Marketing Attribution, or Which of the Daltons Sold More Ice Cream?

Marketing attribution can be explained by a story about Ma Dalton, who sends her 4 boys to drive more clients to her ice cream parlor and wants to reward each of them based on the sales they generate.

Ma Dalton wanted to attract people to her ice cream parlor. Although it wasn’t much to her liking, she had to acquire more customers. She told her boys to get going and drive more people to the shop.

“But how will we convince them?” they asked. She gave them coupons for 15% off and promised she would reward each son proportionately to the number of coupons with their name that had been used for a sale. It was to be a fair deal.

Overview: What is marketing attribution?

In digital marketing, every customer touch point is measurable. Marketing attribution is a model for appreciating the contribution of each touch point to the final sale. Ma Dalton's boys are her marketing machine, and each one will contribute a certain amount to an increase in sales at the ice cream shop. Marking the coupons with each boy's name will help Ma Dalton identify which of her sons contributed the most -- right?

Let's see how each boy did.

Averell made a big banner stating, “Ma Dalton’s ice cream 15% off” and went to the park on the weekend. When people asked him how to get the discount, he handed out a coupon with his name on it.

Jack went to the nearby school and stood by the fence shouting, “Ice cream discounts at Ma Dalton’s,” at every break. When the school kids looked interested, he handed out his coupons.

William stationed himself daily at the corner of the road the ice cream parlor was on and handed out coupons to passersby to entice them into the store.

As for Joe, he was active on Saturday morning but then seemed to be taking a nap most of the time.

At the end of the week, Ma Dalton counted out the coupons with each son’s name and gathered her boys together. Joe came in first since he had left the coupons on a little table by the ice cream parlor door so they were easy to find.

Everyone who went into the shop was happy to find his coupon. His brothers thought he was a genius.

Data doesn’t equal truth and needs to be interpreted to provide real insights. Image source: Author

Marketing attribution in action

In the case of Ma Dalton's ice cream store, the 15% discount coupons she gave to her sons are data points. So, how can we use them to find out which son brought in the most business? Averell made people aware of the ice cream parlor, Jack targeted a qualified audience of ice cream lovers, and William found people nearby.

Each strategy represents a place on the customer journey map. As for Joe, he hacked the system for his own benefit, so let’s put him aside for now.

Imagine a family in the park had seen Averell’s banner over the weekend. This was the first touch point in Ma Dalton’s marketing setup. Imagine the kids from the same family having a second touch point at school when they heard Jack shouting about the promotion.

Finally, the last touch point was when the family went down the road one sunny day and was interrupted by William on the corner suggesting they drop by the ice cream parlor. The family was exposed to all of the Dalton brothers’ sales tactics, which contributed to the family ending up in the shop.

So, which coupon did they present? Averell made them aware of the existence of the shop, but did they ask for his coupon? Jack made the kids come home from school and beg for ice cream, but they might not have brought his coupon.

William caught the family’s attention at the right moment where they were ready to act and handed them his coupon, but they might have ended up presenting Joe’s coupon at the till.

Such is the challenge of marketing attribution. A marketing attribution model presents the marketing data you collect across the customer life cycle in a way that explains the impact of each act of communication and predicts what will work best in the future.

The 5 types of marketing attribution models

Digital marketing attribution is often explained via one of the five models below. The simplest attribution models are based on one single data point. This is the case for first-touch and last-click attribution. After those, you will find the more complex multi-touch attribution models.

Model 1: Last-click

Last-click attribution is when the last tracked action is considered to have been the source of the conversion. This attribution model is the simplest because the last click before a conversion is a data point which can be easily and reliably collected.

When last-click conversion tracking was first introduced, of course it was much superior to no attribution data. But the model assumes no other acts of communication have been important for conversion.

This is rarely true, and last-click attribution has moved from being considered as best-of-breed to worst-in-class as multi-touch attribution models have emerged.

In single-touch attribution, only one data point is considered as the source of a conversion -- the last or the first registered action. Image source: Author

Model 2: First-touch attribution

First-touch attribution is when you consider the first act of communication a user was exposed to as the one eventually responsible for the conversion regardless of any other subsequent pieces of information or elements of communication the user was exposed to.

First-touch attribution is more difficult to track than last-click despite its reliance on one major data point. It assumes that “seeing is believing,” or “discovering is converting.”

Imagine you are marketing a product which is a revelation to users when they discover it, and that no other communication will influence their decision to acquire it once they have become aware of it. That is your perfect scenario for using the first-touch attribution model.

