Interview: Digital tech fuels AutoTrader’s drive into the future

Transformation in stages

The digital transformation has happened gradually. Over time, the company introduced a mobile business, a data business, and now offers platforms for other organisations in its sector to run their businesses from.

Coe says all these developments together form the digital business. “It’s not like those things are sequential steps, they all still exist together. We are still a web business, still a data business, still a mobile business, and we are still adding and augmenting,” he says.

Behind all these projects and plans sits a team of tech staff, which currently accounts for a third of the organisation’s total workforce of 1,200. The tech team comprises engineers, designers, product specialists and data scientists.

“For the past 10 years, we’ve been developing our data and data science capabilities so that we’re able to provide highly scalable, highly performant, AI-driven recommendations on the retailer side”

Nathan Coe, AutoTrader

It’s a demanding and dynamic work environment. AutoTrader’s website is visited by about 1.6 million people a day, who spend a combined 500 million minutes a month on it.

“These aren’t glory stats,” says Coe. “There are 30,000 retailers that use our platform, and on our website, we’ve got 440,000 or so cars, which – unlike products on most e-commerce sites – change every day and are all unique.”

This is a challenge for IT staff, who do more than 1,000 software releases every week, using agile techniques.

“We were early into agile, moving into it about 10 years ago. Now [we operate] continuous delivery and continuous integration,” says Coe. “We’ve got a group of engineers that focus on the infrastructure. That means we can adapt our software, develop our software, and deliver really quickly and reliably. And importantly, we can respond and make many small changes, which reduces the risk of release and increases the velocity of innovation.”

Collaborative culture

It is rare in this day and age for a company so reliant on technology not to use IT outsourcing suppliers. But, according to Coe, this is the very reason the company doesn’t use them.

“We’re very, very principled when it comes to our technology and the way we do it. One of those principles is that because what we do is technology, we want full-time AutoTrader staff working on it. They are people with our culture and our values, people who stay and are not just here for a project,” he says.

“We don’t use any outsourcing providers at all. We have UK staff working in the office more than they are out of the office, in an in-person, highly collaborative environment.”

The organisation’s technology team is based mainly in Manchester, where about 750 of its 1,200 staff are based. It also has an office in London.

IT recruitment at AutoTrader is heavily focused on finding staff in the early years of their careers, says Coe. “We pay very well, but we don’t want to just say, ‘We pay the most so come and work with AutoTrader’, because the proposition is much bigger and better than that,” he says.

Platformisation of fragmented industry

AutoTrader also wants its retail customers and other partner firms to benefit from its IT team’s output. It receives huge volumes of application programming interface (API) requests from organisations that use its platform to run their business. These include garages, lenders, retailers and insurers.

Potentially, AutoTrader’s biggest IT project, according to Coe, is known as “digital retailing”, wherby it offers small organisations in the car dealing sector the opportunity to access the latest technologies through its platform.

“All the technology that we build to run AutoTrader, we now allow others to use to run their business off,” says Coe.

He says this one of the company’s “big goals” at the moment. “Now we’re saying, ‘Actually, you should be able to do more than a transaction on AutoTrader, a brand that you trust’.”

Coe says many industry customers are small and don’t have big technology functions. “This is not us taking interest in selling the cars, but allowing our retailers to get all the advantages of selling in an omnichannel way,” adds Coe.

Perfect problem for AI

While the company is finding answers for its customers, it is creating problems for artificial intelligence (AI) to solve.

AutoTrader is currently engaged in multiple AI projects and, according to Coe, it is a good testing ground for the most discussed technology of today. He says having 440,000 cars on its site, which change frequently, is a “perfect problem for AI” to solve.

“Our products are all unique, so you need to use models to decompose the data and rebuild up to, for example, value products and tell our sellers how long they will take to sell,” says Coe.

But he says while AI is now being talked about by everyone, from big businesses to schools, there are different varieties, much of which is not business-ready. “I think it’s important to rewind a little bit, because AI to people often means plugging ChatGPT into something to write an email, which we think is maybe interesting, but not a particularly intelligent use of AI. “

AutoTrader’s IT team is currently engaged in three separate projects that harness AI technology, but this is nothing new to the company.

“For the past 10 years, we’ve been developing our data and data science capabilities so that we’re able to provide highly scalable, highly performant, AI-driven recommendations on the retailer side,” says Coe.

“We’ve been using AI, mainly machine learning, for years, and we do that because all those vehicles are different and people want to know what the value is. Car dealers want to understand what’s on the cars, what makes them stand out and how quickly they can sell them.”