Promotional Modeling Optimization

With over $90 billion per year being spent against trade funds, consumer goods executives realize that better management of this process could be a huge gateway to success. This month, Consumer Goods Technology caught up with John Bermudez, Vice President of Marketing, Demantra, to discuss the challenges executives are facing around trade promotions and how advanced analytics is a necessary tool to maximize the investments being made.

As consumer goods companies continue to struggle to get a handle on trade funds, what specific challenges is the industry facing?
The biggest trade spending challenge facing the industry is directing funds toward the most profitable events and being able to validate this with retailers. Part of the profitability challenge is being able to execute the promotion cost effectively, which requires the supply chain and manufacturing groups to deliver the increased demand without incurring unnecessary overtime and transportation costs.

what types of solutions are companies deploying? What key deliverables are being overlooked?
Most companies are still executing the most critical steps of the trade promotion management process on spreadsheets -- this includes companies that have implemented CRM systems to manage the trade fund process. These critical steps include predicting the expected lift from the promotion and developing a volume forecast that considers store level difference. Companies relying on spreadsheets typically use gross lift factors (often from last years events calculated at a national level) and require that sales reps manually adjust these forecasts. Lift factors need to be based on current demand patterns and adjusted to the cluster of stores served by a specific distribution center. Consider a lemonade drink mix promotion in May -- it might produce a six-times lift in the southeast where temperatures are already in the 90's and a two-times lift in New England where temperatures are likely still in the 50's. Projecting a four-times lift nationally is going to lead to overstocks at the New England distribution center and out-of-stocks at the Southeastern distribution center.

In the past, the finance department
was the primary user of the trade funds system. sales and marketing are now intimate users as well. Is it a challenge to create a system that is equally shared and effective for all three groups?

It is very challenging to adapt a system designed for finance to reconcile short pay invoices to one that sales will use for promotion planning. This is why there are still so many spreadsheets in use. To make the system useful for sales it must be able to be able to produce accurate event forecasts at the account (customer) level and allow them to simulate the event so they can predict the outcome before they offer to the retailer. Sales account teams are not interested in accurate aggregate forecasts -- they need accurate forecasts for their retailer at a regional or DMA level to be maintain credibility with the buyer. While achieving all of this, the system must also be easy to use as sales reps get paid to be in of front customers not computer screens.

To make the system useful to marketing is even more challenging as it must provide in-depth analytics (see chart) to help them to determine the real impact of marketing programs. TPM systems typically have no real analytics. The analytics capabilities need to analyze all the effects of a promotion including pre- and post-effects, cannibalization, forward buying, category growth and brand switching. It must be able to do this analysis from a variety of data sources including syndicated data, retailer POS data and shipment data. Most importantly, it should be able to use this analysis to optimize the promotion calendar.
What challenges do companies face in effectively forecasting for a promotional event?

The biggest challenge is often overcoming the skepticism that it is possible to deploy a forecasting tool that is capable of providing precise promotion forecasts at an account level without lots of manual intervention. This skepticism often causes a company to do nothing or tolerate poor forecasts from inadequate systems. Following right behind this challenge, is cleaning up promotion history data used for forecasting. To improve the accuracy of history data it is critical that sales maintain an accurate record of the incentives and timing of an event. Integrating the forecasting tool into the sales and account planning system to project account profitability as well as promotion forecasts provides an incentive for sales to maintain promotion history.

It seems the missing piece of the puzzle is advanced analytics, as this would provide field sales with the artificial intelligence needed to accurately produce the forecast. What value can a consumer goods company gain from a TPM solution that combines this analytics piece?

Combing advanced analytics with TPM is the means to produce more accurate forecasts. More importantly, it provides the analytics to determine which promotions produce the greatest return on investment. An accurate forecast on an unprofitable promotion only addresses part of the problem. Having the analytics integrated with the volume forecasting tool ensures that the attributes of an optimized event are converted precisely into event plans at the account level. This is how we avoid the lemonade drink mix problem mentioned above.

Is the concept of Promotion Modeling Optimization the key to consumer goods companies' ability to realize ROI on promotional spend?
Promotion modeling optimization is critical to improving overall ROI on promotion spending. CPG companies run hundreds/thousands of promotion events across hundreds of different customers executed by hundreds of sales reps and brokers. This adds up to tens of thousands of the promotion event variables to evaluate and execute, which is too many to plod through using trial and error what-if analysis. Promotion modeling optimization uses the processing capabilities of today's very powerful servers to sort through the enormous number of possible events to identify those that meet the company's business goals or event objectives.