In a recent InternetRetailing webinar, How AI and predictive analytics combined with marketing automation can increase your customer lifetime value, Matthew Kelleher, chief commercial officer at RedEye shared insights into how brands including Travis Perkins and Footasylum used predictive modelling to boost conversion rates and revenues. Here’s a bulletpoint overview of the event.
Matthew Kelleher started by saying that marketing automation was at a real turning point.
“The potential bringing together of three critical elements means we are at a point where we can start truly thinking about truly leveraging the massive potential of marketing automation as a marketers tool, to improve communications and make multichannel communications more relevant to customers.”
- Customer data platform: bringing all forms of data together into a single customer record. Allowing retailers to get beyond data as a barrier to their businesses.
- Predictive analytics/machine learning
- AI and capability for you to do a lot more with your data than you can today.
“All coming together to make marketing automation a far more exciting capability with a lot more potential.”
• Talking about the potential of AI and predictive machine learning for some time. “There’s a lot of hype but a lack of outcomes” Lots trying to use it but not talking about the benefits of using it. It’s a significant change that might approach the way that businesses approach marketing.
• Big market analysis players have not been wholly critical but they’ve added to the conversation.
Recent Gartner CMO report suggested decrease in investment because marketers wanted to see results and not getting them.
Graph: Predictive analytics drops past hype and into the trough of disillusionment. Talking about it for years now.
• Are we at the vanguard or behind? Last year analysis of B2B marketers found many didn’t have a firm handle on the difference between AI/machine learning/predictive modelling. There is a lot of confusion in the marketplace: I’m pitching a five stage approach that we’ve taken and ideas and approaches from that.
• Data: issue and problems of data to reduce that as a barrier remains a challenge. Customers are more difficult to understand. Marketplace has been shifting around. Next iteration of landscape shift is potentially whatspap question – how are marketers going to deal with it, how get data back and into the landscape. Do it at a holistic level. How tie hat back into an omnichannel strategy?
• Proliferation of channels and potential touchpoints means getting to understand a customer and predict what they’re going to do, understand where they are on lifecycle with your business becomes hard to understand and predict. Sense that we can’t pick up. No way that anything but the biggest organisations are able to leverage enough marketing resource to track what’s going on.
• Speed of change: brands losing touch with customers. So much talk about declining loyalty, sense that being out of touch in the current landscape that we are not in charge of our marketing spend. Consumer loyalty dying away. A new approach may help.
• Customer data platform: one of the problems that the CDP solves is inability to enter or retain first person customer data. Online behaviours: complexity of pulling together online and offline view of customers as perhaps the biggest task: behavioural data, consumer interest in your brand, engagement: if you know who the customer is, they’re a VIP or a prospect. But how are they interacting with your brand?
• CDP pulls all that together and pulls together the model.
• Analytics: Forrester graph: shows potential importance of analytics in the model. Built on analysing patterns of behaviour and feeding that into multichannel campaigns.
• Predictive analytics: how RedEye is putting it to work.
• How AI to help make sense of all that data.
• Letting data decide: we’re all going to have to get used to this. Retailers want platforms to take the guesswork out of it: to identify behaviours and then get on and send messages to those people. How create a segment from the offer.
• Comes back to customer lifecycle. At its core, there are a few core stages. Prospect/multibuyer/churn – can now send campaigns to people behaving in certain way. Can also focus on people at key stages, set up strategies and turn dials in the right direction through specific strategies aimed at those segments.
• Finding person doing a particular action in customer database, traditionally, has been like finding Wally and virtually impossible. AI and predictive analytics begin to answer the question – no more irrelevant offers to someone who you don’t know where they are in the customer journey.
Customer lifetime value
• How do you measure it? Don’t let it be a barrier, find a simple way to measure it. Predict into the future then track performance against it.
• Don’t get hung up on definitions.
• Upsurge in focus on this metric. Beginning to get to a point where we can measure it: insight into customer behaviour. Plays out into we can measure it, influence it, stretch it. If we believe we can influence and improve customer lifetime value then surely we are touching on and improving customer loyalty. GRAPH
• Research: Only 2% see their approach to customer lifetime value as mature, while 77% see driving customer lifetime value as a medium or high priority. People are beginning to think about how to pul these together.
• GRAPH showing how it works. The more data you put in (from personal, demographic, mobile and device data through to app data and lifestyle data), the more valuable.
• Right now email still most important tool. We apply multichannel treatments but pst of the impact is driven from email. Also the most measurable still. Though push still important.
Case studies
Hotel Chocolat: predictive engagement: sent 40% fewer emails, revenue grew by 25%
Allbeauty: engagement enabled to lift lifetime value by 18.9% in six months.
Footasylum: convert model: +27.5% conversion rate. +114.6% increase in sales.
World of Books: multi purchaser model: +83.5% rise in second purchases, month on month
Travis Perkins: VIP model +66% growth in VIP segment, spend per customer up by £194. “Churn can only be analysed o er a long period. You can inadvertently create a short term impact on churn. VIPs is a long-term sell. You want to identify individual behaviour at a stage, identify customers to then use different types of treatment.
Maturing model: reducing rate of churn.
Customer value increased by 38%.
Looked at how it would work to boost metrics and then customer lifetime value
In summary
• Optimise response times. Reduce complexity. How AI works to do that.
• Give campaigns, strategy real purpose. Turn those dials one by one
• Use customer lifetime value to track and measure success at all levels of the business. • It is a KPI that appeals to CEOs as much as to campaign execs.
The webinar ended in a Q&A session. Click here to see this webinar, as well as other previous webinars on the IRTV page, including slides, graphs and the closing Q&A.
