Drinks Tendering Platform

The goal was to design a tendering process that brings bars and drinks producers together. There are two distinct user groups:

• Bar owners create a tender for a drink they need to be supplied.
• Drinks producers submit bids to those tenders, with their own products and pricing conditions.

A marketplace has two sides, and a mistake on either one breaks the other. Before designing anything I needed answers to questions the brief didn't cover: Does something like this already exist, is there real demand, and what does the client assume to be true?

I proposed five steps, each building on the last:

1. Empathize: Learn how the market and its users work today.
2. Research: Interview both groups about goals, desires and pain points with the methods they use now.
3. Ideate: Generate user flows and wireframes from a wide pool of ideas.
4. Prototype: Design solutions to a mid-to-high fidelity stage.
5. Test and learn: Test at least two options against each other (A/B or multivariate) rather than committing to the first idea.

I prepared the questions I would ask at the kickoff meeting:

• Is there an existing or similar product? I'd ask for a list of direct and indirect competitors.
• How do we know there is a need right now?
• What assumptions does the client already hold, and what are their goals?
• What should the research goals be, and what do we want to learn?
• How will success be measured?

I split the research questions into must-knows and nice-to-knows, so that time in interviews went to the questions that could change the design.

A risk I flagged early: Research can produce unexpected results, and the real problem may differ from the original idea. I recommended preparing the client for the possibility that the scope would change if the evidence pointed elsewhere.

I wrote separate interview guides for bar owners and drinks producers. Each guide asked participants to:

• Describe, step by step, how they buy or supply drinks today
• Demo the process on screen if possible, so we see what they do rather than what they remember
• Name the tools they use and how long the process takes
• Say what frustrates them most and what delights them

To make sense of the results I planned a competitor analysis and an affinity map, plus a 2x2 matrix comparing business goals and user goals to show which processes work and which don't.

• How are winning bidders notified, and how do unsuccessful distilleries find out?
• Which platform should come first: Mobile, tablet or desktop? Usage data should decide this.
• The concept will probably need two to four further use cases, mainly how producers place a bid and create an account, and how bar owners create a tender and set up theirs.
• Once the concept is clear, a design system would help test buttons, colours and calls to action consistently.

I designed the first screens around the tasks users return to most often.

I recommended a persona for each target group, with goals, desires, pain points and basic demographics, so the whole team would know who they were designing for. I started with Julia, a fictional bar owner struggling to source suppliers.

• Run the interviews and update the personas with real findings.
• Test at least two versions of the filter pattern.
• Design the bid and tender-creation flows, which carry the most risk.
• Define success metrics before launch so that results can be measured.