Bizinc

What made this hard

The hardest part was Business Profile 2.0 — one profile serving an owner who edits it and a visitor who judges it. Deciding field by field what an owner controls, what a visitor sees instead, and what belongs to neither is the kind of thing that quietly becomes two disconnected products if it is never settled explicitly. Working it out as a written requirements list, business view beside client view, was slower than designing screens directly and it was the only way to make each difference deliberate rather than accidental. Redesigning in place added a second constraint: the platform was already live with businesses on it, so v2.0 had to be a version people could be moved to, not a fresh start.

Research & discovery

Research status — Product-design-led, with stakeholder discovery

Who

Two audiences whose interests only partly overlap — people looking for a service, and the businesses selling it.

Investigated

Where one profile has to serve an owner who edits it and a visitor who judges it, and which signals actually decide an enquiry.

Learned

The two views diverge field by field, and leaving that implicit was quietly producing two disconnected products.

Changed

Business Profile 2.0 was specified as a written requirements list, business view beside client view, before any screen was drawn.

Still open

Which parts of the redesign drove the activation and growth figures — the numbers are aggregate, not attributed.

Next test

Test whether business owners understand the difference between profile editing and customer-facing profile content, and instrument the discovery path to see whether the open-now, distance and category signals shorten the route to an enquiry.

Key decisions

1. 01

   Mapped the two audiences as separate flows before drawing a single screen — sign-up, business-owner and client journeys side by side — then reconciled them into one site map, so where the paths diverge and where they share a screen was a decision rather than an accident.

2. 02

   Specified Business Profile 2.0 as requirements in top-down order, business view beside client view. Laying the two lists against each other made every difference deliberate: what an owner can edit, what a visitor sees instead, and which fields belong to neither.

3. 03

   Wireframed at desktop scale before any visual design — profile, search, dashboard, marketing tools, listings and the community forum — so layout and hierarchy were settled while they were still cheap to change.

4. 04

   Designed discovery for people who are comparing, not browsing. Business cards lead with the signals that decide an enquiry — what the business does, where it is, how far away, whether it is open right now — so a shortlist forms without opening five tabs.

5. 05

   Treated the business dashboard as the product operators run on, not a settings screen: listing management, deals, services and products, and performance sit together, so being discoverable and being operational are not two separate tools.

6. 06

   Kept the marketing and lead-generation surfaces inside that dashboard rather than bolting them on, so the line from a listing to the enquiries it produces to what the owner does next stays visible in one place.

7. 07

   Designed the AI assistant as a way into the marketplace rather than a support widget, and tied it to the map: it asks what you need, reads where you are, and returns real businesses nearby with availability you can act on — so the conversation ends in a booking instead of a link, and a small local business becomes findable to someone who did not know its name.

8. 08

   Built the interface as reusable components with Auto Layout, variants and design tokens, and versioned the file from v1.2 through v2.0 to a development-ready page — so consistency survived the redesign and engineering had one place to build from.

9. 09

   Worked directly with stakeholders to turn business requirements into interface decisions, and kept every journey responsive from the first wireframe, since the customer side is used mostly on a phone and the operator side mostly at a desk.

What I explored

The working file carries the exploration rather than describing it. The v1.2 desktop and mobile pages sit alongside v2.0 instead of being overwritten, separate landing-page and marketplace directions from the designers on the team are kept as parallel frames, and an AI-integration direction was explored as its own branch before the assistant took the form it shipped in. The redesign was chosen against the version it replaced rather than in a vacuum.

What changed

A platform that works from both ends — customers discover and book services, businesses run those bookings and see where their customers came from — held together by one component library across booking, e-commerce and SaaS surfaces. Onboarding was smoothed so fewer people fall out before they have an account, discovery became proximity-aware rather than a flat directory, and the operator dashboard gave owners a reason to return between bookings. The product is live at bizinc.io.

Outcomes & scope

• Design impact

  +45%

  User activation on the dashboard

• Company-reported traction

  +23%

  Customer growth across the redesign period

  Bizinc's own figure, measured over the redesign period

• Design impact

  +12%

  Customer-to-customer service activity

Evidence

The redesign moved the numbers the business tracks: a 23% increase in customers across the redesign period, a 12%+ increase in customer-to-customer service activity, and a 45% improvement in user activation on the dashboard and lead-generation experiences, alongside growth in profile creation and stronger retention. The product is live at bizinc.io, and Bizinc's CEO has written a reference describing the platform rebuild and the progression to UI/UX Manager, published in full on the credentials page.

What I would improve next

The headline numbers tell me the redesign worked; they do not yet tell me which parts did the work. I would instrument the discovery path specifically — whether the open-now, distance and category signals on a business card actually shorten the route to an enquiry — and test the AI assistant with people who have never used the platform, since its value depends entirely on a stranger trusting it enough to start the conversation. On the operator side, I would want to know whether owners find the marketing surfaces inside the dashboard or work around them.

With more time

I would take the operator dashboard further into daily use — a genuine at-a-glance view of the day's bookings and enquiries rather than a set of management surfaces — and give the AI assistant a proper first-run state, since its usefulness depends entirely on a stranger starting a conversation with it.