Why You Need Cross-Channel Attribution

Recently, the multi-million dollar international advertising and public relations company Publicis Groupe announced that it would be using Adobe Marketing Cloud to power its digital marketing platform, now with a special emphasis on multi-touch attribution modeling.

In the last decade, advertising technology has exploded and as the industry continues to expand with billions of people connecting to the internet each day, the entire digital marketing ecosystem is in a stage of constant, ever-advancing flux. What is certain right now is that attribution modeling—especially more sophisticated cross-channel attribution modeling—is where the big game is being played.

It’s Not a Funnel: Don’t Focus on the Beginning or End

Modern web users bounce between multiple social-networking sites, search engine research, and email inboxes, seeing different ad units everywhere they surf. Which banner, video, or search engine result was one of the few out of thousands that stirred interest and ultimately contributed to the sale? When assessing the efficacy of an ad campaign, nuanced conversion path intelligence is essential.

And yet, recent data from eMarketer shows that two thirds of advertisers are still only using first-touch and last-touch attribution metrics; just 22% of marketers say they’re using cross-channel methods to understand the effectiveness of campaigns.

Imagine that you’re a job seeker using networking as your primary strategy. Once you’ve found work, you look back at your process and understand which personal, collegiate, familial, or professional connection from your past proved most significant. Traditional attribution models would either ascribe 100% of the credit to the first touch, which in the case of the networking scenario would be the first person you met on your path to employment; or 100% of the credit to the last touch, which would be the final person with whom you spoke.

First Touch, Last Touch?

Both models have clear limitations. The last touch method undervalues the importance of early awareness building in an ad campaign and overemphasizes the effectiveness of whatever advertising channel a lead interacted with last before conversion. The first-touch method underplays the significance of branding and the slow build of interest as distributed over many subsequent touch points.

Furthermore, it is very difficult to determine the best timespan for modeling an entire conversion path. If a customer interacted with a company’s banner ad two years prior to clicking on another ad and then buying the product, it seems fairly obvious that the very first interaction should not receive 100% of the credit for the conversion, but the line isn’t always that easy to identify. Is it good marketing practice to keep track of a conversion path for a week? A month? Two months?

A Relevant Timeline

Part of the problem with trying to pin down a definitively smart time frame is that the premise is not definitively smart to begin with. It’s quite possible that a touch point from two years ago deserves some small fraction of the conversion credit. After all, building up familiarity and interest through consistent branding is an essential part of the modern advertising industry. However, determining which touch point came first is less of a high stakes game when the first landmark of the conversion journey isn’t receiving ALL of the acknowledgment.

Some Value in Traditional Attribution Modeling

But if these older, more simplistic attribution models are so obviously weak in illustrating campaign efficacy, then why are so many companies still sticking to the bare-bone basics? Well, first touch and last touch models can be useful in very specific limited contexts.

Most importantly, though, cross-channel attribution, while increasingly crucial in today’s very technologically competitive market, poses many challenges and can be quite difficult to get just right.

Breaking Up Is Tricky

One major conundrum of cross-channel models is how best to handle direct traffic, when a user navigates to a website without clicking an ad or running a search. Usually this is a result of previous interaction with conversion touch points, but figuring out which ones isn’t easy. Basic Google Analytics assumes that the last interaction before direct traffic is the one responsible for conversion, but the Multi-Channel Funnel feature of Google Analytics records all direct traffic as its own touch point in a purchase path.

While breaking up a user’s behavior on his or her way to a purchase is valuable, direct traffic is tricky, especially when it’s the final interaction the user has before buying the product. This is because in many cross-channel attribution models, the last touch point gets a larger percentage of the credit, an issue which also relates to a bigger, more general cross-channel challenge—how much credit should be assigned to each touch point.

It is certainly the case that when a user is only visiting a website because of another more important touch point he or she interacted with previously, ascribing all or most of the credit to “direct traffic,” even if it was the last conversion path event, is not a very useful way of understanding campaign effectiveness.

Big Data, Big Algorithms

While a case-by-case basis credit distribution over a conversion path may seem obvious, modern attribution modeling software seeks to automate this analysis and apply similar algorithms across massive bodies of user data. And that means using one method at a time.

There are many different ways to distribute credit for a sale across multiple channels: linear, time-elapse, and position based, to name a few. And while all of these approaches have drawbacks and benefits, they remain the more sophisticated tools for understanding an entire purchase path than first and last click metrics.

Cross-channel attribution modeling is the only way to get a more complete picture of campaign efficacy. Understanding its variations and unique challenges is a crucial way to modernize the way you advertise.

Why You Need Cross-Channel Attribution