Marley Spoon

Delivering by suburb, not postcode: 20K+ new orders in Australia

  • SHIPPED
  • ACQUISITION
  • SALES FUNNEL
  • D2C
  • WEB

TL;DR

A new step in the sign-up funnel that checks delivery by suburb instead of postcode. In regional Australia, one postcode can be 300 km long and hold both empty areas and towns of 30,000 people, so the postcode alone couldn’t say who was within reach. The step runs on two brands, Marley Spoon and Dinnerly.

ROLE
Led the customer-facing design end to end, from discovery to design and testing.
TEAM
1 Product Manager
1 Product Designer
1 Lead Engineer
+4 Software Engineers
TIMELINE
Nov 15, 2024 → Dec 01, 2024
COMPANY
Marley Spoon Group
BRAND
Marley Spoon, Dinnerly
TOOLS
Figma, Jira, Confluence, Miro & Slack

Before

Dinnerly and Marley Spoon delivery steps saying they can't deliver to the address
Dinnerly and Marley Spoon: customers were turned away at the address step.

Unavailable

After

Dinnerly and Marley Spoon modals asking customers to pick their suburb
Dinnerly and Marley Spoon: the suburb check now appears right after the postcode.

Available

Discovery

Competitor analysis

I compared 5 meal-kit competitors in Australia: how customers picked their location, what they were shown, and how the system answered.

What I found

All five left the suburb for last: one more field in the delivery details, filled in by the customer alone. Asking earlier, and helping the customer choose, was the opportunity.

Kick-off

One session with product and engineering to agree on requirements and edge cases before scoping.

What we agreed

We dropped the quick fix of adding the suburb to the delivery details page and agreed to ask earlier. “No slots yet” and “no delivery at all” would get different answers.

Define

Customer journey

I mapped the sign-up funnel step by step to see where delivery was decided and where the suburb was asked.

Marley Spoon sign-up funnel before: postcode at step 2, delivery slot at step 3, suburb in the delivery information form
The funnel before: postcode at step 2, suburb in the delivery details form at step 3.

The suburb was asked too late

Customers got a yes or no on delivery from the postcode alone. The suburb was only asked later, in a long form, when the answer had already been given. It had to come right after the postcode.

User flow

I mapped every path a customer could take from a partly covered postcode, to settle the logic before designing any screen.

One question, only when it’s needed

Fully covered postcodes skip the new step. Partly covered ones ask for the suburb right after the postcode. Once the customer picks it, postcode and suburb are already filled in on the delivery form.

User flow: if the postcode is only partly covered, the customer picks a suburb, changes the postcode or closes; if it isn't covered yet, they can change the postcode
The new path: postcode first, then the suburb, only in partly covered postcodes.

Design

Hypothesis

If the suburb is asked right after the postcode, more customers can be served without losing sign-ups along the way.

One question, in a familiar shape

After postcode and email, a bottom sheet slides up and asks for the suburb. It only appears in partly covered postcodes, and it works with a familiar pattern that the customers already use on their phones.

The risk: one more step

An extra step can make people leave. So postcodes with a single suburb skip it, and the title says why the question is being asked. A wrong postcode sends the customer straight back to the postcode field, ready to type, not to the start.

  • Marley Spoon location step with postcode and email
    Location
  • Marley Spoon suburb bottom sheet
    Modal
  • Marley Spoon suburb search with the keyboard open
    Search
  • Marley Spoon suburb selected, ready to confirm
    Confirmation
  • Marley Spoon delivery step with slots for an available suburb
    Available suburbs
  • Marley Spoon delivery step with no delivery dates for the suburb
    Unavailable suburbs

Build for searchability

Suburb combo-box open above the iOS keyboard, with the close button and the list's bottom edge marked
  1. X / “Close button”

    The margin is 24px from the right and top edges. The X stays fixed, keeping its margin, when the user scrolls.

  2. Combo-box search and dropdowns

    The bottom line of this component must be 16px from the keyboard.

  • Dinnerly location step with postcode and email
    Location
  • Dinnerly suburb bottom sheet
    Modal
  • Dinnerly suburb selected, ready to confirm
    Confirmation
  • Dinnerly delivery step with slots for an available suburb
    Available suburbs
  • Dinnerly delivery step saying it can’t deliver to the address yet
    Unavailable suburbs

What the design actually moved

Checking delivery by suburb, right after the postcode, answered both sides of the problem. A two-week A/B test confirmed the new step didn’t hurt sign-ups, and the full rollout followed.

  • US$2MNet revenue

    In 12 months, from customers within reach.

    Before: they were turned away when their postcode was closed.

  • 5K+New customers

    And 20K orders in the same 12 months.

    Before: no way to order inside a closed postcode.

  • Nearly zeroUndeliverable orders

    Customers out of reach find out before ordering.

    Before: an estimated US$200K+ a year in failed orders.

Metrics are rounded for compliance.

  • 22%Became subscribers

    More than one in five customers who saw the suburb step.

  • 18%Closed the modal

    Most likely customers whose suburb isn’t covered yet.

  • < 1%Left at this step

    The extra question cost almost no sign-ups.

  • Question the default early

    The first idea was to add the suburb to a page the team already had. It was the cheapest fix, but it would have told customers too late. Comparing it with the customer’s path in the first session changed the plan before any screen was designed.

  • The 18% is the next question

    Almost one in five customers closed the panel. Most likely their suburb isn’t covered yet, and the MVP only let them change the postcode. The next step is to let them say they’re interested, so the business can see where demand is.

  • Share what the data shows

    Customers from these suburbs spent about US$98 per order, in line with the business case and steady across 12 months. I passed this to marketing, so paid acquisition could target those suburbs.

Unifying the leading Latin America credit & account limit management, now 70% faster

Customer support teams had to chase 150+ credit limit fields across a mainframe, screen by screen, while handling live...

Read case

Owning the seller experience of a delivery platform across LATAM & South Africa

A trustworthy, scalable experience. The store manager tool is part of a platform now live in 13 countries.

Read case