1. Decision Model Development Stephan Sorger www.StephanSorger.com © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 1www.StephanSorger.com

2. Model Development Process: Overview Interpret Results Define Problem Select Model Diagram Process Build Model Gather Data Run Model Present Findings © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 2www.StephanSorger.com

3. Model Development Process: Define Problem Interpret Results Define Problem Select Model Diagram Process Build Model Gather Data Run Model Present Findings Define ProblemDescription Product featuresWhat features should be included in the product? PricingWhat should be price of product/ service? PositioningHow should product/ service be positioned? SegmentationWhat are the salient segments in the market? TargetingWhich segments should be targeted? BudgetHow much advertising budget should be allocated? © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 3www.StephanSorger.com

4. Model Development Process: Select Model Interpret Results Define Problem Select Model Diagram Process Build Model Gather Data Run Model Present Findings Select ModelDescription ConjointDetermining which features matter most RegressionForecast future sales using historical data BassForecast future sales without historical data Decision TreeMaximizing expected value of different choices © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 4www.StephanSorger.com

5. Model Development Process: Diagram Process Interpret Results Define Problem Select Model Diagram Process Build Model Gather Data Run Model Present Findings Inputs Model Outputs Historical Data Survey Results © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 5www.StephanSorger.com

6. Model Development Process: Build Model Interpret Results Define Problem Select Model Diagram Process Build Model Gather Data Run Model Present Findings InputsOutputs Procedure: How model works Sheets/ Tabs (examples) - Tab 1: Main page - Tab 2: Historical data - Tab 3: Graph/ Plot Model process area - Algorithms for model - Excel functions Usage: How to use/ interpret outputs Calibration: Making model “fit” data © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 6www.StephanSorger.com

7. Model Development Process: Build Model Interpret Results Define Problem Select Model Diagram Process Build Model Gather Data Run Model Present Findings Excel FunctionDescription Arithmetic (+, -, x, /)Sum up survey results, divide to get average, etc. Ratios(Marketing Spend) / (Annual Revenue), etc. Analysis ToolPakSet of analysis functions accessible via Tools menu Regression: Find coefficients to fit line into data Moving Averages: Filter noise out of data Statistics: Sampling, z-tests and t-tests AlgorithmsSearch “model (e.g. Bass) algorithm excel” http://andorraweb.com/bass/http://andorraweb.com/bass/: equations, usage, etc. Pivot TablesAnalyze data by changing (pivoting) data structure © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 7www.StephanSorger.com

8. Model Development Process: Gather Data Interpret Results Define Problem Select Model Diagram Process Build Model Gather Data Run Model Present Findings Gather DataDescription HistoricalPast sales data Data by customer, product, sales rep, date, etc. SurveyMarket research data Survey target market to get feedback on proposed idea Qualitative research: Open-ended discussions Quantitative research: Send out questionnaire CalibrationTweak model to “fit” data Use regression analysis to find model coefficients Use Excel Solver tool to optimize model for data © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 8www.StephanSorger.com

9. Model Development Process: Run Model Interpret Results Define Problem Select Model Diagram Process Build Model Gather Data Run Model Present Findings Run ModelDescription AssumptionsDefine assumptions; Examples: “Assumes economy will stay at current state” “Assumes no reaction from competitors” ScenariosDefine multiple scenarios; Examples: “Advertising at Current Levels (Baseline)” “Advertising at +10% Increase” What-IfRun different scenarios Determine results with different inputs DemoEnsure smooth running demonstration © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 9www.StephanSorger.com

10. Model Development Process: Interpret Results Interpret Results Define Problem Select Model Diagram Process Build Model Gather Data Run Model Present Findings Interpret ResultsDescription TablesCreate tables showing results of different scenarios GraphsPlot out data in tables Visual representations assist in forming insights ProblemRelate findings to original problem How does model help to solve problem? What additional insights have we learned? AdaptabilityApplicability of model to other marketing problems Where else can we use this model? © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 10www.StephanSorger.com

11. Model Development Process: Present Findings Interpret Results Define Problem Select Model Diagram Process Build Model Gather Data Run Model Present Findings Present FindingsDescription DocumentationAdd labels to Excel model to assist others in using User GuideCreate MS PowerPoint / Word guide to using model: -Problem statement and background -Model selection rationale -Research methodology and model calibration -Procedure: How to use model and interpret results -Use case: Example of how model is applied GraphicsPlots/ Charts/ Tables to show results Visual results make more powerful arguments InsightsSummarize insights learned and how problem solved © Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics. Model Development. Page 11www.StephanSorger.com