This post is by Jon Buss, Managing Director EMEA at Criteo.
Anyone working in the marketing industry knows all too well the necessity of proving the worth of corporate communications to those at the top. With competition increasing in all market sectors, businesses are starting to bring all activities down to the bottom line and qualitative measures of impact are no longer enough for the c-suite.
Attribution modelling appears at first to be a simple solution to the problem; introducing a method of measuring the financial impact of communications in terms of business objectives, such as revenue, profits, customer retention and new business. However, the process of measuring the effects of advertising, marketing and corporate messaging on the bottom line is not a simple task, and requires multiple tools and techniques in order to establish a quantitative representation.
Communications have traditionally been measured by qualitative means; including variables like the business' share of voice within the industry, the number of visits to the corporate website, click through rates and impressions. Whilst these are legitimate aspects of the marketer's toolbox, their importance rarely translates to the c-suite where executives speak in terms of financial return on investment (ROI). Therefore, attribution models provide marketers with a tool to assist in justifying their activities and budget in terms that can be clearly understood and appreciated by the decision makers of the organisation.
The biggest barrier to understanding attribution modelling and choosing the most appropriate method for your business is getting to grips with the tools that convert qualitative metrics into quantitative figures across different channels and platforms. A summary of the most common options available is below:
- First Click: attributes 100 percent of revenue to the first consumer touch point
- Last Click: attributes 100 percent of revenue to the final touch point prior to the purchase
- Last Non-Direct Click: attributes 100 percent of revenue to the last click which can be identified as an influencer, and so excludes direct links to the purchasing site
- Linear: attributes an equal portion of credit to each point along the customer journey
- Positional: combines the approaches of First Click, Last Click and Linear, by granting a certain fixed percent to the first and last touch points, and then dividing up the remaining revenue percentage equally between the remaining intermediary points
- Time Decay: attributes increasing value to touch points the closer they are to the purchase in time.
From these options we not only have a clear understanding of the type and purpose of each tool, but we are also able to make sense of industry buzzwords such as multi-click (multiple touch points) and multi-channel (multiple communication channels). Perhaps the only term which is often used but doesn't necessarily have an obvious definition is fractional attribution. This includes more sophisticated approaches, such as linear, positional and time decay where values are attributed to each stage of the journey.
But how to choose the right model for your business?
First and foremost, it is important to take stock of your existing campaign. If you are only utilising one channel and limited communications, it is only necessary to implement a simple model, such as first click. You could try last click but whilst last click is currently one of the most common models used by businesses across the world, it is also one of the least accurate and is therefore less likely to provide representative figures for your ROI.
For ad campaigns featuring offers and discounts, it would be sensible to apply the last non-direct click model, as it is easier to identify the point at which the customer was influenced to make the purchase. If instead, you would like to gauge impulse buying, time decay will provide an accurate representation of the effects of repeated brand and product exposure. Finally, for a long-term, sustained campaign, the linear model accounts for each touch point along the customer journey, demonstrating the value of communication spread across every channel.
Ultimately each individual campaign will likely feature several of the solutions detailed above and none of the stock-standard options are intended to be a perfect fit. Rather, most effective use of attribution modelling comes from customisation; applying the most relevant measurement tools to communications data to generate the best representation of the impact of marketing activities on the financial bottom line.