These days, digital marketing and analytics is ingrained in every organization’s marketing blood. Every day, a new channel, platform or vendor pops up in this domain providing a niche offering. A digital marketer is daunted with the task of understanding the relevance of each of these channels to include in her marketing strategy. You may think that, once the right channels are identified and included in the plan; everything is set. Unfortunately it’s just the beginning. Considering the fact that each of these channels or vendors report their own conversions metrics; the marketer is faced with the challenge of getting a holistic picture. For example, AdWords may report 100 conversions, Facebook may report 50 conversions and your display vendor may say they drove another 50 conversions; but your sales report will just say you had 75 conversions leaving you in a puzzled situation of why there aren’t 200 conversions (the sum of all vendors’ conversions count). This is exactly the problem of marketing attribution.
What is Marketing Attribution?
Any prospective customer is exposed to a plethora of marketing channels – starting from TV commercials to print advertisements to online media channels. In the attribution world, these are called as touchpoints. For example, an user may have seen a banner ad, searched for something and clicked on a paid search ad before making a purchase. In this scenario, the paid search vendor claims complete credit; but we know a banner ad was shown to the same person before making the purchase and she also had completed an organic search action. Ideally, some credit need not be given to each of these other channels (and paid search doesn’t deserve the complete 100% credit). This is marketing attribution – how the credits are assigned depends on the method used. For example, it could be a completely arbitrary approach or based on some algorithms. Before giving credit to each of the touchpoint, it is ideally to understand the influence they had on the final purchase. For example, the influence by a banner ad may be less compared to the paid search ad or vice versa. It depends on various factors like when these channels were exposed, the time difference between engagement points, the time lapse between the touch point and the final conversion and so on. The ultimate aim of attribution is to help the marketer understand the channels or media that can earn the highest ROI. This is becoming more appealing these days with the advent of niche media channels, and cross-device positioning.
Attribution has become the buzzword in digital marketing world; and so is each and every vendor claiming their offering on attribution. Minimal level of attribution can be achieved with even site side tools like Google Analytics. More sophisticated attribution solutions are provided by full-scale players like Visual IQ, Convertro, Adometry etc. These sophisticated vendors build their attribution solution based on machine learning algorithms. The resources provided in the end gives a good detailed introduction and thoughts on marketing attribution. Let’s quickly have a bird’s eye view of various attribution methods in the industry.
Types of Marketing Attribution
Most of the simple attribution models are based which even to give the credit to – whether the last click or the first click. Last click based attribution was once the de-facto standard. In these the entire credit is given to either the channel which generated the last click or the first click. This can be expanded to the touchpoint considering impressions as well. The second set of attribution models builds on how to assign credit to all the touchpoints in a customer journey. It could be as simple as giving equal credits to all the touchpoints in the customer journey to customized ones like giving more credits to clicks viz. impressions and so on. Algorithm based attribution builds on top of this assigning credits for various touchpoints taking into account multiple factors like time lag, nature of the touchpoint, time lapse between touchpoints and using various statistical methods. Various machine learning algorithms may also be used to attribute the right credit for each touchpoint. Sophisticated attribution market players build on top of this providing other capabilities as well like online-offline media optimization, cross-device attribution and TV-Online-Offline mix.
Techniques used in marketing attribution
As mentioned already, marketing attribution has evolved and become sophisticated these days with niche players entering and existing tools/players trying to add features and capabilities. For example, Google Analytics provide various attribution models like last interaction, first interaction, Model Explorer and so on. Attribution players that offer off-the-shelf solutions uses probabilistic and deterministic methods to derive at the credits. Algorithms used by such tools are sophisticated with various learning components; but the underlying pillars may be basic statistical methods like regression models.