Forecasting promotions: You don't have to use a feather to crack a coconut

Forecasting promotions: You don't have to use a feather to crack a coconut

Yevgeny Abramovich examines why forecasting the impact or “lift” of promotions is becoming increasingly complicated - and how to pick the right tools to address the challenge.

How many times have you heard about a tempting two-for-one promotion for your favourite organic pasta only to be confronted with empty shelves when you arrived at the supermarket? I certainly have and the more times it happens, the less I trust the supermarket and consider alternatives. Now that social media provides a channel for consumers to vent disappointment and commiserate with fellow sufferers, trust in brands can erode faster than ever before. I’m sure most people reading this will have complained about bad service or stock-outs on Facebook or Twitter at least once.
There’s some interesting behavioural science behind this known as “the prospect theory”. Also referred to as “the loss-aversion theory,” this says that people tend to overweigh their perception of loss. Applying this to the supermarket example, not being able to buy the promoted pasta disappoints a customer much more than the comparative satisfaction of being able to purchase it (The Prospect Theory, Kahneman & Tversky). It’s also why you see so many more complaints being propagated through social media sites, than positive testimonials.
This phenomenon is all too familiar to consumer goods industry executives working hard to prevent customer disappointment during trade promotions. Forecasting the impact or “lift” of promotions is becoming increasingly complicated. Digital media, social networking, mobile devices and the decline of print media are all conspiring to change the consumer landscape. Even large market leaders are finding themselves at sea and losing money on promotions, missing their targets and having to cope with PR crises caused by angry customers. With the cost of promotions coming in at as much as 12-15% of gross revenues, this becomes a very expensive problem. Yet despite this enormous cost, most companies have little understanding of the impact promotions have on their most critical business driver, customer demand.
Why? Forecasting promotions is slow and onerous work because so many different variables with complex interactions between them are buried in huge data stores amidst lots of irrelevant information, or ‘noise’. Even highly expert planners are foiled by the process. Individual campaign complexity is multiplied by thousands of product-channel combinations every year.
This extraordinary challenge requires new tools and new approaches. According to Gartner, 59% of companies rely on basic spreadsheets for trade promotion forecasting and analysis. This is akin to using a feather to crack a coconut!
Fortunately, there are many alternatives to spreadsheets available to trade promotion planners. New technology is emerging on the scene that is finally making accurate promotion forecasting possible. For example, it is now possible to detect the shared characteristics of promotional events and identify the impact they have on normal product sales. Specialised trade promotion technology extracts knowledge about which variables most impact demand and produces a set of simple intelligible rules, easily understood by the company. Here are a few factors to consider when evaluating these:
  1. How well is the system oriented to the business? – although the system needs to be highly sophisticated and powerful, it should still be easy enough to use so that planners can focus on the business rather than toiling with complex technology.
  2. Does the system provide accurate planning information from the demand signal? – the systemshould translate the demand signal into a reliable inventory plan that guarantees the service levels you need to keep your customers happy.
  3. Can the system forecast promotion ‘lift’? – the system should be able to separate promotion and media lift from true baseline while effectively shaping demand at the customer / account level.
  4. Does the system support Big Data? – the system should allow vast amounts of raw data in different formats (for example corporate relational data and online data from social media sites) to be correlated and analysed. This should be scalable enough to meet current and future needs as the data universe expands. The analysis should be automated and fast so that planners can respond quickly to demand signals.
Regarding new approaches, the most successful companies are adjusting their operations to the challenges the market presents. To provide an efficient consumer response, I believe it is inevitable that companies will need to create an integrated SI&OP (Sales, Inventory and Operations Planning) process. Creating this ‘bridge’ and eliminating the functional silos provides companies with much better visibility of the multiple aspects of promotions and brings them closer to the holy grail of flawless execution supported by better inventory management.
By combining new tools and approaches, companies can run as many trade promotions campaigns as they require, achieve much higher returns and sustain long-term brand value through higher customer satisfaction and loyalty. Remember that some customers do make the effort to vent their approval through social media channels so that should provide some incentive to get the organic pasta back on the shelves!
Yevgeny Abramovich is business development manager at ToolsGroup.
Back to top Back to top