The Digital Crystal Ball: 3 Ways to Use Predictive Analytics to Grow Business
The crystal ball––oft spoken of in folkloric terms, but never available when you need it––has entered the realm of possibility. In our switched-on world, where digital interaction is present virtually every moment of an individual’s life, we as marketers now have the tools to peer into the future using data––not crystalline––to gaze forward at where our business is going.
Predictive analytics––the process of using new and historical data to foresee the result, activity, behavior, and trends of our consumer base––is the key that is making successful businesses, well, successful. Enterprises primed for growth in today’s hyper-competitive marketplace are using predictive analytics to gain a deep understanding of the customer base to maximize revenue, efficacy of marketing budgets, and, of course, profits.
So how can you unlock the benefits of predictive analytics for your business? Let’s look at some of the key predictive tools and how they can be deployed to help your business.
1. Predictive modeling of customer behavior:
Using data points gleaned from previous campaigns (particularly, those data that help us understand what worked and what didn’t), plus all demographic information known about your customer base, you can build predictive models to draw correlations to link past behavior and demographics. This model endeavors to score each customer according to their likelihood to buy certain products, and projects when and how to best approach this individual. In the wild, you may have seen tactics such as suggested products being offered to you during your online purchase checkout. This is an example of how this model works in execution.
2. Qualification and prioritization of leads:
Chasing a lead that is not likely to convert can be expensive. Applying predictive analytics to lead modeling can get you more “bang” for your lead investment buck. It uses an algorithm to score leads based on known interest, authority to buy, need, urgency, and available funds. The algorithm — using public and proprietary information – analyzes, compares and contrasts customers who converted with those who did not, and then find “alikes” among the incoming leads. The higher the score, the more qualified the lead. The highest-scoring prospects should be directed to sales or offered immediate incentives to convert; medium scores deserve a drip campaign; low scores…forget it.
3. Customer targeting and segmentation:
Among the most common use of predictive analytics, customer targeting and segmentation takes three basic forms:
- Affinity analysis refers to the process of clustering/segmenting the customer base according to attributes they have in common, facilitating “fine tune” targeting;
- Response modeling looks at past stimulus presented to customers, as well as the response generated (converted or not) to predict the likelihood of a certain approach to get positive response;
- Attrition rate (or churn analysis) provides a look at the percentage of customers lost during a certain period of time, as well as the opportunity cost/potential revenue lost with their departure.
With the deliberate use of these predictive analytics tools (and others), a business can then predict the Customer Lifetime Value (CLV). This measurement looks at several aspects of historical behavior to identify:
- the most profitable customers over time,
- acquisition spending trends around which activities generate the best ROI, and
- types of customers that are loyal (retention traits).
This model then adds an estimate of expected retention to the equation as a means of estimating future value. Once you understand the CLV, you can right-size the cost of acquisition and your marketing budget to reach the desired ROI.
One last note: When applying predictive analytics, it’s absolutely critical to A/B test your approaches to inform your output. Known as casual inference, A/B testing of the same target audience allows us to infer the WHY behind the WHAT customers are doing.
With these steps and measurements in place, you have earned your role as fortune teller – overseeing a true Predictive Analytics Organization. This is a tight ship, where marketing, sales, operations and finance work hand in hand, constantly providing feedback into the “data-outcome-analysis” loop.
Finally, the future of predictive analytics rests on ethics. Yes ethics. Instead of “sneaking into” peoples’ technology to follow their behaviors and disrupt their buying pattern to increase market share, the future of predictive analytics is to engage the consumers so they SHARE their preferences. That is what led Nike to acquire Boston based AI Platform Company Celect. By embedding predictive algorithms in their own website and aps, Nike will be able to better predict which models are getting traction, where consumers want to buy them, and when they are likely to buy.
Remember, it all starts with the clear articulation of the business strategy. With all parties in alignment, the chips should fall in place:
- predictive modeling of customer behavior helps educate campaigns to drive loyalty or generate leads;
- lead qualification modeling helps the sales team focus on the most probable customers to buy/close the deals;
- both of these together help finance understand the CLV and educate the whole organization on the acceptable customer acquisition cost to drive the targeted ROI.
If you’re not predicting, you’re losing ground.
Adriana Lynch is CMO with Chief Outsiders, the nation’s leading fractional CMO firm focused on mid-size company growth. She works with companies to differentiate, drive customer loyalty, and unlock profitable growth.
July 9, 2021