Maximising business value from marketing expenditure – should you view it like a waterfall?

How easy is it to accurately evaluate the real business value of marketing expenditure? And how important is it to recognise the scale of the gap between different approaches to marketing attribution?

It’s something we at UniFida have been thinking about a lot, and discussing with our clients. This ‘Value Waterfall by Channel’ chart we put together for one of our clients helps answer these questions.

At the top of the Waterfall is the view taken by Last Click or single source exponents who expect us to believe that because a certain event took place in a customer journey before a sale was made, the total value of the sale was contingent on that particular event.

For online activities, Last Click attribution is still used by default in Google Analytics and is often employed by Affiliates. For offline activity, the equivalent of Last Click reporting is single source coding, such as when a call centre allocates a single source code to a sale, e.g. a Press code, or in direct mail where a sale is counted as a response because it happened so many days after the customer was mailed.

Increasingly inaccurate

Last Click and single source coding are becoming increasingly inaccurate because customer journeys to a sale are more complex and involve an increasing number of marketing touchpoints.

The next layer down in the Waterfall, and closer to the actual water level, is multi-touch attribution, or MTA. This approach takes into account all the steps in the customer journey, not just the last, and uses a scoring system to attribute value to each of the steps according to the role they played in the sales funnel.

Different MTA platforms use different algorithms to attribute value cross the steps, but the result is much closer to reality than Last Click.

Although excellent in theory, MTA ignores the fact that indirect channels such as TV, OOH and Press can be very important contributors to brand building and supporting a purchase decision. And many other factors will also be influential, such as a person’s natural purchase cycle, the time of year, their level of disposable income, the weather, and so on.

The role of econometrics

Getting closer to reality, and much closer to the actual water level, is the approach to marketing attribution that allows these other factors to be considered. This is best done by combining econometrics (MMM) with MTA.

The econometrics model can be used to scale back the value contributed by the direct channels, as reported by MTA, to a lower level. Econometrics describes the incremental impact of marketing in the context of how all factors are influencing purchasing decisions, combining all marketing channels’ impact with the influence of all external factors. In reality this will usually involve cutting back the reported impact of marketing by a significant fraction.

However, a final step needs to be introduced to get right down to the actual financial benefit contributed by marketing, and hence to the true water level, which is to allow for the cost of sale. Many marketing reports show the impact of marketing on sales rather than on the correct business metric, which is its impact on value contributed. The cost of sales, or cost of goods as it is often called in retail, will be what the company has to pay to create the product that it’s selling.

Live example

Our Waterfall chart is a summary of a recent live attribution modelling exercise we conducted for a large retailer, who combine online sales with bricks and mortar retailing. It shows a massive difference between the value reported by Last Click and the actual impact of marketing on the bottom line.

It’s interesting to note that for Google Pay per Click the indexed value was reduced when MTA is introduced compared to Last Click, whereas for Direct Mail the value increases. This is explained by the fact that Pay per Click is much more likely to feature in the customer’s journey near the end of the sales funnel, whereas Direct Mail is more likely to feature at the start of the funnel where it is very good at initiating a sale.

Our conclusion is simple. If you want to truthfully report the business value of marketing expenditure, it is absolutely essential to avoid false metrics such as Last Click or single source tracking. 

It is important to get close to the bottom of the waterfall where all channels, plus external factors and the cost of sale, are factored into your reporting metrics. To do this properly you may have to combine both MTA and econometrics!

For more information visit https://www.unifida.com.au/marketing-attribution