Client's new product forecasts produced highly inconsistent results, deviating as much as 50% from actual sales, despite the use of a well-known simulated test market technique.
Produce forecasts that are more closely aligned with actual in-market performance. Increase forecast consistency (reduce variance) so that estimates can be interpreted and compared; provide a basis for evaluating the relative sales potential across new products in the development pipeline.
A multi-stage solution to the new product sales estimation problem was designed. The answer did not lie in remedying technical inadequacies of any specific analytical tool. Rather, by making assumptions explicit within a stepwise process, forecasts were better understood, and where deviations from expectation occurred, diagnosis was possible.
Thinking Framework -- integrating client's pre-existing new product research in a decision structure
Thinkcast Simulation -- incorporating management assumptions about unmeasured factors
Integrated New Product Tracking Scorecard -- monitoring post-launch performance vs. forecast
Copyright 2013, Applied Thinking LLC
Few real-world problems are clearly defined. As a result, our solutions often entail a customized configuration of analytical tools.
Case 1: Improving the new product forecasting process
Client believed that the inability to generate empirically-based insights about the business easily hindered planning and placed their company at a competitive disadvantage when change occurred. Traditionally, managers at the company relied almost exclusively on their "deep smarts" about the workings of the marketplace when quick decisions were needed.
Develop a decision-support environment that promotes agile planning. Increase "speed to insight" by providing easy access to strategic information.
- Assemble the "right" data needed to support today's business
- Discontinue legacy practices, where appropriate
- Determine realistic constraints given the marketing information budget
The approach to building the new planning platform began with defining management's underlying mental models about the business (i.e. what factors were critical to success), as well as the extent of internal alignment across functional areas. An inventory of existing information revealed several gaps that required data acquisition from outside sources as well as internal collection. A knowledge base was designed, and problem solving frameworks were put in place that provided guidance on "how to use" the new data-based system effectively. Subsequent to the initial database creation, scorecards and standard metrics were introduced to support demand planning.
Analytic Audit -- determining management assumptions, needs, and current resource availabilty
Knowledge Base Design -- providing the architecture that supports the data-based environment
Thinking Framework -- setting guidelines for information use
Thinkware -- delivering key information in easy-to-use dashboards
Case 2: Developing a knowledge-based planning platform
Client wanted to determine whether an across-the-board price increase would be detrimental to its major product line. Since trends suggested that prior price reductions had little or no positive impact on sales, the client wanted to regain lost profit by increasing margin.
Estimate the sales and profit effects of increasing case price by 3 - 7%.
Since the client did not have an empirical measure of price elasticity, an econometric model was developed to assess the degree to which sales would respond to a price change. Given historical sales patterns, however, it appeared that other factors played a role in driving the client's business. The resulting model indicated that economic conditions had varying effects across the line. Further, since negative retailer reaction to a price increase was suspected, simulations that played out "what if" scenarios representing a range of possible in-store environments were created.
Given the complexity of pricing effects, a systems model that included management assumptions as well as data-based elasticity estimates was developed to provide realistic planning guidelines. Since many of the factors that could interact with price effects were uncontrolled by the client, simulations reflected both probabilistic as well as deterministic factors.
The net result of this approach was a range of potential outcomes that encompassed more than consumer response to a change in shelf price. Possible actions by the retailer, a key intermediary in pricing execution, were explicitly acknowledged and taken into account as part of the decision.
Market Response Analysis -- measuring the sales impact of retail price and other environmental factors
Thinkcast Simulation -- incorporating management assumptions about potential retailer actions
Case 3: Setting a pricing strategy
Learn more about ...
Learn more about ...
Learn more about ...