Marketing Metrics forDigital Retail MediaExperimental Design for Campaign Measurement
Bio20-year career in media and marketing researchAdvertising, media, retail and technology expertiseFormer VP, Research & Shopper Insights at InStore BroadcastingDesigned media planning systems at Nielsen IMSMedia planner at Hal Riney & PartnersJohn MorganFounder and President
OverviewMarketers are increasingly demanding accountabilityProof of performanceConsumer insightsAd optimizationReturn on investment
Digital Retail Media:A Perfect LaboratoryLocated at point-of-purchaseCan be interactiveAddressableLow-cost productionEfficientCategory exclusivity
Experimental Designs for ROI AnalysisMatched panel designTrue experimental design that measures the effect of an advertising variable compared to a control during two observation periodsLatin square designUsed to measure the effect of multiple advertising variables without the use of a control
Matched Panel ExperimentsA method for testing the effect of a campaign variable (advertising creative) on brand metrics (brand equity, purchase intent, sales) in a real-world environmentSmall scale testing minimizes the risk and costs associated with a full roll-out
Matched Panel MethodologyLocations are matched on traffic, category sales, demographic variablesLocations are grouped Control cell (no advertising)Test cell(s) (advertising variables) Sales are compared on a                               pre vs. post / test vs. control basis Control StoreMatch!Test Store
Matched Panel Example:Convenience Store TVProduct: Energy drink brandVenue: Regional c-store video networkObjective: Positive ROIMeasure: Net incremental revenues (increased profit minus media and production costs)
Matched Panel Example:Convenience Store TVPre-PeriodCampaign-Period+ $6,000Net RevenuesDifference in Sales
Matched Panel Example:Convenience Store TVAdvertising ROI AnalysisNet Revenue Difference        $6,000- Media & Production Cost   $5,000= Return on Investment        $1,000
Latin Square ExperimentsA method for testing the effect of multiple campaign variable (advertising creative) on consumer response without using a controlMinimizes the number of venues needed for testing AdABC
Latin Square MethodologyLocations do not have to be matched, but should be representativeEach advertising variable is introduced in every test cell during a different period of timeCustomer response is measured for each combination of advertising variable and time
Latin Square ExampleProduct: Checking account promotion for a national bankVenue: Doctor’s office waiting area video networkObjectives: QR code scans direct to account application, new checking accounts openedMeasures: Number of QR code scans
Latin Square Example:Comparing OutcomesTotal QR ScansCreative A:  60Creative B:  75Creative C:  45Creative B is most effective!
The Proof is in The NumbersOur Services Audience Measurement and SegmentationAdvertising Sales and Marketing StrategyConsumer Insights for Content GuidanceAdvertising Campaign ROI AnalysisResearch MethodsSyndicated Research SourcesMatched Panel AnalysesConsumer Intercept SurveysObservational StudiesClients we ServeAdvertisers and AgenciesMedia Network OperatorsRetailers and Venue OwnersInvestorsJohn MorganFounder and PresidentRetail Media Insightsjohn@retailmediainsights.comwww.retailmediainsights.com

Retail Media Insights - Marketing Metrics for Digital Retail Media

  • 1.
    Marketing Metrics forDigitalRetail MediaExperimental Design for Campaign Measurement
  • 2.
    Bio20-year career inmedia and marketing researchAdvertising, media, retail and technology expertiseFormer VP, Research & Shopper Insights at InStore BroadcastingDesigned media planning systems at Nielsen IMSMedia planner at Hal Riney & PartnersJohn MorganFounder and President
  • 3.
    OverviewMarketers are increasinglydemanding accountabilityProof of performanceConsumer insightsAd optimizationReturn on investment
  • 4.
    Digital Retail Media:APerfect LaboratoryLocated at point-of-purchaseCan be interactiveAddressableLow-cost productionEfficientCategory exclusivity
  • 5.
    Experimental Designs forROI AnalysisMatched panel designTrue experimental design that measures the effect of an advertising variable compared to a control during two observation periodsLatin square designUsed to measure the effect of multiple advertising variables without the use of a control
  • 6.
    Matched Panel ExperimentsAmethod for testing the effect of a campaign variable (advertising creative) on brand metrics (brand equity, purchase intent, sales) in a real-world environmentSmall scale testing minimizes the risk and costs associated with a full roll-out
  • 7.
    Matched Panel MethodologyLocationsare matched on traffic, category sales, demographic variablesLocations are grouped Control cell (no advertising)Test cell(s) (advertising variables) Sales are compared on a pre vs. post / test vs. control basis Control StoreMatch!Test Store
  • 8.
    Matched Panel Example:ConvenienceStore TVProduct: Energy drink brandVenue: Regional c-store video networkObjective: Positive ROIMeasure: Net incremental revenues (increased profit minus media and production costs)
  • 9.
    Matched Panel Example:ConvenienceStore TVPre-PeriodCampaign-Period+ $6,000Net RevenuesDifference in Sales
  • 10.
    Matched Panel Example:ConvenienceStore TVAdvertising ROI AnalysisNet Revenue Difference $6,000- Media & Production Cost $5,000= Return on Investment $1,000
  • 11.
    Latin Square ExperimentsAmethod for testing the effect of multiple campaign variable (advertising creative) on consumer response without using a controlMinimizes the number of venues needed for testing AdABC
  • 12.
    Latin Square MethodologyLocationsdo not have to be matched, but should be representativeEach advertising variable is introduced in every test cell during a different period of timeCustomer response is measured for each combination of advertising variable and time
  • 13.
    Latin Square ExampleProduct:Checking account promotion for a national bankVenue: Doctor’s office waiting area video networkObjectives: QR code scans direct to account application, new checking accounts openedMeasures: Number of QR code scans
  • 14.
    Latin Square Example:ComparingOutcomesTotal QR ScansCreative A: 60Creative B: 75Creative C: 45Creative B is most effective!
  • 15.
    The Proof isin The NumbersOur Services Audience Measurement and SegmentationAdvertising Sales and Marketing StrategyConsumer Insights for Content GuidanceAdvertising Campaign ROI AnalysisResearch MethodsSyndicated Research SourcesMatched Panel AnalysesConsumer Intercept SurveysObservational StudiesClients we ServeAdvertisers and AgenciesMedia Network OperatorsRetailers and Venue OwnersInvestorsJohn MorganFounder and PresidentRetail Media [email protected]