Shine Krishna
Resume

Spark NZ · 2019 · 4 min read

Redesigning Spark's broadband sign-up so more people finish

Spark's online broadband journey showed plans before checking the address, and mixed moving house with new connections. I changed the order of decisions so people only saw what they could actually get.

  • Sign-up journey
  • Mobile first
  • Telco
Role
Senior Experience Designer
Team
Service designer, Wireless, Fibre, Copper and Billing squads
Timeline
2019, 3+ months
Platform
Responsive web, mobile first
The Spark broadband landing page on a laptop, with the headline "Unplan your broadband" and an address check field below.
  1. My part

    I led discovery and designed the new flows, after 73,834 people started and only 619 finished

  2. Key decision

    Ask for the address first, and split the journey by what people want to do

  3. Result

    Key changes shipped after I left: reordered pages and clear messages up front

01 · Overview

Almost nobody finished

In two months, 73,834 people started buying broadband on Spark’s website. Only 619 finished. Many of the people calling the contact centre had already tried online and given up.

In 2019 Spark had just renamed its broadband plans and was shifting focus to wireless broadband. It wanted to sell more broadband online, make moving house easier, and take pressure off the call centre, for both personal and business customers. The existing online journey wasn’t coping.

73,834

new customers started the online journey

619

made it to the end

Two months of the existing journey. Fewer than 1 in 100 people who started finished. Bars are drawn to scale.

02 · My role

What I owned, and who I worked with

I was the UX designer in the Broadband tribe, working with a service designer and with squads across Wireless, Fibre, Copper and Billing.

  • I led discovery, journey mapping, personas and scenarios.
  • I designed the new flows, from sketches to an interactive prototype, and tested them with customers.
  • We planned the changes with the squads, because they touched billing, network and product at the same time.

03 · The problem

Why people gave up

  • 10interviews with existing customers and prospects
  • 6personas, like new movers and price-conscious shoppers
  • 65%of traffic on mobile

Alongside the interviews, I looked at analytics and contact centre data, and listened to calls. The calls told the same story as the numbers.

The problems were less about any one screen and more about the order things happened in:

  • Plans were shown before the address was checked, so price and availability often changed halfway through.
  • Moving house and new connections shared one journey.
  • The full price and technology details only appeared after signing in or signing up, which is exactly where many people left.

The language didn’t help. People searched for “unlimited” or “high-speed fibre”, which didn’t match the new plan names. And the business wanted to lead with wireless, a technology most customers didn’t know much about.

A flow diagram of the old buy journey, with decision points for checking the address, signing in and paying monthly scattered across the flow.
The old buy flow. The address check and sign-in sat in the middle, after people had already chosen a plan.
A wall of colour-coded sticky notes on brown paper, organised into rows for research, buy and move, with persona cards on the left and screenshots along the top.
Mapping the Join stage with Spark’s JUCCI framework. Research, buy and move journeys, with persona needs and pain points against every step.

04 · Key decisions

Changing the order of things

We agreed four principles to guide every call: show clearly what’s available at the address, lead with the plans most relevant to each person, make choosing simple, and only reveal add-ons when they’re needed.

Challenge

Movers, new customers and people changing plans were all pushed through one complex journey.

Decision

Split the journey by intent at the start. The first question asks whether someone is joining Spark, moving house or changing plan, and each gets its own focused path.

Challenge

Plans and prices changed once people entered their address, which felt like a bait and switch.

Decision

Ask for the address first. People then only see plans they can actually get, with upfront costs and contract terms clear from the start.

Before

  1. See plans and prices
  2. Enter address
  3. Plans and prices change
  4. Sign in or sign up to see the full price
  5. Checkout

New connections and moving house share one journey.

After

  1. Join, move house, or change plan?
  2. Enter address
  3. Only plans available here
  4. One decision per screen
  5. Review with costs up front

Each intent gets its own focused path.

The same job, reordered. Steps marked with a dashed outline are where people got surprised. Steps with a solid mark are the new decisions that come first.

Challenge

Customers were asked to pick a technology they didn’t understand, like wireless or fibre.

Decision

Offer options based on how people use the internet, and cut plan cards down to speed, data, contract and price, with plain-language info icons for technical terms.

I also designed for the awkward cases: addresses with no coverage, wireless as a fallback, bringing your own modem, and switching paths halfway through. Each one got proper error states.

Three hand-drawn wireframes: the Unplan landing page with an address field, a plan selection page for a specific address, and a review order page.
Early sketches. The address field comes first, and the plan page is already specific to one address.
Six high-fidelity prototype screens: the landing page, a recommended plan for an address, landline configuration, accessories, the cart, and sign in or create a Spark ID.
The interactive Figma prototype we tested with target customers. One decision per screen, which mattered most on mobile.

05 · Outcome

What happened

Usability testing and feedback from internal teams showed a clear improvement in how easy the journey was to follow. We iterated on anything that wasn’t clear, then planned the changes into sprints with the squads.

This was my last project at Spark. After I left, the team made the key changes to the existing journey: they reordered the pages and put clear messages up front to reduce drop-off. Later, Spark redesigned the whole journey as part of a wider brand and product change, and I was told my work was its starting point. I don’t have a conversion figure to share.

Sprint planning walls with coloured cards for broadband journey features, grouped by team and sprint.
Sprint planning with the squads. Billing, network and product changes had to move together.

06 · Reflection

What I took from it

Plan names and steps need to be obvious, not branded. And working out what someone is trying to do at the very start saves a lot of complexity later on.