Predictive Analytics For Profit

Predictive Analytics For Profit

You’ve certainly heard the buzzword “predictive analytics,” but how many of you have actually harnessed the power of predictive analytics to improve your bottom line?

Prediction Drives Performance
Research from the management consulting firm Bain shows companies that use advanced analytics are twice as likely to be in the top 25% of their market than those who do not. In my experience, it’s not uncommon to see at least a 10-20% increase in profits when applying advanced analytics to a problem. Many times the benefit is much more.

Before we get into things, let me give you a basic definition. At the root, predictive analytics use historical data to predict future events. The idea is, if you can predict the future, you can intervene and change the future for the better.

There are dozens of ways for media companies, publishers, and other content creators to use productive analytics to grow revenues and profits. Here are just eight ways to do so.

1. Increase Success Rate of New Product Development and Content Creation Efforts
A study by Harvard Business School showed that in the US, 85% of new product launches failed due to poor market segmentation. In other words, the company did not build a product that met the needs and wants of the people they targeted. Predictive analytics can group similar customers based on their prior behavior. This allows you to be more precise in your product development, content creation, and marketing activities. Read The Sterling Woods Group’s article on the benefits of segmentation here. For illustrative purposes, let’s say we are launching a new travel magazine. We can use data to determine which customers in our marketing database are business travelers, versus family travelers, versus adventure travelers. We can use data such as which articles they have clicked on in the past to make such classifications. Once we have properly segmented our customers, we can present content, products, and offers that appeal more directly to each group.

2. Increase Direct Mail Acquisition ROI
Most marketers accept an extremely low direct mail response rate: 2-3%. Why not raise the bar with predictive analytics? Use data insights to send more appealing offers, or at least save time and money by not marketing to those unlikely to respond. Many publishers purchase scoring models from list vendors, but have you asked when the model was last updated? Models lose effectiveness over time. Why not try building a proprietary model with your own data and see how it competes with purchased models? Or try using both.

3. Expand Scope of Email Tests and Increase Conversion
Many marketers use the concept of A/B testing to improve conversion rates of emails. This means they send out two versions of an email, Version A and Version B, and see which one does better. However, you probably want to test more than just two options at a time. The problem is math gets complicated. For example, if you want to vary six different things (for example: greeting, graphic, first sentence, copy, shipping cost, promotional price), you would need to set up 64 different emails (2^6) to test. Using an advanced predictive analytics technique (fractional factorial design), you can create an experiment with only 16 different versions. You learn just as much as 64 tests with only 25% of the effort. I’ve led this project for a media company that experienced a 50% improvement in conversion rate as a result.

4. Increase Renewal Rates
You have probably seen comical decision trees circulating in social media (one of my favorites is “Which Fast Food Chain Should You Choose”). Marketers can build serious decision trees using advanced analytical tools to predict future outcomes based on actions the customer has taken. I’ve used decision trees to identify people that were unlikely to renew, then put in place communication strategies to save at-risk subscribers. One such campaign saved us tens of millions of dollars.

5. Drive Cross-Sales
Core Metrics found that companies that use recommendation engines as part of their e-commerce strategy could expect a 5-15% increase in total revenue. Predictive analytics are behind the Amazon recommendation engine. Why not use it on your platform?

6. Increase Average Pricing
Segmenting your customers by their willingness to pay helps you better target promotional campaigns. Offer the deepest discounts to the most price sensitive segments to get them in the door, and wow them with the quality of your product so they renew. I have experienced a double-digit lift in average pricing by using this methodology with several different clients.

7. Optimize Acquisition Budget Through Understanding Customer Lifetime Value
Analytics can predict a lifetime value based on things you discover about a customer in their early days – where they come from, what their first few activities are, etc. Understanding a new customer’s likely lifetime value allows companies to properly allocate limited marketing resources and ensure the proper ROI. Let’s say I knew someone acquired through direct mail would generate $30 in lifetime value, thanks to our model. So, I’d keep increasing the investment in direct mail until we hit the point that the cost per acquisition was low enough that the company could earn a satisfactory return on the marketing investment.

8. Reactivate Expired Customers
Want to get back some of your former customers within the constraints of your marketing budget? Use a predictive model to determine who is most likely to come back based on their behavior when they were customers. At one company, we were able to cut reactivation marketing spend by $300,000 while maintaining total revenue.

How The Sterling Woods Group Uses Predictive Analytics
We use predictive analytics to help our clients find new products and services to launch. Applying analysis early in the product development process increases the odds of success. We continue to apply advanced analytics to maximize profits from these new revenue streams post-launch.

Close Menu