In marketing analytics, a marketing and sales funnel is the set of steps a visitor goes through before making a purchase. More and more, this customer’s journey includes a diverse set of touchpoints, involving both push and pull:
The most accurate and reliable way to find the optimal marketing touchpoints for your own funnel is to use machine learning— clever algorithms that quickly search through your data. These algorithms learn by example, finding common patterns within customers’ demographics and touchpoints in order to predict sales.
While each customer journey looks different, it can be conceptually represented by the diagram below (Figure 1).
Figure 1 : A Customer Journey Can Include Many Touchpoints
The challenge for marketers is to optimize this journey, selecting the touchpoints that bring the customer to a decision to purchase as quickly and smoothly as possible. Some customers may need only 1 or 2 touchpoints, while others may need 30 or more touchpoints. Some customers prefer push-based touchpoints, while others prefer self-service touchpoints. Some customers prefer passive media, while others want a more active experience where they interact with websites and their peers on social media.
Historically, organizations have sent the same messages to all customers, using the same media channels. But, in the modern world, this may be perceived as SPAM that annoys customers and pushes them away. The diagram below (Figure 2) shows the traditional approach in action – every customer sees the same sequence of touchpoints.
Figure 2 : When All Customers Receive the Same Outbound Marketing Content
More sophisticated marketers have applied customer segmentation to their customer database. They group customers into segments according to their demographics, creating different customer journeys within each customer segment. Then, they send the same messages, using the same media, to all customers within each segment. This is a marked improvement on the previous approach, but not every customer within a segment is the same. The customer segments are typically grouped by demographics and not according to the content preferences of the customers.
Figure 3 : When All Customers Within a Segment Receive the Same Marketing Content
The diagram above (Figure 3) shows different marketing touchpoints for different customer segments. While the match between the marketing touchpoints and the needs of customers has been improved by segmentation, there nevertheless remains a mismatch at the individual customer level.
With a large pool of customers and range of communication channels and messages, how do we know who to target with which channel and which message?
The solution is next best action, widely acknowledged as best practice in modern marketing. Next best action considers historical data about each individual customer, such as:
Using this type of customer and behavioral data, it is possible to find patterns between characteristics of customers, the sequence of their customer journeys, and which customers ultimately purchased. For example, it may be that customers under the age of 30, who received an email followed by an SMS reminder the next day, were much more likely to accept an invitation to attend a seminar. On the other hand, female customers aged 45 to 47 who live in the north responded better when sent newsletters.
The best way to find these patterns is to use machine learning — where algorithms learn by example, searching for common patterns in the customers’ demographics and touchpoints in order to predict sales. Once the algorithm has learned these patterns, you can use it to select which marketing content to use next. Tell the algorithm which touchpoint you are considering, and it will predict the probability that this customer will ultimately purchase. When you consider a different touchpoint, the probability changes. The “next best action” is the touchpoint with the highest probability.
Figure 4 : Evaluating the Value of Each Possible Touchpoint
The diagram above (Figure 4) shows a worked example of a next best action algorithm. In this example, you would send the second touchpoint for your next communication to this customer, because adding that touchpoint brings the highest probability of a sale (55%) and therefore is the one that would most likely move this particular customer closest to a decision to purchase.
Apply this to every customer and you will get the next best action for each customer. As each customer’s behaviour changes, the data also changes. For example, they visit your website and browse the pricing pages, so the next best action should be recalculated to optimize this customer’s experience. By responding to the individual behavior of each customer (Figure 5), next best action ensures that each customer experiences their own personal journey.
Figure 5 : With Next Best Action, Every Customer Has A Personalized Journey
Next best action models can be complex and time-consuming to build, but with the DataRobot automated machine learning platform, next best action models can be built with just one click. DataRobot helps you quickly generate accurate predictions to find the most appropriate touchpoint for each customer, focusing your marketing efforts, and optimize your revenue and profit.
About the Author:
Colin Priest is the Director of Product Marketing for DataRobot, where he advises businesses on how to build business cases and successfully manage data science projects. Colin has held a number of CEO and general management roles, where he has championed data science initiatives in financial services, healthcare, security, oil and gas, government and marketing. Colin is a firm believer in data-based decision making and applying automation to improve customer experience. He is passionate about the science of healthcare and does pro-bono work to support cancer research.