"Re:infer gives us a clear signal of what conversations result in a purchase and which don’t. At the scale we’re operating at, this provides hard, statistical evidence on what is going on in these conversations and attributes that to the most effective sales behaviours. We can finally tell our partners what behaviours are driving positive outcomes."
<small>John Veichmanis, Chief Operating Officer at carwow<small>
The challenges
The new and secondhand car market is experiencing a period of rapid acceleration towards online, enabled by trends such as electric vehicles, greater connectivity and consumer preferences. This steep incline in online adoption has only been accelerated by COVID-19, and it’s created a pressing need for car sellers to learn and adapt quickly.
For carwow and its partners, the most important question has been what is the best way to sell a car online. OEMs and physical car dealerships have long been adept at in-person selling, but carwow is still discovering what makes an online sale successful.
carwow was determined to help its partners understand how they can deliver the best online customer experiences, close more sales and generate higher profits. This was crucial for helping the carwow marketplace to grow and become more profitable.
carwow’s customers use its platform to compare offers and buy new cars directly from trusted dealers. They can also upload an appraisal of their old car and dealers will then bid to buy it. This gives the company access to a base of 12 million customers and over 3,000 partners interested in researching, buying and selling cars.
The company wanted to explore using AI to understand these relationships and increase revenue for dealerships by giving them actionable insights to help optimise enquiry conversion. With huge volumes of comms and sales data, they needed a means of bringing them together to identify the drivers of successful customer-dealer interactions.
carwow processes 20 million customer enquires a year and sells approximately 13% of all cars in the UK.
The solution
Once a customer has settled on a car, the carwow platform enables them to call or message the dealer for more information, to configure the car, or to set up a test drive. As a replacement for the typical face-to-face conversation in a car dealership, carwow recognised that these messaging interactions were crucial to a successful sale.
carwow manages 20 million of these messages a year, across three markets and sells approximately 13% of all cars in the UK. The scale of communications at the company was enormous and could contain valuable insight into customer motivations and preferences. However, these interactions only produced unstructured data. carwow lacked a viable way to mine and analyse this information at the scale and speed required.
That’s why carwow adopted Re:infer, a UiPath company, and its Conversational Data Intelligence platform. Re:infer provides a powerful, no-code NLP platform that connects into all business communications in real-time. Users can rapidly train NLP models that give them complete insight and coverage over their customer conversations. This allows them to assess the quality of their customer experience and service in unprecedented detail.
carwow leveraged Re:infer to rapidly extract the customer intents and sentiments from the sales conversations happening on its platform. Re:infer provided them with the tools to do this quickly and cost-effectively. As a result, carwow gained valuable insights and statistical evidence of the buyer and dealer behaviours with the greatest impact (both positive and negative) on sales conversions.
The results
Using Re:infer, carwow discovered over 120 intents that had a predictive impact on sales conversion. This included factors such as agent responsiveness, the mention of competing offers by customers, and the frequency of dealer follow-ups.
carwow now had the hard data to recommend best practices to its partners, and the means to discover entirely new ones. These insights enabled the company to create engaging marketing content, and share insights with dealers to ultimately improve enquiry conversion.
From the intents extracted, carwow could make reliable predictions on how specific sales behaviours would influence a sale. These included:
- Establish needs - 20% higher conversion
Understanding what people want helps you offer them better alternatives.
- Manage expectations - 40% higher conversion
Let people know how long things are going to take. When things take longer than expected, them them updated.
- Make it personal - 10% higher conversion
Responses that are personalised and tailored to the buyers specifics.
- Don't be pushy
Sending offers to early decreases conversions by 15%. Talking about financing unprompted at start of interaction decease conversion by 20%.
- Buyers that reference competing offers are 140% more likely to buy
carwow has also been able to roll out real-time automated assessment of agent performance across all touchpoints and for every stakeholder. Every dealer, sales manager and agent can see how they are performing against these ideal sales behaviours, receiving feedback that is accurate and immediately actionable.
This gives carwow a scalable means to provide coaching and actionable insights to improve dealership enquiries, without adding lots of additional headcount. The system adds to their value proposition, providing better value for money for dealerships. Consistent and efficient enquiries also mean a better customer experience. Carwow now has a repeatable approach that it can scale to all regions with almost no additional staff.
“Re:infer conversational intelligence gives our teams that work with our dealer partners every day the confidence to provide a great service. They love sitting in the tool and understanding how they can provide excellent, actionable advice and support. That empowerment is really important to us. Having our people and the AI working together provides a better experience for customers but also, most importantly, to scale.”
<small>John Veichmanis, Chief Operating Officer at carwow<small>
