The Death of Last Click Attribution & Its Impact
For this revealing session we’ve got moderator Craig Macdonald, VP of Marketing & Product Development, Covario. Our speakers are:
- Adam Goldberg, Chief Innovation Officer, ClearSaleing
- Robin Smith, Global Search Marketing Manager, Life Technologies (formerly Invitrogen)
- Gary Milner, Global Marketing Director, Interactive and Teleweb Marketing, Lenovo
- Mark Grote, Sr. Search Advertising Manager, Microsoft
Craig says there’s an old adage: success has many founders but failure is an orphan. If you asked a marketing manager what the result of their campaigns is, everyone would want to take credit for the wins. The issue is attribution modeling and being able to assign credit in multi-channel programs. How do you know you have the right metrics in place to put the right resources into the funnel?
They don’t expect to give us an answer today, but there’s a lot of solid work being done in attribution modeling. How do you coordinate systems and data for attribution modeling. How do you do testing to see the impact of what you’re doing on the conversion funnel? And one overlooked issue they’ll try to approach is governance.
Is paid search overrated and what are the panelists’ organizations doing around attribution modeling. Gary starts by saying if you go back to John Wannamaker’s quote, the traditional ad industry has been rife with wastage. The strategy is to target ad dollars on perceived targets but they’re not always the right ones.
About 15 percent of media investment is in digital. Forrester also expects the number to go to 25 percent in the next few years, but Gary thinks that’s an underestimation. Last click measurement may lead to poor investment decisions and slow the digital investment. Paid search is getting too much or too little credit.
- Metrics set up as first/last/multi-click
- Organic: is it set up as a campaign?
- How do you measure affiliate?
They did a study to get an idea of how significant the cross-vehicle interaction is in terms of visitors and revenue exchange between marketing programs. They found about half of visitors came directly to the site. They probably heard about the URL elsewhere. About 30 percent is coming from a single touch point, like an affiliate program. 20 percent comes from multi-touch point approach, with the visitor seeing the brand in several places before coming to the site.
So what they learned is that they’re tracking last click as well as an attribution revenue stream based on lifetime value. Set your systems to measure last click and pervasive lifetime value of traffic. Consider third-party solutions to track. Implement a dashboard system for last click and participation revenue. Consider a split dashboard with revenue generation and influence programs.
Robin is up next. She says at her company the standard last-click ROI crediting is used. But she asked, “Is paid search selling or is it an order taker?” She assumes that most spend is in platform or brand terms. She wondered where to spend money if she got additional budget. She wanted a holistic approach to campaigns. She believed she wasn’t making educated decisions and that allocation could be critical to the success or failure to the campaigns. What looked highly profitable could turn out very unprofitable, and vice versa.
In order to calculate ROI correctly and make better decisions, she set out to:
- Justify large budgets for entire purchase funnel and to bring more budgeting dollars online
- Gather all visit data and assigned every visit a unique ID
- Analyze the data and initial findings
- Choose a reallocation method and look for another tool that could do it for us
- Understand that balancing the redirection of revenue is for improving ROI
The reality of ROI crediting should be multi-click. She found that in reality they’d been ignoring over 80 percent of the touch points leading to the conversion.
There are always challenges and risks. It’s very complex because of the amount of data you have to look at. Even with analytics today, it’s hard to make actionable decisions. Sometimes you’ll never get answers to some questions.
- Technical and resource limitations
- You can easily find yourself in a data tsunami
- Creating clear goals on what you want to achieve
- Minor changes in weight or a variable in formula can make significant changes in performance results
- As a search marketer you many not like what you find
Mark will talk next about Microsoft’s approach to last-click attribution. He asks who in does offline marketing? It’s about 25 percent of the audience. He thinks paid search is indeed overrated. He thinks that search does a little branding, a little awareness, a little selling, but it’s really at the bottom of the funnel. What we’re really looking at is SEM on steroids. We all like the 100 percent accountability of search. He touts that benefit all day long. Nowhere else in marketing is there such clear justification.
But what’s unique about Microsoft, they have multiple goals and layers in a given campaign. Search is one layer, but what’s happening elsewhere in the campaign? There’s a display ad, but we don’t know how that affected a viewer. From a search marketer’s perspective, how do you even begin to analyze that? Add to that multiple languages, multiple campaigns and products, it quickly becomes more complex.
He asked who was in offline marketing because he says any perception shifts are not being measured or even recognized. The solution will have to be through technology. Cookies can probably be used better. But anyone going offsite or off-network, that’s problematic because they lose control of the user. Webtrends and Omniture are used for analytics, and that adds another layer of complexity.
Adam will wrap up the presentations for us. He says that parts of search are and aren’t overrated. To move up the chain of better attribution, he researched the fears and concerns marketers have in attribution management.
Evolution of Attribution Management:
When a client moves from last click to even attribution, they see a lift in performance. The mathematical model is very complex and still developing. Uncertainty is a factor that accounts for things that can’t be tracked yet.
They start by looking at the purchase path of a consumer. Problem recognition leads to information search, then an evaluation of alternatives. The purchase decision is made, then the purchase. When the person navigates to the site for that purchase, all the previous searches and influences aren’t seen. But influencers, introducers, etc., are assigned value in attribution modeling.
When you first start with attribution, you may think you don’t have introducers or influencers. But that’s often because you’ve killed everything but the closers due to the last-click model. You may have to reintroduce them so you can measure their effect.
What’s the biggest challenge?
Adam says first it’s a tech problem. You have to be able to see your path. You’re going to have to put tracking in place in order to get beyond the last click. With the tech in place it could be a testing problem.
Mark says you have to figure out the problem you’re trying to solve and then you have to engineer the tech to help you manage the data. Start slowly with just a few campaigns, improve and streamline the processes and expand.
How’d you get the organization warmed up to this?
Robin says there’s definitely a barrier there but how it started with her organization was through the sales team. She had to set out to prove that the Web does assist in the sale. Gary says it’s not just about click-based modeling, it’s about impression-based modeling. Craig recognizes that there’s a certain maturity that the technology has to reach before attribution modeling can develop.