Drawing from exisiting literature and findings provided from companies such as KANTAR and Google, as well as conducting our own user reasech. We were able to pull out some key findings to help support the need for this functionality on the site.

These findings only solidified anecdotal knowledge and evidence. A large number of potential customers are already engaging in comparing vehicles by themselves or from aggregator sites such as Autotrader. The desire from customers is there. So we want to take away or ease any pain points of manually comparing vehicles. There is also business requirements of keeping potential customers in our eco-system and attempt to limit their need to use competitor sites.

The majority of Arnold Clark's online traffic is on mobile devices, so applying a mobile-first approach to the design work is a must. Making sure we nail the experience at this limited viewport before scaling up to larger views. There is a particular challenge in creating a comparison experience at this limited width. We had a lot of questions to answer while designing this feature.

Q. Do we limit the number of vehicles you can compare at any time?
A. Yes. It was decided to limit the number of vehicles a customer can compare to 4. This was to combat any information overload that could be experienced with a larger number. We wanted to help the customer narrow down their choices, when they are making their purchasing decision. Giving them this limit of 4 would encourage them to start looking more specifically at which vehicles they might be committed to buy.
‍
Q. Do we limit the number of options visible on the screen at any one time?
A. With the limited viewport width, its not viable to have all of the comparison options be visible. However we felt that having a fixed number of 2 items on the screen at a time, only added frustration and rigidity to the experience.
‍
Q. What information do we show within the comparison table?
A. Ideally we would show all aspects of the vehicles such as their technical specification, standard specifications, and vehicles summaries. However due to data limitations, the MVP had to have only the vehicle summary to begin with. As the data we are supplied with overtime more data will be fed into the comparison table.
‍
What are our call-to-actions?
A. The CTA for the tool should be the prominent CTA on the vehicle page. For used cars and block cars we have in-stock this would be the reserve CTA, for new cars the enquire CTA would be used.
‍
And so on...

Different potential concepts were designed with varying levels of functionality and put through rounds of testing with users.

From developer investigation on the feasibility of the feature from their point of view, they estimated a lot of development time would be needed to build out more complex functionality. As a result we as a team agreed pushing out the MVP of the feature was a must. This would allow the developers to begin work on building out the code. Which could then have layers be added on top as we gathered feedback and data of its usage to then further design for.

The MVP design was given to user testers on usertesting.com. The aim of the testing was to gauge if the proposed design was usable and testers understood how to get to the comparison table from shortlist. Along with gathering input on its value when someone might be looking to purchase a car, and further comments.

The results from the testing were positive, with majority of the testers being able to discover the comparison tool in shortlist. There were no discoverable usability error or bugs. Which then allowed us to begin to release this new tool to Arnold Clark's web traffic, starting at around 20% of traffic before ramping up to 100%.

As the traffic level rose the number of interactions with the tool steadily rose as well. Tens of thousands of customers have used the tool since launching to full traffic in the end of April. The main feedback we are receiving from our widget is the fact the data provided is limited and they would like to see more specs.