Model 3: Linear distribution

In a linear attribution model, you will attribute the same importance to each touch point the user has had with your communication in the user journey. This works well for a very structured marketing process in which each user touch point is part of a journey to the conversion.

Since many decision processes are in reality rather chaotic, the linear attribution model falls short of many marketers’ needs, but it can be an excellent basis for starting a multi-touch attribution process.

In multi-touch attribution, all significant and available data points are considered in the model and rewarded according to each model: linear, time decay, and U-shaped. Image source: Author

Model 4: Time decay

The time decay attribution model is another evolution from the last-click model to multi-touch attribution. It considers the last action to be the most important but also rewards other touch points. Earlier actions are discounted proportionally to their distance in time from the conversion.

This model is used in relation with intensive advertising campaigns -- for example, to trigger impulse buys -- and is one of the more complex models to implement.

Model 5: U-shaped (position-based)

The position-based attribution model considers the first and last action to be the most important: The action that made the user discover the product, and the action that made the user convert. Actions in between are considered to have each contributed in a linear fashion and with the same importance.

It is called “position-based” because each action is rewarded as a function of its order in the line of tracked actions. And as the first and the last actions have a higher importance than the ones in between, they can be seen to form the letter “U.”

3 benefits of marketing attribution

Attribution in marketing is not for everyone. If you don’t have large data sets and you don’t deploy campaigns in various digital marketing channels, don’t bother. It will be a waste of your time.

But if you do manage large data sets, have come to acquire reliable data, and deploy campaigns at various stages in the user journey, the right attribution model can provide a lot of insight and help optimize your marketing tactics. It is an investment which can drive efficiencies and improve your digital strategy.

Let’s look at some of the benefits.

1. Joe doesn’t get to pocket all the money

Joe’s name was on most of the coupons Ma Dalton collected at the end of her marketing campaign, but he hadn’t generated one single sale. If loyal customers walked into the ice cream parlor and picked up his 15% off coupon, he actually caused Ma Dalton to generate a net loss on the ice cream she sold them.

Had Ma Dalton been able to track all the touch points during the customer journey for each sale instead of looking at the coupons, she would have been able to reward Averell, Jack, and William rather than Joe.

It would, however, require her to change the rules of the game, which is one of the major obstacles for implementing marketing attribution in large organizations.

2. Better understand users

Looking at more touch points than simply one action allows marketers to better understand user behavior. Did Averell’s banner in the park eventually drive users to the ice cream parlor? If yes, perhaps a permanent banner, or a billboard close to the park, could be an interesting marketing investment.

For ongoing client relationships, customer relationship management tools can provide a structured solution for customer attribution. These tools register customer touch points and can help identify obstacles in the way your company interacts with clients and eventually attribute value to each action.

3. Improve budgeting for advertising

Marketing attribution helps you comparatively evaluate marketing campaigns. It is being used to improve budgeting between marketing channels.

Discount coupons at the door of your shop was a bad idea. Disinvest. Perhaps William’s actions to divert traffic from bypassers generated sales, and it could make sense to send Joe to the opposite corner of the road to drive more sales. Invest more.

FAQs

• What’s the best attribution model?

  The best attribution model for you is the one which most reliably appreciates the value of each of your communication channels. While last-click is better than nothing, it rarely brings that balanced view.

  First-touch and time-decay will be useful for those who recognize themselves as having a wow product or are doing massive advertising campaigns to generate sales.

  To others, the linear or U-shaped model might be the best fit. Some advanced marketers use custom attribution models or data-driven attribution based on machine learning.

• What’s wrong with last-click attribution?

  The last-click attribution model is Joe’s coupon. In some cases, the last action before the sale will have been conclusive. In other cases, it will have been navigational and, in extreme cases, it will have been fraudulent.

  A common example is the use of brand keywords in search engines. When a user searches for a brand and then buys online, the last-click attribution model will tell you the brand search generated the sale

  In reality, the user had already made up their mind and was simply navigating via the search engine to your website. Last-click attribution simply tells a very small part of the story.

• Should you change your attribution model?

  For many marketers, the question is whether you should have one in the first place. Using last-click attribution in some of your marketing campaigns doesn’t necessarily mean you forget the wider picture.

  When you collect data, you don’t collect truth. You have to establish the model you interpret the data with and which one brings you the most truthful view of your marketing.

Appreciate each for what they contribute

Marketing attribution is the art of building a model of the world on the basis of data points you have available. You need reliable data points and a certain volume of data for attribution to bring value to your marketing.

When it does, it can help you see further into the effects of your digital marketing, uncover inefficiencies, discover opportunities, and expand where it matters.