Provides an overview of two experiment methodologies typically used to test the effectiveness of advertising, specifically in the retail media or place-based media environment.
2. Bio 20-year career in media and marketing research Advertising, media, retail and technology expertise Former VP, Research & Shopper Insights at InStore Broadcasting Designed media planning systems at Nielsen IMS Media planner at Hal Riney & Partners John Morgan Founder and President
3. Overview Marketers are increasingly demanding accountability Proof of performance Consumer insights Ad optimization Return on investment
4. Digital Retail Media:A Perfect Laboratory Located at point-of-purchase Can be interactive Addressable Low-cost production Efficient Category exclusivity
5. Experimental Designs for ROI Analysis Matched panel design True experimental design that measures the effect of an advertising variable compared to a control during two observation periods Latin square design Used to measure the effect of multiple advertising variables without the use of a control
6. Matched Panel Experiments A method for testing the effect of a campaign variable (advertising creative) on brand metrics (brand equity, purchase intent, sales) in a real-world environment Small scale testing minimizes the risk and costs associated with a full roll-out
7. Matched Panel Methodology Locations are matched on traffic, category sales, demographic variables Locations are grouped Control cell (no advertising) Test cell(s) (advertising variables) Sales are compared on a pre vs. post / test vs. control basis Control Store Match! Test Store
8. Matched Panel Example:Convenience Store TV Product: Energy drink brand Venue: Regional c-store video network Objective: Positive ROI Measure: Net incremental revenues (increased profit minus media and production costs)
10. Matched Panel Example:Convenience Store TV Advertising ROI Analysis Net Revenue Difference $6,000 - Media & Production Cost $5,000 = Return on Investment $1,000
11. Latin Square Experiments A method for testing the effect of multiple campaign variable (advertising creative) on consumer response without using a control Minimizes the number of venues needed for testing Ad A B C
12. Latin Square Methodology Locations do not have to be matched, but should be representative Each advertising variable is introduced in every test cell during a different period of time Customer response is measured for each combination of advertising variable and time
13. Latin Square Example Product: Checking account promotion for a national bank Venue: Doctor’s office waiting area video network Objectives: QR code scans direct to account application, new checking accounts opened Measures: Number of QR code scans
14. Latin Square Example:Comparing Outcomes Total QR Scans Creative A: 60 Creative B: 75 Creative C: 45 Creative B is most effective!
15. The Proof is in The Numbers Our Services Audience Measurement and SegmentationAdvertising Sales and Marketing StrategyConsumer Insights for Content GuidanceAdvertising Campaign ROI Analysis Research Methods Syndicated Research Sources Matched Panel AnalysesConsumer Intercept SurveysObservational Studies Clients we Serve Advertisers and AgenciesMedia Network OperatorsRetailers and Venue OwnersInvestors John Morgan Founder and President Retail Media Insights john@retailmediainsights.com www.retailmediainsights.com