In theory, Multi-Touch Attribution (MTA) sounds like a dream of a buyer come true.
Since digital media leave a trace message or interaction, and since machine learning can find patterns in all this data, the idea is that MTA can determine the impact the differential of each ad, key phrase, video, or downloaded white paper that led to a sale or other action.
This could be a boon, given the increasingly complex customer journey, that potential buyers go through online and offline contact points, often researching and making decisions about their business. 39 purchase before meeting a seller.
"If [MTA] was" Everyone would like the concept, "said Jeff Bickel," everyone would like the concept, "said Jeff Bickel, but it is complex to understand and implement he stressed.
Accounting for the immeasurable
On the one hand, he noted, the MTA is a workable concept as long as "everything is contained online". But many marketing campaigns and interactions include offline contacts, where it becomes complicated to capture and integrate this data with digital interactions.
Given this complexity and the need to understand the value of marketing expenses, he said, marketers are constantly evaluating whether their investment in a particular attribution solution is profitable , noted Bickel.
This raises a recurring question for the MTA, he said: "Is the juice worth it?"
Offline data is not the only problem with MTA, though sure. Brian Baumgart, Co-Founder and CEO of the Logic Conversion Assignment Provider, points out that different industries, such as manufacturers, have different attribution requirements when you evaluate the individual impacts of the touch points.
He also points out the problem. what he calls "incommensurability" – factors such as existing brand equity that influence a purchase decision. Apple has an intrinsic advantage in influencing a decision to purchase compared to a brand new manufacturer of consumer products, for example, simply because it's well known.
To account for this factor and other factors, Conversion Logic's MTA implementation involves custom data models for each type of enterprise and advanced algorithms using a advanced learning and the own Conversion Logic identity graph, complemented by Tapad, LiveRamp and Oracle graphs.
Tony Marlow, CMO Data and Marketing Services Company Infogroup, believes that MTA is "essential" for his business. Working with attribution partners, he particularly enjoys the temporal degradation model, as he gives more credit to recent messages and helps to report brand messages over the years, such as those of automakers. There are several standard attribution models .
MTA and Customization
But he warns that MTA is not "a quick fix", indicating additional complicating factors to be taken into account.
Marketing personalization, for example, aspires to convey a specific message to a user at a given time and place, often optimized in real time.
"If you personalize Marlow said:" This creates a challenge for MTA because the model must take into account the impact of an ad for a specific consumer ice cream shop. around the corner, while he's just engaged in losing weight. The finer the customization, the less it seems to fit into an attribution category.
However, for Bayer, vice president of media strategy and platforms, Josh Palau, the impact of even a highly personalized creation can be measured. "If [the attribution] covers the creations, we can see that an ad Watson [personalized] offers twice as much as a graphic ad."
While MTA was sparking "a lot of mischief." "Claritas Bickel's excitement explained that this required both massive and granular data and that it was" complex enough to understand, "explains Bickel de Claritas, two reasons why he is gaining ground especially in large companies , he says, because they have the required data and qualified personnel.
But medium-sized and small businesses "do not do it all", he said, because the first, the last or the first and the last touches "are sufficient for many places." The path of least resistance, he added, is to go with "what that they can touch and feel. "
First, last, to Yourself
Mark Sullivan, CallRail's Demand Generation Manager, agrees with Claritas member Bickell, citing his belief that there was "not a lot of understanding "or using the MTA in small and medium enterprises.
His own company, which offers telephone analysis, is challenged to obtain and merge data from trade shows with online distribution and marketing. While CallRail's marketing includes Facebook, where it tracks the tracks via UTM tracking on links in ad campaigns, it's mostly focused on the first and last contact.
Typically, he explained, a user interacts with CallRail. at a trade show, you may later click on a Google ad and visit the company's website, fill out a form and, two weeks later, click on the company's listing on the site. G2 Crowd review site that led to registration for a free trial. But only the salon and the G2 / free trial list are taken into account.
Similarly, the Radius B2B customer data platform uses the assignment to the first and last keys, but with the addition of a credit to the customer. team or member of the team.
Resources on Paid Media and Attribution
Values of Commitment
Merijn te Booij, Director of Genesys Call Center Provider , I've told how his long distance business company created his own own MTA solution.
At first she came and went with the assignment first and last contact, which did not provide the desired indication of return on investment. Then he gave 40% credit to the first touch, 40% to the last and 20% to everything in the middle.
Finally, he decided to create commitment values, each touch having an impact of 10%.
Using machine learning, Genesys has processed its extensive set of contact-against-results data over the years, so that a comparative engagement value can be determined for each touch or sequence of contacts. Similar sets of engagement paths have often led to similar results, he said, even if there could be 30 contacts in a course. The new sales to existing accounts, of course, have followed a series of better known contacts than, for example, the course of engagement of a new account.
Similarly, the .ME domain extension provider has also concocted its own multipoint mix. it's "somewhere between looking at the numbers and trying to interpret their numbers," according to Natasha Djukanovic, director of marketing,
The specific complications of the .ME society are due to the fact that many of its sales are made through partners. also sell similar products and carry out joint campaigns.
For .ME, Google Ads campaigns generate the best conversions, but they are only worth 10% of credit because they work "more like a trend predictor" of multiple campaigns through other channels, she said. Google Analytics ends up being the most valuable tool because it allows .ME to determine referral paths from partner websites.
"The Fool's Run": Enter Incrementality
Given its complexity and data needs, however, some scholars marketers see MTA as a stalemate.
"The multitactile is a dupe race that will die in the next few years," Patrick Tripp, a customer data platform, told me. by e-mail.
"Because many points of contact with consumers are digital," he said, the idea is that they "are all traceable".
A sufficient amount of data would allow marketers to "create models to tell you exactly what impact each touch point (paid or organic) had on you," he said, and the result would be a measure of the resulting ROI for each marketing component of a complicated customer journey.
The problem, Tripp Sai d, is that "gathering the depth of data required to measure each point Contact information influencing the decision to purchase a consumer is an insurmountable task in the real world. "
In addition, he adds," the most intelligent brands and the most more sophisticated "I want to know:" How much new net revenue can I generate through specific marketing efforts? ""
Measuring the gradual increase in your campaigns will replace MTA, "he explains. , adding t that incrementality was previously ignored because difficult to implement.
Such a scientific experiment, incrementality in marketing executes a test group that receives the campaign or a variant thereof, while a control group receives no campaign or another. This A / B test can then determine the net effect of the entire campaign, without having to assign values, such as brand value, hard-to-grasp data, or highly customized messaging.
"Multi-touch Attribution was a dream of marketers," says Tripp, "but it's impossible to implement them in reality.
"Incrementality can not only be implemented, but it actually solves the problem of marketers trying to understand: how many activities do I have when I'm doing an X activity through compared to the number of cases I get when I do not do activity X? & # 39; "
This story was first published on MarTech Today For more information on marketing technology, click here.
About the Author
Barry Levine covers the marketing technology of Third Door Media and previously covered this space as a senior writer for VentureBeat, and he has written on such technical topics, among others, for publications such as CMSWire and NewsFactor, he has founded and directed the website / unit of PBS Thirteen / WNET, and has worked as a senior producer / scriptwriter in line for Viacom, created a successful interactive game, PLAY IT BY EAR: The first CD game, founded and directed an independent film, CENTER SCREEN, based at Harvard and MIT, and served for five years as a consultant with the MIT Media Lab, you can find it on Li nkedIn and on Twitter on xBarryLevine.