Triangulate econometrics, experiments and attribution over long time frames
Triangulate measurement: combine econometrics, experiments and attribution, let each inform the others, and run them over long time frames. Long-run effects can only be measured with long runs of data.
More from this source
StatSpend drives far more of effectiveness than CMOs thinkPrincipleYou control the budget, not the ROIPrincipleAds mostly don't wear out; wear-out is usually a frequency problemCaseAmazon re-ran a two-year-old Christmas ad and it tested exactly as wellCaseA Knorr ad seen about 120 times in Scotland stayed one of the country's best-loved adsPrincipleDisguised repetition is one of the highest orders of creativityWarningCalling one ad a 'campaign' loses what a campaign isWarningBig brands that market like small brands become small brandsWarning'Do more with less' usually means doing less with lessPrincipleSet the size of the marketing budget with finance, not just its allocationCaseOptrex went dark for ten years in one region and the brand nearly died therePrincipleSell brand building to finance in cash-flow language, not brand-love language