Vanarama turns to chatbots and machine learning to better serve customers

Vanarama, the popular van leasing website famous for its catchy TV and radio advert jingle and its sponsorship of the National League football division, has brought its IT back in-house as it looks to deliver new capabilities like chatbots and machine learning to its customers at speed.

Speaking to Computerworld UK by phone, lead architect and CTO designate at Vanarama, Jamie Buchanan, who has been in the role for nearly a year now, talked about how he was tasked with bringing the technology stack back in -house after years of outsourcing, allowing the company to be more agile in how it delivers new capabilities to customers.

At the behest of CEO Andy Alderson, he was given the remit to build an IT team from scratch, implementing an Agile development methodology and moving away from the .NET framework to a more modern environment running Ruby on Rails.

Buchanan did admit that is was a challenge to find the right kind of tech talent to move out to Vanarama's Hertfordshire headquarters, however. "Being outside of central London meant that finding those skills tends to mean we have to search hard for people willing to take on a challenge and move out," he said.

The various Vanarama websites - for vans, cars, insurance and fleet - are primarily mobile, targeting customers "on their sofa" according to Buchanan. This means speed and simplicity are imperative if they are to drive up engagement and, crucially, sales.

Machine learning

Vanarama is working on a number of projects to try and target new customers and help close more deals once people enter its sales funnel.

This starts with smarter personalisation and matching, meaning customers coming to the site can only see vehicles they are eligible to lease, whether that is for insurance or affordability reasons. This sort of work generally relies on machine learning.

Read next: How was built on machine learning

Under the covers Vanarama is a price aggregation service, like Expedia is for hotels, but with the added complication of customers having to be approved to lease certain vehicles because of insurance or their credit rating.

Traditionally when leasing a vehicle a dealership would give you two options: their own financing option and a third party as a fall back. Vanarama instead searches across a panel of around a dozen funders to give applicants the best price via its algorithms.

Vanarama has also recently developed a pre-scoring tool for a quick credit check, allowing customers to search for relevant vehicles without having to go through the whole credit checking process.

"The idea is to give customers the ability to compare the market for finance funders and products instantly from their sofa, but without leaving a footprint on their credit file, a kind of pre-score," Buchanan explained. "We work with various data providers to gather a wide range of publicly available data, we combine this with our own in-house data models. We are working on these constantly."

By tightening up its pre-scoring ability Vanarama can create efficiencies by only sending credit applications that are likely to be approved to the funders. "With funders our aim is to send perfectly completed finance proposals for customers who have a high probability of getting credit, this will greatly reduce their cost to serve and enable us to drive more volume," he said.

Read next: Auto Trader uses MongoDB and machine learning to improve car valuation accuracy

Traditionally Vanarama's data science team, headed up by director of digital analytics and customer insights Damien Fortune, have turned to statistical modelling and BI tools from the likes of SAS and Qlik to drive these models.

Now it is looking to use Amazon Web Services' new self-serve machine learning platform SageMaker "to make our existing models and algorithms more accurate and near real-time," Buchanan said.

New technologies

Aside from tweaking its existing technology and modernising its stack, under Buchanan the company is looking for new ways to ease the customer journey for its users.

One way to really speed up the application process would be to use the newly public open banking APIs to allow customers to give Vanarama access to their financial information, saving them from filing paperwork to get an affordability assessment.

Read next: What is open banking? What does it mean for banks, fintech startups & consumers?

This is very much a work in progress. "We are working closely with our credit agency data partners and will most likely need to register as an AISP [Account Information Service Provider as approved by the Financial Conduct Authority] as part of those projects," Buchanan said.

Another new initiative is the introduction of a chatbot to help customers find the right vehicle by answering a set of questions, instead of the traditional filter method. Working with a specialist UK vendor, Vanarama was able to quickly build and launch the iVan chatbot on the site.

The chatbot iVan sits in the bottom right of the page but also pops up when you try to abandon the site to ask if it can help, it takes your name and presents a range of questions to help you find what you were looking for, before handing you off to an agent or emailing you a personalised quote.

"The priority is the self-serve journey for customers and making that as easy as possible," he said. "We would like to build more car chooser tools for people that don't have the time or interest to do the research and can do lifestyle searches."

Vanarama claims that iVan is already driving 100 leads per day through the website and Facebook Messenger.

Copyright © 2018 IDG Communications, Inc.

Shop Tech Products at Amazon