The context

With about one million online users, one of the UK’s biggest energy suppliers wanted to provide its customers an easy way to give meter reads by installing smart meters in their home for free. However the booking for smart meters installation was low which led to either estimated reads when customers forgot to give meter read or incorrect meter reads. The project explored ways of increasing smart meter installation bookings.

My process

Heuristic analysis

• Three questions are asked to establish eligibility but it is not explicit to the user the reasons for asking these questions
• This question is not clear in specifying that if the meter is in a metal box the customer is not eligible to get a smart meter.
• Once the booking is confirmed there is no way to cancel or rebook on the journey.

• “None of these dates suits you?” The question is misleading to the user as their queries are not recorded in the system but encourages the user to register interest.
• This page is text heavy which can slow the booking process.
• “What is blocking your meter?” is an open-ended question and can confuse the user.

Data from Mixpanel

Mixpanel is a tool that gives data insights about customer interaction on the site (The following figures are indicative and not company data. This is to protect company data.)

• 66.5% users drops at identity screen
• 43.5% users dropped at eligibility screen
• 40% users dropped at booking screen

Fullstory observations

Fullstory is a tool that records user interaction with a website or portal. To explore user interactions I viewed 50 videos. The insights were:

• 1. Many users left when asked to input account number on the Identity screen
• 2. In Book an appointment page most users were missing the time slot section and clicking next button without choosing it, which lead to error warning. But users struggled to understand where the error was
• 3. Most user clicked on print this page instead of CTA

Data insights - Speech analytics

We had a speech analytics tool that would analyse calls from the customers. (The following figures are indicative and not company data. This is to protect company data.) The insights were:

• 26% Calls to customer services were regarding smart meters
• 12.3 min Average call for smart meter
• £ 330,000 Cost of calls for 4 months

User interviews

• “I missed the time slot and was clicking continue”
• “I got alarmed at choose how meter talks to us?.It felt like I had to read something important, like it is an important decision with charts and data”
• “When I answered what is blocking my meter, I was expecting the engineer to help me move the huge cupboard but instead when he saw it, he cancelled the install”
• “I missed the appointment as I forgot about it on the day and went out”
• “This booking journey was too long, I am not sure when it would finish”

Service map

To understand the service better and find more opportunities for improvement I ran a workshop to map the service of smart meter booking and installation. The team members included a marketing manager, customer service manager,smart meter engineers , development team, user experience designer and a product manager.

Research analysis

• 1. Many confusing eligibility questions that block the user to proceed further
• 2. Text heavy journey, that takes longer for user to read and can cause drop off
• 3. Difficult to find appointment slots
• 4. Difficult for customer to remember their appointment date
• 5. Not possible to cancel and rebook appointment on digital platform
• 6. Confusion around what an "In-Home Display" is, which leads to customer calling the customer service

Wireframes and User flow

User testing results

I tested the designs early on with 10 customers who had not yet booked a smart meter. There were few issues with copy like “Tell us how your meter will talk to us page.”.
It was not clear if meter will talk to the customer or OVO, which worried the users.

Final design

Impact

Measured in January 2020 - 2 weeks after launching the app (The following figures are indicative and not company data. This is to protect company data.):

• 1. Booking increased from 19% to 31%
• 2. Customers complete the journey on an average in 3.2 minutes from 4.1 minutes previously
• 3. 58% clicked on new calendar feature
• 4. 23% drop in appointment rate from 40% previously

Next steps

• Build the feature where users can say where their meters are located to avoid situation where installation is not possible due to technical limitations.
• Conduct A/B test to compare the user journey with and without smart meter location feature
• Monitor Booking appointment page to understand how to improve it further