Cassie Lancellotti-Young is executive vice president of customer success at personalisation and marketing firm Sailthru.
cassie lancellotti-young at Sailthru
Personalisation has been a priority for marketers for many years now, but has now reached buzzword status. Gartner Research predicts that marketers fully invested in personalisation will outsell their competitors by 20% by 2018.
So what does personalisation actually mean? The truth about personalisation is that there is not one single meaning, rather it’s a business strategy that exists – and continues to evolve – on a maturity curve. There are varying definitions used by marketers to communicate how they can become increasingly more sophisticated.
sailthru personalisation curve
As marketers move up and to the right on this curve, the return on investment (ROI) becomes that much more palpable: when we’re speaking of basic implementations such as field insertion in subject lines, maybe we see a modest lift in opens, whereas content and product recommendations based on demographic segments might yield more clicks and a modest improvement to conversions. Recommendations leveraging behavioural data would be the holy grail of sorts in driving more sales.
Marketers who have experienced the benefits of personalisation have done so through planned, tested programming. Increasing their maturity and use of personalisation has become the focus on their CRM and retention programs. And rightfully so – Bain and Company reports that increasing retention by 5% can increase profits anywhere from 5% - 95%.
Up until very recently, best-in-class marketers were those leveraging advanced segmentation, using frameworks like recency-frequency-monetary (RFM) to inject behavioural elements into their marketing programs. But times (and technologies) are changing. Top consumer marketers recognise the gains to be had well beyond classical segmentation, namely in the form of customer predictions.
What’s perhaps most unique about customer predictions is that they benefit both the consumer and the marketer. Predictions take relevance to the next level for the customer as marketing programs become more than just about product promotion, but actually going so far as to target customers with items in their anticipated buying window and price range (if I typically spend £100, surely I am more likely to transact when I am given recommendations in that price range than substantially above it), or sending content-focused treatment over product treatments when an individual is not likely to buy, as well as influencing messaging cadence and channel.
On the marketing side, retailers can stretch expected spend (e.g. only offer an incentive if the customer meets a floor of 110% of what her buying prediction says she will) and optimise margin (e.g. suppress customers who are 95%+ likely to buy today from an incentive or discount offer). At the intersection of both, a retailer might automatically suppress any customer with a 90%+ likelihood to opt out from that day’s sends, which helps the consumer (not bombarded with messaging) while also helping the brand (decreases opt-outs, thereby protecting customer lifetime value). The caveat to these use cases is that marketers must have access to predictions made at the individual customer level and available in that format for messaging.
And while the primary focus for predictions has been on retention, there is an application for the top of funnel in a way that works for consumers and brands alike: consumers see more relevant ads, and brands reduce their acquisition costs. Country Outfitter is a US retailer who recently leveraged likelihood to buy predictions to build lookalike audiences on Facebook to identify new customers who looked just like those with the highest near-term transaction value. The result was a 94% reduction in cost per checkout over the control method of simply using backward-looking customer value figures.
Again, it’s important to note that any step forward on the personalisation curve will drive returns, but as we move up the maturity model, the magnitude of those returns becomes that much more material. The reason marketers have relied on models like RFM for decades is because they work, but predictions that bake in significantly more customer data points, all captured in real-time take the returns to the next level. On average, when we look at the expected revenue of marketers using our own predictive model vs. RFM, marketers can tap into a potential 40% increase in revenue based on their ability to optimise the experience from the inside-out and the outside-in.
As we move into 2016 and beyond, it will be interesting to see where else customer predictions will take us. We might see service centres sorting call queues based on likelihood to buy; in-store sales associates leveraging beacons and predictions to offer products that fit a customer’s interests and predicted basket size; or cross-channel coordination of offers tied to individual behaviours in email, web and mobile channels. As with any solution, there will be evolution, but what we do know is predictions will be a transformative force in marketing programs moving forward.