My role: Lead UX Designer & Frontend Developer · 2019–2025
The Modern Milkman

Turning a broken sign-up into one end-to-end journey

How replacing a disconnected browse-and-hope experience with the Milkround Builder, a guided five-step flow, connected new customers to their first order in a single session and drove consistent week-on-week sign-up growth.

01 - The Problem

Sign up, then hope for the best

The original Modern Milkman sign-up flow did one thing: created an account. What happened next was left entirely to the customer. They’d land on the site, browse the products, and, if everything went well, eventually build a basket and check out. In practice, many didn’t.

 

The gap between creating an account and placing a first order was wide, and a significant number of new customers never crossed it. The experience had a structural flaw: sign-up and first purchase were treated as separate events separated by an open-ended browsing session with no guidance, no context, and no momentum.

"We were handing customers the keys and pointing at the shop floor. No introduction to how the milkround works, no help building a basket, no reason to check out today rather than tomorrow."

t a company built around subscription and habit, that first order mattered enormously. Customers who placed it were far more likely to become long-term subscribers. Customers who didn’t rarely came back.

02 - Discovery

Where people were dropping off, and why

The data showed a clear pattern: drop-off was highest in the space between account creation and basket building. New customers were arriving, looking around, and leaving without ordering. The site assumed they’d know what to do, but most people signing up to a milkround delivery service for the first time had no frame of reference for how it worked.

 

Several questions went unanswered during the original sign-up experience:

Questions left unanswered

Answered by the builder

The insight was straightforward: if we could answer those questions as part of the sign-up itself, and take the customer all the way through to checkout in one guided journey, drop-off would fall and conversions would rise.

03 - Design

One flow. Account to checkout without leaving.

The solution was a five-step progress-based sign-up flow that guided new customers from account creation to their first completed order in a single session. A progress bar at the top made the journey visible from the start, customers could see exactly where they were and how far they had to go.

Step 1. Account

Standard account creation, name, email, address, password. The progress bar set expectations immediately. Five steps, clearly laid out. Customers could see the finish line from the start.

Step 2. Delivery

Customers were shown their local driver's delivery schedule, for example, "Your local driver delivers by 7:30am every Tuesday, Thursday and Saturday", and asked when they'd like deliveries to begin. This step made the service feel local and tangible, and committed the customer to a start date before they'd built their basket.

Step 3. Milkround Builder

The centrepiece of the new flow. Customers answered three questions: how many people are at home (1–2 / 3–4 / 5–6 / 7+), when deliveries should begin, and which products they'd like. Household size informed suggested quantities. Product selection gave customers a personalised starting point rather than a blank catalogue to navigate alone.

Step 4. Your Basket

Customers reviewed their suggested milkround, adjusted quantities, and added or removed items. The heavy lifting had already been done, this step was refinement, not discovery. Arriving at a populated basket rather than an empty one made the next step feel natural.

Step 5. Checkout

Standard checkout. The customer had gone from nothing to a completed first recurring order in one sitting, without leaving the flow, without navigating the site, without having to figure anything out themselves.

04 - Constraints & Trade-offs

Shipping an MVP, not a masterpiece

The first version of the Milkround Builder was deliberately scoped to test one hypothesis: would guiding customers through a structured flow improve conversion compared to the freeform browse-and-build experience? The answer came back quickly in the data.

From there the team iterated, adding, removing, and refining elements based on real usage patterns. Some steps that seemed useful in design proved unnecessary once customers were actually moving through the flow. Others needed friction removed. A few assumptions about what customers needed in the builder turned out to be wrong in practice, and the data made those corrections clearly.

 

The phased approach meant every subsequent decision was grounded in evidence rather than opinion, which is the only environment in which good product iteration actually happens.

05 - Outcomes

Consistent growth from the moment it launched

Sign-ups increased week on week from launch. Connecting account creation directly to first checkout, without the customer needing to navigate away, come back, find products, and decide to proceed, proved to be the meaningful change the drop-off data had been pointing toward.

Week-on-week sign-up growth

From launch, new customer sign-ups increased consistently. Removing the gap between account creation and first order removed the biggest single point of drop-off in the acquisition funnel.

Drop-off reduced at the critical moment

By the time a customer completed Step 3, they had a basket, a delivery schedule, and a start date. The commitment was made. The distance between sign-up and first checkout was eliminated rather than shortened.

A better first impression

The builder educated customers on how the service worked as they moved through it. By checkout, they understood their delivery days, their driver's schedule, and how their recurring order functioned, reducing post-sign-up confusion and CS contacts.

A scalable onboarding foundation

The five-step framework became the structural basis for future sign-up improvements, including the later integration of the membership step, because the architecture was already in place to add and remove steps as the product evolved.

06 - Reflections

What I took away from this project