As a marketer, attribution is one of the most important challenges you face today in a multi-channel world. Your customers interact with multiple campaigns: They start with a TV ad, see a Facebook ad, see a Google Ad two days later, discover an SEO-driven blog article, and finally click on an email discount offer to make a purchase. Traditional marketing attribution will tell you that the email campaign converts the best and that you should drop the rest. So, following the data, you might be tempted to do just that for the next campaign, only to find that your conversion rate has inexplicably tanked. 

Until recently, marketers used a default “last-touch” attribution model for sales, attributing them to the last touch, or last click before purchase. However, in today’s multi-channel environment, this model can lead to misunderstanding the customer journey, resulting in misallocation of budgets and suboptimal tactics. The problem for marketers is how much weight to attribute to each channel to determine budget allocation and ROI. According to a survey by eMarketer, cross-device attribution (42%) and accurate measurement (36%) are some of the biggest challenges for digital media professionals in 2021. 

In the post-COVID world, the path from prospect to conversion is becoming increasingly complex. Customers navigate multiple touchpoints before reaching a sales representative or making a purchase. The pandemic revealed new buyer behavior that emphasized a self-directed discovery and education journey, forcing marketers to increase their touchpoints across channels and devices, complicating attribution. 

Multi-touch attribution (MTA) accounts for the complexity of the customer’s journey with its multiple channels, devices, and touchpoints. With MTA models, marketers have a better understanding of which touchpoints are influential through the entire journey.

Get a unified view of all your marketing data sources.

Learn more

MTA introduces a new way of thinking

Marketers have not been quick to embrace multi-touch attribution modeling, but now they are under pressure to show ROI or face decreased budgets. The reluctance to adopt multi-touch attribution is partly because of the lack of a tactical plan, said John Loury, President, Cause and Effect Strategy (CE Strategy), a business intelligence and analytics firm that was founded to provide insight-led marketing strategies for clients. CE Strategy has been helping clients lean into this new paradigm in the world of data-led marketing and has expanded into other business units such as sales, operations, finance, and HR to provide companies with a holistic understanding of how business key performance indicators (KPIs) impact one another.

“As marketers, we’re still stuck with the old ways of thinking that our campaigns must be perfect and have fixed goal posts,” John said. “But reality shows us that we have moving targets. We have to adopt iterative and agile learning methods for campaigns, and we can do this with marketing analytics. That’s why I find this idea of building algorithms around multi-touch attribution so exciting, because we can train these models and they can continue to evolve over time using machine learning.” 

He gave an example of how an iterative approach changes the way campaigns are planned in the nonprofit fundraising world, where it’s necessary to determine the right channels and the right amount of touches to get people to either show up to an event or to donate to a cause. With social and electronic channels thrown into the mix, nonprofits are looking to replace direct mail (due to growing costs) with social and digital for their campaigns.

“But campaigns often are impacted negatively from a response standpoint when the direct mail channel is removed,” explained John. “So the idea is to devise small test cases. We add a channel or take away a channel to a specific segment to measure the impact that it has and overall response. Similar concepts apply to the retail world as well. It’s all about continuously fine-tuning our mix (touches, channels, and offers) to generate the ideal response from the donor or the customer, which is different for each audience segment and evolves over time.”

Multi-touch attribution models explained

There are several MTA models, which weigh various touchpoints along the customer journey differently, and marketers must evaluate and assess which model works best for their business. Some of them are:

Linear model

With the linear model, you apply equal credit to every touchpoint along the customer journey. This is an unbiased approach, but it also fails to recognize which touchpoint was the most influential and presents a simplified view of the customer journey. This is one of the more popular MTA models for early adopters.   

Linear model

Time decay

The time decay model is typically used for promotional campaigns, with increasing weight given to progressive touchpoints along the customer journey, with the last touchpoint receiving the most weight. 

Full path 

This model places equal weight on the first, lead, opportunity, and final touchpoints, giving them each 22.5% of the credit for the sale, while the remaining 10% is distributed among the other touchpoints. This model offers a better understanding of the customer journey at the bottom of the funnel.  

Full path

U-shaped

The U-shaped model attributes 40% of the credit to the first and last touchpoints, and distributes the rest equally among the touchpoints in between. This model is suitable for those who have a short sales cycle and want to focus on the first and last touch of the campaign. 

U-shaped

W-shaped

This model emphasizes the role of the first, lead, and opportunity touchpoints over the others; these are each assigned 30%, and the other touchpoints get the remaining 10%. This model is suited for longer sales cycles with plenty of touchpoints in between to nurture prospects. 

W-shaped

Custom 

Experienced multi-touch attribution users can build their own custom model to suit their needs. It requires a higher level of expertise but offers marketers greater control and takes into account the complexity of the customer journey. 

How to implement a good MTA strategy

In the presence of multiple data sources, it’s easy to go down a rabbit hole, with no clear business outcomes in sight. John recommended starting with a business goal and tracking back the marketing touchpoints that lead to it. When you have defined your KPIs, they can lead you to understanding which attribution model will suit your business. 

Setting realistic expectations is important because marketing professionals new to multi-touch attribution often expect improvements on optimization immediately. But the reality is, it could take three to six months of observing and optimizing to see significant improvement. Marketers should also consider factors like length of sales cycles and consider all the touchpoints in the customer journey. 

Building a good data foundation for successful multi-touch attribution

Without clean data, marketers cannot get the quality reporting required to optimize advertising campaigns. Accurate, but not necessarily complete, data (AI and machine learning can help us round things out in many cases) and taxonomy are the essential building blocks of a successful multi-touch attribution model, said John. For example, unique campaign IDs are a good start but often missed by marketers. This results in bad tracking/reporting and inaccurate insights. The key to building a good strategy, said John, is to map out the data sources that will lead to the accomplishment of predetermined business goals. 

“Every campaign like this should have a checklist of the types of data that are required in order for us to gain that type of insight which will achieve goals,” John advised.

Multi-touch attribution marketing offers an agile way for marketers to respond to customer behavior and optimize campaigns on the go. Data-driven marketing is a long game. There is much learning with each campaign, and the iterative process enables marketers to combine short-term value and deliverables with long-term strategy.  

Get a unified view of all your marketing data sources.

Learn more

Vrushali Haldipur is a Content Strategist at Sisense. With a background in global media and digital marketing, she specializes in bringing tech stories and marketing campaigns to life.

Tags: |