top of page

Marshmallow App

Managing our app strategy and initiatives to deliver an industry leading customer experience.

miles app.png


Marshmallow focuses on delivering best in class customer experience through their insurance app. All initiatives within the business are aimed at being app led due to the engaging relationship we are looking to build with our customers.

Over the last 12 months I have led the product strategy for a host of initiatives across product squads in order to scale up our offering within the app. We currently have one of the highest app/play store ratings in insurance sitting at 4.8/5 which shows we must be doing something right. The following projects demonstrate core initiatives we have delivered over the last year.


- Strategy lead 

- Product design

- Product management


12 months+

The background

There were a host of new projects that were planned across our product squads which cover claims, retention, growth and risk mitigation. However there was no consistent oversight of the how these would holistically be managed across our end to end customer experience. I led the working group that would focus on the following areas:

  • Product strategy of all app based initiatives in order to ensure clear scalability

  • Information architecture reviews 

  • Setting all tracking events and dashboards to measure impact of releases

  • App & Play store reviews & updates

  • Setting up customer feedback loops

  • Migration away from web views to fully native experiences

The current experience

When I joined at its core the app showed information about a users policy and allows them to make changes and seek basic help for common queries. I will be covering the following initiatives to demonstrate how we have evolved the app to deliver measurable value to the business. 

  • Scaling our information architecture to view new features and other insurance products

  • Allow users to view multiple policies intuitively 

  • Improve access and ability to make changes 

  • Reduce live chats to operations and increase self service in the app

Customer mindset & Sign-up

App overview (Initial features)

App overview (Future feature view)

Workshop alignment

From our H1/H2 planning process each product squad is aware of the features we will be delivering throughout the year. In order to ensure we are aligned internally on how this will fit within the app architecture I setup a series of workshops.


The goal here was to agree on a set information architecture to allow engineering to write up tech plans and unblock research/design which a clear vision of the product mapping.

Workshop outcomes

  • Agreed direction on future IA and feature structure throughout the app

  • High level prioritisation of release roadmap for the app across squads

  • Cross-team visibility and alignment of upcoming initiatives 

  • Buy-in from senior leadership on engineering work involved to move this work forward

Analysing customer behaviour

Before moving any further into the initial design phase I worked closely with engineers to understand current usage across the app. The purpose here was to understand:

  • What were our most used feature areas, who by and why was that?

  • What areas were not being used today, did these need to be more intuitive or actually have no user value?

  • What were the most frequently requested queries to our COPs teams that could be self service in the future? 

Key learnings

  • Customers were not aware of MTA's available and could not locate them in the app causing thousands of unnecessary calls to customer service (Also validated through user testing)

  • The existing help section had very low usage only 2-3% of active users were visiting, this could be due to a poor value exchange to users

  • Viewing policy data and making changes made up 50% of journeys, we needed to facilitate these journeys in a more intuitive way

Design & Testing

Over the following 4 months there were multiple design projects that were focused on to deliver value alongside the work involved to reach the app vision that we had defined. The key design goals here were:

  • To improve general self service in the app for customers to make changes and get help

  • Create a global dashboard that can accommodate viewing multiple policies and insurance products

  • To allow customers to view their policy information more intuitively to reduce queries to live chats

The background

User testing & Findings (Dashboard + Manage policy)

User testing & Findings (Help & Support)

MacBook Air - 14.png

Core design features

Below you can see the core features that were built through the app vision initiatives. If you click through on the images you can view how the designs have changed from the previous stage. Ultimately this allowed us to  future proof the app based initiatives due to having a clear design direction.

Global dashboard


The improvements include adding a scalable carousel based policy selector, shortcut actions, simplifying the core navigation and bringing through vehicle based insights.


  • Weekly active users increased by 9.3%

  • MTA's in app increased by 7.1% (Reducing request through live chat to agents)

  • Interaction with our new feature releases increased by 14.4%

Policy overview

Manage policy

Grouping themed information and utilising progressive disclosure to reduce the length of the page by 200%+. Reducing this area to two tabs and improving heading content to be explicit with users on what they can find.


  • Reduction in queries to agents by 6.3% for miscellaneous questions in locating information

  • Increase in customers updating their vehicle and home address by 7.4% (Improving accuracy of user profiles)

Help & Support

Help area

Improvements include updating the categorisation of existing FAQs, introducing a keyword search to locate articles. As well as recording article effectiveness to track areas in need of development.


  • Weekly FAQs viewed in app increased by 13%

  • Reduction in weekly live chats started through the app by 10% (120+)

  • An average of 75 article rating responses recorded a week to help understand what FAQs to improve

Building feedback loops

We observed that a lot of customers would share requests and feedback to our cops team that would rarely pass its way down to the product squads in a structured format. We built out a new intercept feedback feature in the app for the following reasons:

  • To begin to analyse the sentiment to identify any themes we could start addressing

  • To prompt users to give a rating on the play/app stores if they were having a positive experience

  • Our main KPI was around number of positive/negative reviews on the stores and trustpilot (App ratings increased from 4.6-4.8)

Screenshot 2023-05-08 at 18.21.17.png

App store rating

Qualitative feedback

Key learnings

Planning ahead but not too far..

It was important to understand what features were upcoming and needed to be accommodated within the app to set the initial strategy. However its impossible to understand and predict changes in direction and roadmap updates along the way. Learning here was to not spend to much time planning/scoping into the detail.

Design vision and impact

It had been challenging to get buy in from senior leadership for time to be allocated for general improvements in the app as mentioned above in the navigation, policy and help areas. By sharing a combination the app vision features, the impact the changes would have on reducing queries to agents. Alongside being able to maintain more accurate customer data helped build a compelling rationale.

Technical limitations

We had an shortage of back end engineers for long periods in the mobile team so we had to look for FE initiatives to work on until that resource was available. Working with engineers early to help build out a suitable roadmap that would not have any dependencies was key to releasing continual iterations

bottom of page