Why Attribution is passé, even for online marketers
Written by :
For some years now, Attribution analysis has been king, and digital is being measured using Attribution models, which evolved over a period of time from simple attribution to the website of the transaction, to giving some credit to the others in the funnel (judgemental); and finally, to some statistical methods used to weight the impact of each website along the purchase path. Despite all of these advances, Attribution always suffered from a fatal flaw of not taking mass media, in store activity and promotion or indeed brand equity into account. This meant that all credit was given to advertising online, and the credit duly inflated. Not the ideal situation for your brand.
Moreover, the entire method depended on serving cookies to individual devices and then tracking the path to sales. This has now been upended by privacy laws, and slowly but surely, the ability to track individual paths needed for attribution is reduced.
Does this mean that marketers, especially those that spend a bulk of their budget on online, operate without even the limited knowledge that one got from attribution analysis?
Facebook, Google and other players have recognised the need for another measurement system that does not depend on individual data, but can operate with aggregate data not only from online but every other influencer of sales.
Marketing Mix Modelling (MMM) has been a trusty old work-horse of FMCG companies, who rarely have individual data of consumers (online sales being a small fraction).The data available at weekly levels for both sales and inputs that include other important factors such as promotions, competitive activity apart from offline and online advertising. This method when used by online marketers will give a better picture, as this technique can figure out the diminishing effectiveness of each media in the context of the overall environment.
This can lead to better allocation of budget between media, understanding synergistic effects of media played together, improved return on marketing through optimization, and a comparison at individual website, type of input (video vs static) for every dollar spent.
Is MMM easy to deploy?
The answer is that it takes more discipline on the marketer’s part to collect data across the board, refresh the data at regular intervals, and also have the capability to run statistical models. It requires more effort but as the old adage goes” nothing worthwhile is achieved without hard-work”. The good news is the pain goes away as the systems set in and more processes such as structured data collection, validation of models and so on are automated. The tools exist, the adoption is the next step.
Is there a right and royal RoAS?
Written by :
Krishna Kumar CS
RoAS - the single most powerful metric known to advertisers. It can enrich or ruin your ad campaign. It can make or break your brand. It is an essential ingredient to optimise your Ad spend. According to some, it is the holy grail of marketing.
Data science offers a more precise guide – computing the price elasticity at a key SKU level can suddenly make this exercise a lot easier. While we still use judgement and intuition, now our gut is better informed and therefore the decisions are much better as they are based on hard data.