Case 03 · Tubr 0 → 1 · hospitality data · web app

Turning raw hospitality data into a product.

Tubr is a 0→1 analytics product for hospitality businesses. I partnered with the founder and a data scientist to turn a stream of raw point-of-sale data into something a bakery owner - and their staff - could actually act on: journeys, IA, dashboards and a brand, from scratch.

Drop a Tubr product screen
TUBR hospitality analytics
Role Lead product designer · 0→1
Timeline 2024
Scope Web app · manager + employee
Status Shipped MVP
0→1
idea to shipped MVP
2
audiences served
1
brand built from scratch
3
founder + data scientist + me
01 - Context
The why behind the work.

A bakery owner is sitting on data they can’t read.

Hospitality businesses generate a constant stream of sales data - every transaction, every item, every hour. But it lives in a point-of-sale system and a pile of spreadsheets, and reading it means flicking between tools after a long day on your feet.

Tubr had the data pipeline and a pilot bakery, but no product. My job was to turn that raw signal into something a busy owner would open every morning - and that their staff would understand without any training.

02 - The problem

Raw data is not the same as a decision.

The challenge was three things at once: make the numbers mean something, serve two very different users, and do it under a brand that did not yet exist.

Raw data

Numbers without meaning

Sales totals and transaction counts don’t tell you what to do. The product had to translate data into a next action, not just a chart.

Two audiences

A manager and their staff

The owner needs forecasting and the week in full; an employee needs one timely nudge mid-shift. Same data, opposite needs.

Zero brand

A product with no face

There was no identity to design within. The look, the voice and the name had to be built alongside the product itself.

03 - My role

Lead designer on a blank page.

I led design end to end alongside the founder and a data scientist. I ran the discovery, mapped the journeys, defined the information architecture and the split between manager and employee, designed the MVP, and created the brand - turning a data pipeline into a coherent, usable product.

What I owned
Discovery & research User journeys Information architecture Manager + employee UX MVP UI design Brand & identity Data visualisation Prototyping
Team
Founder · data scientist · me
Disciplines
Research · product · brand
Surfaces
Responsive web · manager + employee
Tools
Figma · founder + data partnership
04 - Process

From a pipeline to a product people open.

01 Discover

Walk a day in the bakery

Founder sessions Journey mapping Contextual research Opportunity framing

I mapped the owner’s day end to end - point of sale, ordering ingredients, planning, managing staff - capturing the moments of truth, the pain points and the opportunities at each step. The recurring theme: people were wasting time and guessing where data should have given them an answer.

Journey map
Journey map - the bakery owner’s day, with pain points and opportunities at every phase.
02 Define

Two products in one

Sitemap & IA Role definition Brand foundations Design system

The journey made the split obvious: a manager experience for forecasting and the full week, and an employee experience for a single in-the-moment nudge. I defined the sitemap and IA around that divide, then built the brand foundations - colour, type, spacing and icons - so the data could feel friendly rather than corporate.

Sitemap
Sitemap - the manager and employee experiences mapped from one data source.
03 Design

Plain language first

UI design Data visualisation Prototyping Brand application

The MVP leads with insight, not charts. The home is a weekly summary; a plain-language headline tells you, in a sentence, how the week is going; the factors that contributed sit underneath, and the deeper charts are there when you want them. The employee app strips this right back to a single, timely prompt.

MVP
MVP - the manager dashboard and employee views, built on the new brand.
05 - Key decisions

The choices that made data feel human.

A handful of decisions turned a data feed into a product a non-technical owner would actually use.

01

Split the manager and the employee

Same data, opposite needs. Two tailored experiences beat one compromised app that serves neither well.

02

Lead with plain language

A sentence - you’ve sold more than usual by now - before any chart. Insight first, evidence second.

03

Show the why, not just the what

Every headline number is backed by the factors that drove it, so the owner can act with confidence.

04

Lightweight for staff

Employees are on their feet - their view is one glanceable prompt, responsive right down to a phone screen.

05

A brand from zero

Purple, friendly type and a warm voice - so dense analytics feel approachable, never intimidating.

06

Weekly summary as home

The week, not the day, is the unit a bakery plans around. The home opens on the rhythm that matters.

Tried

One app serving both the owner and their staff.

Chose

Two role-specific experiences from one data source.

Why →

A manager and an employee want opposite things. One app would have been busy for staff and shallow for owners.

Tried

A dashboard of charts and tables - the analytics default.

Chose

A plain-language headline first, charts underneath.

Why →

A time-poor owner needs the takeaway, not a data-exploration tool. Charts support the story; they are not the story.

Reading one screen - the analytics dashboard
Tubr analytics dashboard, annotated by Hayley
01
Insight before evidence

The chart is the hero - largest, top-left - because it answers the daily question, “am I on track?”, in seconds. The plain-language takeaway sits above the chart, not below it.

02
Honest uncertainty

Solid lines are actual, dotted are predicted, the grey band is the confidence range. I kept the forecast visibly uncertain - hiding that a projection is a projection would make the tool feel dishonest.

03
Clear logic labels

Three metrics, three fixed colours, used consistently so the legend is never re-learned. Finance terms - CM1, CM2 - are defined inline, so a non-finance owner learns them without feeling tested.

04
Progressive disclosure

Simplest view by default, depth one tap away via the tabs, the “why” narrative collapsed below. Built for two readers at once: the one who trusts the headline, and the sceptic who checks the reasoning.

05
Read it by colour

Predicted and actual sit side by side with a green-up / red-down chip, so “did we beat the forecast?” reads by colour before you read a number. Cards are a uniform size on purpose - like compares with like.

06
Personalisation

The “set your weekly goal” CTA is the visible entry point to personalisation - from there the platform tailors its recommendations to the user over time.

07
Hierarchy of info

Individual order detail sits last - it is the least decision-critical, so it is available but never competes with the headline metrics.

06 - Craft & system

The brand that made data feel friendly.

With no identity to start from, I built the brand alongside the product - a confident purple, warm type and plain-spoken copy that makes analytics approachable for a non-technical owner.

Tubr design system raw data → decisions
You’ve sold more than you usually do.
Grotesk · display + UI

Two audiences, one product. A manager reads the week at a glance; their staff get a single, plain-language nudge - never a wall of charts.

Numerals · data · £1,800 · 110%
Purple
#865BFF
Primary · brand
Aubergine
#341C42
Ink · deep
Orange
#FE6F34
Accent · alerts
Lime
#8AF33F
Positive · growth
Lilac
#BCA4FF
Tint · surfaces
Component & token library · manager + employee
Weekly summary KPI rings Charts Plain-language insight Category breakdown Targets & breakeven Manager view Employee view Status badges Spacers Icons Buttons
Product screen
Product screen
Product screen
07 - Outcome & impact

A data pipeline, turned into a real product.

Tubr went from a raw data feed to a coherent MVP - a manager dashboard, an employee experience, a clear information architecture and a brand, defined and designed alongside the founder and data scientist. It gave a non-technical owner a reason to open the product every morning.

What I'm carrying forward

Working at 0→1 with a founder and a data scientist taught me to design from the data outward - to start with the signal that exists and shape it into a decision, rather than designing screens and hoping the data fits.

0→1
from raw data feed to a shipped MVP
2
tailored experiences - manager and employee
1
brand and design system, built from scratch
MVP
shipped with a pilot bakery
Next case study
Unifying a European fintech platform
Qwist

Building something? Let's talk.