The ROI attribution that was done for digital advertising was never anywhere near perfect. The main issue was that it attributed “credit” based on various algorithms, each of which had its own share of problems. More importantly it had the issue of last mile attribution. Since people may have checked out a brand and then bought it online, all credit was given to digital advertising. It was always patently clear that mass media that accompanies or precedes digital advertising greatly improves digital effectiveness as more people click on the link and more people also convert as they have been presold on the premise. The marketers who got carried away (many still do) with digital and shifted more budgets to it, paid the price. While digital advertising obviously works, it does not work beyond a point, unless accompanied with other media support. Pepsi marketers realized this when they went much more with digital and promptly lost share. So many digital only clients (especially with good budgets) have no idea how effective their strategy really is and probably have a good amount of wastage built in,
Now with privacy laws coming into place, cookies collecting individual customer journey is going to be illegal from end of 2023, with some browsers having already implemented it, digital attribution as we know it will be history. How then, are brands to make decisions based on data? The answer lies in what the brick and mortar businesses like CPG have been practicing all along. Colgate for example does not know who its customer is or the customer journey to the store, but the brand is still able to attribute success at a media or tactic level. How do they do this? Good old statistics and machine learning. What’s more, this method does not suffer from last mile attribution and therefore can estimate the impact of media across the funnel better as well as a synergy effect. Look up Marketing Mix Modelling, a tried and tested approach and employ your budgets more effectively.