The consumers’ path to purchase has grown increasingly complex, with the vast majority of paths to conversion occurring over multiple sessions and multiple marketing channels. So how can you effectively attribute value to each channel?
Measuring the effectiveness of marketing channels has always been a necessity. But with the rise of multi-channel marketing  – and the growing split between online, offline, earned and paid media – it is now even harder to track a consumer’s full journey.
Many businesses are essentially marketing blind, unable to answer the question: which channel produces the highest ROI?
Consider the number of digital marketing channels that most brands use: organic search, paid search, referral, social media, email marketing… and you can start to imagine the various number of ways that consumers may complete a purchase.
Take for example customer A, who wants to buy makeup.
- She searches for ‘bronzer’ on Google. Your brand pops up first (courtesy of your PPC campaign), and she clicks through and has a browse on your site, eventually deciding not to purchase. This action constitutes consumer A’s first exposure to your brand.
- Because of this, a few days later, a targeted Facebook ad pops up on her newsfeed which, again, she clicks through to browse your website, even adding products to the cart, but decides not to purchase.
- Finally, she enters your website through a direct search and purchases her products.
Here we see three touchpoints in this customer’s path to purchase. But which one is ultimately responsible for her decision to finally purchase?
And, in a monetary sense, which of these touchpoints can we consider most worth the money that has been spent on it?
Some may argue that the PPC is worthless because it didn’t lead to an immediate purchase, whilst others (particularly Google AdWords) will contend that spending money at this stage is vital to increasing brand awareness.
So how can companies figure out the right answer?
Why use an attribution model?
When done right, attribution modelling can provide holistic, accurate information about the financial return your activities are delivering – allowing you to adjust what you’re doing, and utilize your budget effectively to deliver more value to your business and your customers.
However, many of the current out-of-the-box attribution models don’t actually enable you to achieve this. Let’s take a look at some of these standard models, so we can understand where they fall down.
To keep this entertaining, let’s use soccer as an analogy to compare the different types of attribution.
Last Click: This model attributes 100% of the goal to the last player who interacted with the ball.
So, Player F gets 100% of the value.
First Click: This model attributes 100% of the goal to the first player that interacted with the ball.
So, Player A gets 100% of the value.
Time Decay: The players who touched the ball last get the largest amount of credit attributed to them, whilst the further back players get smaller values.
E.g. Player A gets 2.5% of the value, B gets 4.5%, C is attributed 8%, D gets 10%, E gets 25%, and player F gets 50% of the value assigned to them.
Linear: This model gives equal attribution to each player on the path to the goal i.e. every player who interacted with the ball is given equal credit.
E.g. All players are attributed 0.16% of the credit for the goal.
Position Based: The first and last players (A and F) are credited with 40% of the goal, and all other players (B, C, D and E) sharing the remaining 20%.
These various models disregard the fact that a game up is made up of thousands of touches, and that we can only understand the value of each player by questioning what happens to all the other players when we take them out.
It is the relationship between players that’s key, not necessarily any one individual player.
Introducing data-driven  multi-channel attribution
By identifying a set of unique user events that contribute in some manner to a conversion, this model assigns the actual value and cost to each of these events. This means that you are able to truly understand the revenue generated from individual channels, and subsequently understand whether they are cost effective.
It is this ability to drill down to the granular detail that makes multi-channel attribution so impressive. In doing so, you gain a true reflection of your costs and revenue; rather than relying on the proxy measure of how your marketing channels are performing that out-of-the-box attribution models gives you.
So, in a game of football, a multi-channel attribution model takes into account the relationship between each player, and how they each contributed to the final goal. It can also account for any offline influences on the team, such as the manager’s chosen strategy, or the team’s financing.
And, in the case of customer A, multi-channel attribution will take into account the relationship between all three touchpoints, and assign revenue accordingly.
Understanding the knock-on effect
When we compare data-driven, multi-channel attribution to things such as first and last click attribution, it becomes clear that the current out-of-the-box models don’t provide anywhere near as much information as businesses need when making decisions about their marketing channels.
With this level of granular detail, multi-channel attribution modelling allows marketers to truly understand the knock-on effect that each of their channels has on the others. To do anything else is to make decisions about the entirety of a marketing campaign or budget, using information from only a small part of it.
If you changed your PPC campaign, how would it affect Customer A’s natural search and purchase decision? If you took Ronaldo off the pitch, how many goals would your team score?
A 2011 study by Forrester and iProspect illustrated that consumers who see a brand’s display ads are more likely to search for that brand and/or category afterwards.
It is really only by looking at a situation – be it a football match or a consumer’s path to purchase – in a multi-channel way that you will start to gain a granular understanding on how customers are progressing through their multi-channel, multi-device journey.
It emphasizes the interplay between channels, rather than choosing an arbitrary measure of credit.
A multi-channel attribution model can help marketers avoid fixing things that aren’t broken, and instead focus their attention on those that are.
It allows them to identify which channels they are wasting money on, which ones are pushing customers through the purchase funnel, and where their budget can be re-distributed to optimize conversion rates.
This is Part 1 in a series of articles on using data-driven attribution in your marketing campaigns. Check back next week for Part 2, where we’ll look at the barriers to attribution and how to tackle them.
‘Customer engagement’ is a concept so nebulous, and so often used, that it’s lost meaning in the marketing world. What is customer engagement, what metrics should you use to measure it, and what tools are available to help you capture the right data?
In 2015, Swedish manufacturer Scania decided to overhaul their digital capabilities. Two years on, their Head of Online reveals the benefits and challenges of digital transformation, and provides some useful advice for brands in the same position.
As a tool for understanding the customer experience, feedback should be integral to your marketing strategy, but making sure that feedback leads to organizational change isn’t easy. How can businesses effectively collect, interpret and action customer feedback on a large scale?
Kevin Krone, ClickZ’s Featured Digital Leader for May, talks about his 25-year rise through the ranks of Southwest Airlines: from intern to the company’s second ever CMO. He delves into the difficulties of keeping up with a rapidly-changing digital world, and which technological developments get him most excited for the future of marketing.