In a nutshell, that is what predictive modelling does for you. We, as marketers, have evolved our digital communications to a point where they are outstanding.
We have personalised, optimised, and tested our automated communications to enhance the customer experience and improve conversion. The final piece of the puzzle is finding the right time to deploy them to the individual.
Most marketing automation platforms restrict marketers by only allowing for a one size fits all approach when it comes to the timing of communications and behavioural lifecycle campaigns.
However, with predictive modelling the timing of each communication is determined at an individual level, so it is relevant to each person and their individual behaviours.
Predictive modelling uses machine learning to predict the likelihood that a contact will perform a specified behaviour
The Aberdeen Group found that predictive analytics users are twice as likely to identify high value customers by influencing individual’s behaviours to nurture and retain their existing customers, as well as converting prospective customers.
At Redeye we’ve built our predictive suite based around the customer lifecycle, predicting whether a contact is likely to make their first purchase, become a VIP or begin lapsing based on their individual purchase and browser behaviour.
The beauty of adding machine learning onto your customer database is that it will continue to learn and become more accurate as your customers continue to interact and engage with your brand. The more data it must analyse, the more accurate the model becomes.
If customers are in the growth stage of their journey with your brand, the RedEye predictive models can help to identify those who are most likely to make a purchase, allowing you to facilitate and encourage conversion. Likewise, if they are looking likely to lapse, you can proactively plan and activate a lapsing campaign to stop them churning or unsubscribing.
Prospects typically make up the majority of customer databases for most retailers
If you could convert just 1% of your prospects, what additional revenue would that generate for your business? These are contacts in your database that you have not yet managed to convert, perhaps even despite your existing welcome programme and day to day communications.
That’s where predictive modelling comes in. The models score and identify groups of contacts that are highly likely through to not very likely to purchase, allowing brands to focus their marketing efforts effectively and efficiently.
This could involve including “recommended for you” content and existing customer reviews within email campaigns to those more likely to purchase, or utilising website personalisation to help nudge them towards a purchase when the contact next visits your website.
To maximise efficiency and protect your margin, consider removing discount codes or offers for customers who are already likely to purchase. Instead, consider a spend stretch strategy by displaying up-sell and cross-sell products to increase their basket size.
Implementing predictive modelling helps you move towards a more individual personalised approach rather than one size fits all solution
Armed with the knowledge of when a contact is likely to purchase you can start looking into areas, such as your check-out journey, for any common drop off points to build an automated, re-targeting campaign to encourage purchase.
Contacts who are less likely to purchase may require an additional incentive, and in this example from buyagift they have included a unique voucher code to ensure a one-time use code is displayed and cannot be shared or used by anyone else.
To make this email truly personalised, within the recommended for you section it also displays relevant products, based on the individual’s browse behaviour.
Another great tactic to aid conversion is to use time sensitive offers, such as the free delivery offer shown in this example from My Protein.
By limiting the window of conversion time, this adds a sense of urgency and can help convert customers who are predicted as less likely to make a purchase.
To enhance this email further you could also include interactive content such as a live countdown clock which would tick down the time left to redeem the delivery offer.
Outside of offering discounts, it is also worth considering what else could be a blocker to making a purchase for these contacts. It could be giving greater reassurance around delivery, a variety of payment methods, or perhaps displaying social proof and customer reviews to give them the confidence they need to make a purchase.
How can predictive modelling assist with my peak strategy?
Predictive modelling really gives you the power to take your segmentation strategy and automated lifecycle programmes to the next level. Implementing this approach to your day to day activity now could be a perfect opportunity to create an uplift in engagement prior to the peak period.
With Black Friday and an influx of marketing activity just around the corner, if you have already done the hard work of converting more of your prospects into purchasers, the likelihood of a customer purchasing a second time increases to 27%, giving your Black Friday campaigns a greater chance of conversion success.