Moving beyond the era of traditional mass marketing, today AI offers businesses an opportunity to understand and analyse customer behaviour at a very granular level. This enables them to have a degree of predictability around customer behaviour, and to hyper-personalise and contextualise the entire multi-channel customer experience.
How can businesses best leverage this powerful tool to gain sustainable competitive advantage in the long term
At a recent masterclass at TechSparks 2018, Smartech’s Pradyut Hande, Senior Growth Marketer and Product Evangelist and Kedar Parikh, Product Architect, took the audience through the various ways in which AI is transforming the data-driven marketing environment on a day-to-day basis. Here are some highlights from the interactive session.
For businesses today, both digital and/or offline, the challenge is to deliver extremely hyper-personalised and contextually engaging campaigns based on customers’ behaviour. This becomes important to increase customer lifetime value. For instance, the traffic to your website could be spiking at a particular day and time in the week. You run analytics to see what precedes this spike. Were there targeted social media campaigns that brought the traffic? Here is where a combination of predictive analytics, prescriptive analytics and trend analytics come in.
Pradyut added, “While AI can give you a great degree of insight, sometimes you need to amalgamate common sense and gut instinct with AI to arrive at the best and fastest decision.”
Campaigns are effective only when the right content reaches the right audience at the right time through the right channel. There are a number of factors which contribute to content personalisation. Smart segmentation helps you target the right audience; recommendation engines and subject line personalisation help you send out the right content. The user’s preferred channel of communication is the right channel. Everyone has a sweet spot when they are most likely to check their email. Send time optimisation can help marketers reach the prospect’s inbox at a time their email or notification is most likely to be opened.
They went on to cite the example of a business that was able to see a consistent 20 to 25 percent increase in the email open rates just by being able to capitalise on customers’ preferences of time.
While it is impossible for any business to retain 100 percent of their acquired customers, churn prediction can help fuel engagement and retention strategies. Over time, it can track user actions, inactions, and drop in engagement rates across channels. For instance, it can tell you the number of users uninstalling apps, unsubscribing from newsletters or publications. This data is critical to marketers because it highlights a red flag, and they need to craft engagement campaigns that are extremely contextual, and personalised to re-engage with these customers in an effective manner to ensure that churn is at least arrested, if not eliminated.
A leading payment wallet app that was seeing an extremely high churn rate of about 45 to 50 percent over a 30-day period started running campaigns based on data derived from churn prediction. The result? Their churn rate reduced by almost 15 percent.
Today, nearly 30 percent of sales for Amazon come from recommendations, which is a significant number. While a customer browses, searches, or adds a particular product to his or her cart, a business might not want the engagement to end there. The idea is to maximise revenue from a particular user. For online businesses, recommendation engines work by leveraging a customer’s past purchase behaviour and capturing preferences in the back end, and displaying products that the customer is likely to engage with, based on that data.
How do you actually get somebody to open your email? Subject lines play a key role in gaining customers’ attention. If the subject line doesn’t appeal to them, it is unlikely that they will open the mail. Based on the kind of subject lines of mails that a user has been engaging with, businesses can come up with a simple model using NLP and sentiment analysis to tailor the subject line at the user level, to use particular keywords. The more contextualised and personalised the subject lines are, for emails, or push notifications, the higher the potential conversion rate.
How do you reach more customers on the right channel is also important. Engage with users on their preferred channel. Look at all the historical interactions across channels, whether they prefer app push notifications, browser push notifications, email, SMS, or in person interaction.
A simple machine learning model can slice through huge chunks of data on users and identify natural segments based on users having similar attributes, which makes it easier to customise the outreach for each particular segment, for instance the customers with high margin transactions.
Today chatbots are a convenient option for businesses for customer engagement. It may not be a sustainable option for a lot of players to have human agents on the job 24 x 7. Chatbots solve a variety of use cases, from customer support to gathering feedback, resolving frequently asked questions, across industries.