Enterprise Marketing & Promotions Management
Predictive Analytics

4 Ways To Improve Retail Promotions Planning with Predictive Analytics

By / May 2016


How do you decide what products to put on the front page of your promotional mailer, or in your weekly customer email, at what price point? Are your promotions actually effective – or could you be hurting your own performance?

Maybe you have some basic analytics on past performance. Maybe you have some category performance targets you need to hit, or some competitive positioning goals. But with all of the data you have available on customers and their shopping behavior, it can be overwhelming. You need to find ways to sift and gain insights quickly under pressure.

And the risks are high: getting promotions right can improve traffic and basket size, not to mention promotions play a huge role in a customer’s perception of value. Choosing the wrong offers can lead to financial losses that are hidden in other metrics.

Enter a new realm of big data analysis called “predictive analytics” or predictive modeling. As machines learn to look at the historical trends and patterns in your data, they can essentially predict or model where you’ll have the most success. Here are just four ways predictive analytics can help you in planning smarter retail promotions.

1. Targeting offers to the right customers
While predicting product sales is critical, predicting sales by customer segment is essential. How do you know whether your promotions are great for cherry pickers but awful for your most valuable customers? Integrating behavioral and attitudinal information to create a consumer-centric view of your data establishes a clear path for leveraging customer science and predictive analytics.

For example, these advanced analytic approaches can be used to develop consumer segmentations, predict consumer churn, and calculate a retailer’s share of wallet. You can identify price elasticity and know what shoppers are willing to pay – and move away from simple competitor price matching. If you can’t predict your performance by customer group, you are throwing promotions against a wall and hoping some will stick.

2. Using decomposition to score and rank promotions
How do you define how a promotion is “successful?” It depends a lot on your objectives (You have defined those, right?) Some promotions work well for an overall category; others work well to bring in new shoppers and improve household penetration; others serve to increase market basket diversity. Science-based predictive analytics tools can help you score and rank promotions by their function and objective.

Predictive analytics can tell you how long to run a promotion, what tactics will work best, and what combinations will best help you achieve the results you want. What’s the benefit of moving a SKU from Page 2 of the circular to Page 1? If you don’t know, you’re disappointing your customers and losing potential trade funding from your suppliers.

3. Making faster business decisions with interactive visualization
It used to be that data analytics remained just the purview of a select set of analysts crunching Microsoft Access databases or using basic Business Intelligence tools and then presenting their findings to the business in monthly or quarterly meetings.

Too many retailers are still making these decisions with a combination of spreadsheets and hours of human effort. You can shave months of analysis of complex data down to minutes in some cases.

If that’s still your process, then you’re missing the opportunity for faster decision making. Predictive analytics in combination with new visualization capabilities means business users in merchandising, marketing and category management can make quicker choices – and make them in real time based on the most recent data available.

4. Improving relevance through personalization
Who doesn’t want their promotions to be more relevant? Once you understand the foundational analytics, why consumers make choices and draw insights through predictive analytics, you can begin to make decisions focused on consumer strategy and engagement and develop personalized promotions.

These personalized data driven engagement strategies create the ability to deliver relevant content at the optimal frequency and cadence through the channels that the customer prefers. This will increase engagement rate, as well as develop loyalty in an increasingly competitive environment.

In summary, if you’re not taking advantage of predictive analytics, quite simply you could be missing the chance to capitalize on revenue opportunities. And, you could be hurting long-term growth. Innovation in the area of analytics tools means you can see into the future like never before – so it may be time to explore how your business can benefit.

Contributed by Howard Langer, Managing Director, dunnhumby Price and Promotion
Howard leads the Global Price & Promotions Practice at dunnhumby leveraging experience in working with leading retailers in North and South America as well as mainland Europe and Asia. A former retailer, Howard has spent much of his career delivering Price and Promotions programs from senior merchandising and marketing roles in the UK while at the Kingfisher Group, one of the largest retail groups in Europe. He implemented the UK’s first ever price optimization and has successfully delivered many multi-million dollar price investment programs. Howard is the Founder and Chair of the Retail Pricing Forum in the UK a group dedicated to improving the Price & Promotions capability in retailers. He is a visiting lecturer at both the University of Southampton and Warwick Business School in the UK.


View More Results