Using big data analytics to power customer lifetime value

It’s hard to see a customer’s lifetime value if you’re only taking a 30-day view of engagement.

This was a key conclusion drawn by HotelClub president, Nicolas Chu, who spoke at the first annual Data Strategy Symposium in the Hunter Valley. The leader of the online hotel booking website spoke about the company’s approach to utilising data to better understand the lifetime and retention value of customers across its network, as well as better target advertising.

Chu said understanding a customer’s lifetime value is a major headache for marketers. “The traditional metrics we use are based on short-term vision [30-day metrics], meaning you can’t see the customer who then comes back in six months to buy something,” he said.

One of the problems is that it’s really hard to attribute the value and sale to a customer to a specific channel. “Even if you’re not in the online space, you have the same issue when you’re managing an integrated campaign and you’re trying to understand the value of each channel,” Chu said.

“Everyone is spending millions to drive demand. We are at the stage where we have an idea of the value of each channel, but how we measure the online space is very short term and narrow.”

With traditional 30-day measurement, marketers tend to rely on multi-touch or last-touch attribution models to understand the value of each channel. But they also need to understand how to acquire customers who are more likely to spend money with their organisations in the future, Chu said.

To address this, HotelClub is compiling data over a six-month period to map out the customer journey, as well as better comprehend where credit should be attributed for the first purchase across its many communication and advertising channels.

“We’re starting to model our investments to recognise if someone comes back through another channel six months later to purchase something,” Chu said. “We also try to give some credit for new customers. You might be willing to spend more to acquire new customers in the first place, because you can increase your repeat rate long term.”

Through its longer-term data analysis, the company quickly discovered that some channels, such as affiliate websites, have a lower repeat rate and cost more of the marketing team’s money over time. This fact wasn’t apparent using traditional 30-day attribution metrics, Chu said.

“You can spend $2 to get $1 back online because you know that in six months’ time, you’ll get more than that back and have customers spending more with you over time,” he said.

HotelClub also learnt customers who first purchased a lower margin product have a higher value than originally anticipated. “We realised people are trialling us with a product, but after a certain period of time and if the experience was good, they were coming back to our site and the length of stay was longer, so the value was higher,” Chu said.

These multi-dimensional insights are now assisting HotelClub to make more strategic investments in communication channels and are helping improve that customer lifetime value, Chu said.

Another way HotelClub is leveraging big data is to understand the retention lifetime value of individual customers by estimating their future worth. To do this, the company performed analytics on a sample of 200 million customer as well as more than 200 variables, coming up with a model for predicting a customer’s value.

The lifetime retention information is now being used in real time in HotelClub’s call centre to determine how much time should be spent on on-boarding and assisting customers who call in, versus directing them to other channels for information.

“Without big data it’s extremely difficult – you can look at the purchase history and what they’ve spent with you, but it’s difficult to predict what their future value is,” Chu said. “Just because someone spent $100,000 with you, it doesn’t mean they’ll spend another $100,000.

“We now use this model to understand whether we should spend 15 minutes, 20 minutes or an hour to retain you because we know if you’re most likely to book with us in the future, or if at the end of the day you won’t be a loyal customer.”

Big data analytics: The new black magic of marketing?
Five big ideas to profit from analytics and big data
Big data, budgets and predictive analytics

This information is also being used when individuals log in to a site to serve up more relevant display advertising. Chu said the 10 per cent of highest lifetime value customers are only served HotelClub information banners, not third-party advertising, because that is of more value to the business.

Follow CMO on Twitter: @CMOAustralia, take part in the CMO Australia conversation on LinkedIn: CMO Australia, or join us on Facebook: https://www.facebook.com/CMOAustralia