, competitors — and the system processes several dozen metrics from multiple sources, rendering the result on a map divided into hexagons.

The product had been built to a near-MVP state by two co-founders — data scientists who vibe-coded the interface without design support. They brought me in for a UX review and a prioritized set of recommendations.

Vibe-coded MVPs almost always mix three different layers of problems. Separating them upfront makes prioritization possible:

Cosmetic layer — inconsistent buttons, spacing, typography.
UX layer — friction in flows, cognitive overload, navigation.
Product layer — decisions that affect whether the user reaches the product's value at all.

I logged the cosmetic issues but deliberately pushed them down the list. The most serious problems in Mapinamo lived at the product layer — and that's where the audit needed to start.

The most serious finding wasn't a UX issue in the strict sense.

After completing a multi-step onboarding that already taxes the user cognitively, the map opens with no result. A panel of settings appears, and the implicit expectation is that the user will start playing with filters and eventually arrive somewhere.

From a product standpoint, this is risky. The user came for a specific answer — where should I open my business — and their effort budget is limited. If after serious work in onboarding they still see no suggested locations, motivation drops, and a meaningful portion of users will churn before reaching value the product is actually capable of delivering.

Recommendation: add a dedicated Top Spots tab in the right sidebar that opens immediately after onboarding — five highest-ranked hexagons under default metric weights, each explorable. Subsequent filter work refines an existing result instead of building one from scratch.

This gives users an Aha moment within seconds of finishing onboarding — which is the entire reason they sat through the survey in the first place.

Each onboarding step asked users to compare dozens of options presented as flat text.

The underlying principle: someone opening their first café doesn't know what matters for their business. Forcing them to figure it out from zero offloads work the product can do for them. Presets and smart defaults are a form of product accountability to the user, not a shortcut.

The map UI was built the way a data scientist would naturally build it: a left panel of metrics, each with its own range slider. Self-evident to the people who built it; opaque to a first-time user.

A non-trivial IA question: what exactly is the user saving — the onboarding configuration (business type + metrics + competitors), individual hexagons, or the state of filters at the moment of the click?
The logic I proposed:

The architecture preserves context where context matters, without making the user think about it.

Cosmetics — UI inconsistencies, typography, spacing. Logged in a separate document. At the product's current stage, fixing these has limited impact on key metrics compared with the product-layer changes; they're worth revisiting once the foundational recommendations are in.

Vibe-coded MVPs are an increasingly common artifact. They share a signature: technically functional products whose interface reflects the builders' mental model rather than the user's. Auditing them isn't cosmetic work — it's the transition from tool that works to product that lands.
For Mapinamo, I mapped the problems across three layers, prioritized by product impact, and packaged the recommendations so the team can ship them incrementally, starting with the highest-leverage changes.