Marketing Attribution

Overview

The marketing strategy in today’s world almost always involves a combination of multiple online and offline channels. Each of these channels plays a certain role in bringing a prospect close to the brand. A typical customers journey would involve exposure to one or more ads across multiple platforms over a period. How do you then know which platform has contributed to the conversion and how much? 
Without these insights, you have no idea if you’re investing your money in the right platform or not.

Multi-touch attribution (MTA) or Marketing Attribution is a marketing measurement method that evaluates each touchpoint’s impact in driving a conversion, thereby determining the value of that specific touchpoint. It is essentially a bottom-up approach to measuring marketing efficacy (as against market-mix modelling, which is more of a Top-down approach).

The goal is to understand the performance of your campaign across multiple channels and decide where to spend marketing funds to improve conversion. The granular approach of attribution modelling means that data is analyzed frequently, in real-time, or as close as possible.
MTA focuses on digital channels that directly connect with customers/prospects in the form of a click or impression. The series of clicks and Impressions events are stitched together to work out the customer’s journey. 

The last touchpoint in the journey is the success event or conversion. This could be any desired action taken by the user.  Giving contact details, downloading a white paper, or buying a service or product are all examples of conversion.

Why should you consider MTA?

You can use multi-touch attribution to streamline your marketing tactics and improve ROI. Given below are some of the benefits.

• Optimize Marketing Spend  – Get insights on what marketing spend has contributed to conversions and use these insights to plan your spending across platforms.
• Improve Your ROI  – Understand what campaign/creative resonated well with what audience, carve out more targeted campaigns and increase your ROI
• Understand the role different channels play – Some media may be more effective in building initial awareness and others giving the final nudge. Understanding the role played by these channels helps you plan your marketing strategy more effectively
• Improve your marketing Tactics – Understand nuances such as average time taken or the average number of steps taken by customers to convert and include these insights to plan your marketing campaigns accordingly
• Make Marketing adjustment more Frequently – MTA models can be executed more frequently, e.g. every week, enabling you to change strategy more often, as required.

Attribution Approach

Several types of attribution models are available. The choice of the model may differ depending on your business type, sales objective, type of product etc.  Each attribution model has different strengths, weaknesses, and focus areas.  The models can be grouped into the following three areas:

• Single Event Attribution 
• Rule Based Attribution
• Algorithmic Attribution

Consider the following sample scenario and then evaluate each of these models in that context.

Let’s assume that a user First saw the ad on paid search (Google Paid). After few days, he came across the same ad on Facebook. He searched for your business on Google (organic search) engine a week later and clicked on it.  The next day, he browsed the information on the website by coming directly to your site, registered himself but did not purchase. After a week, he clicked on your email offer and made the purchase 

Single Event Attribution 

In this attribution, credit is assigned to a single event along a path to conversion

• First Touch – The first event receives entire 100% credit for the conversion. No other events are assigned credit. So in our example above, ‘google paid search’ gets 100% credit for the conversion
• Last Touch – The last event receives entire 100% credit for the conversion. No other events are assigned credit. So in our example above, email gets 100% credit for the conversion

Rule Based Attribution 

In this approach, credit is assigned to multiple events along a path to conversion based on a predetermined set of rules. Examples of rulesets include even weighting, time decay, and u-shaped: 

• Equal-Weight – Credit is applied equally across all events and/or channels measured along a path to conversion. In the above scenario, each of the 4 events will get equal credit (25%) for the conversion. 
• Time Decay – In this type of model, Credit is applied to events at increasing or decreasing intervals along a path to conversion. Event values are usually altered based on specific time windows when the events occur. Example: 40% of credit could be given to events within 24 hours of conversion, 30% to events within 1-3 days, 20% to events within 3-7 days, and 10% to events within 7-14 days. 

• U-shaped – Credit is disproportionally applied to events at the beginning and end of a path of conversion. For example –  40% of credit could be given to events occurring on the last day before the desired outcome, 30% to events occurring between days 1-13, and the remaining 30% to events that happened at the beginning.

Algorithmic Attribution

So far, we have seen the heuristic models; they use a specific, straightforward approach in understanding the complex marketing cycle. While this may accelerate the analysis, some organizations may want to go beyond to an algorithmic approach to increase the accuracy of attribution 
Markov models are one of the best options from those available for attribution modelling. A Markov model is a probabilistic model, which focuses on specific calculations of the chance that interaction in one channel will transition to a different state, such as a conversion. Markov model helps you quantify the statistical impact each of the channels or campaign had on conversion.
 

Conclusion

Multi-touch attribution can help you improve the efficacy of your marketing and sharpen your game. Multiple approaches can be adopted depending on your industry, needs and data availability.  Get in touch with us to work out the right approach and the right solution for your requirements.