Intro Guide to Attribution [Online Marketing]

Google’s Zero Moment of Truth (“ZMOT”) Illustrated

What is Attribution?

Attribution is the process of identifying all potential “Touchpoints” in a User’s journey to purchase, and assigning credit to an Advertiser’s various marketing channels based on how each channel independently and collectively influenced the User’s decision to purchase.

A “Touchpoint” is typically defined as an Impression and/or Click.

What are “View-Thru” Conversions?

View-Thru Conversions are when the User views an Ad, but doesn’t click on the Ad, and then converts anyway. The argument here is that the Impression(s) alone compelled the user to navigate to an Advertiser’s site and make a purchase.

What is Last-Touch Attribution? Why is it used by so many advertisers, and how is it inadequate?

Last-Touch Attribution basically assigns 100% of the conversion value to whichever paid or unpaid channel drove the session during which a User completed a purchase.

Last-Touch Attribution is typically based on tracking parameters that are appended to the incoming URL (e.g. ?utm_source=Taboola&utm_medium=Bloomberg).

Last-Touch Attribution can be incredibly misleading, however. For example, if a User clicks on an Advertiser’s campaign with Taboola (on Day 0), and the same User then converts after clicking on a retargeting campaign (on Day 5), the Advertiser would attribute 100% of the conversion value to its retargeting campaign, not taking into account the fact that Taboola (and potentially others) played a role in driving that User into the funnel in the first place.

Last-Touch Attribution is widely used, despite its shortcomings, for a number of reasons.

  • More advanced attribution systems are very expensive, whereas the most common tracking solution (Google Analytics) is free and easy to install.
  • Marketers new to the industry basically “grow up” on Google Analytics (or something similar), and never have a chance to gain exposure with more advanced tracking solutions.
  • Leveraging and making changes based on data from more advanced attribution systems would require a very sophisticated marketing team.

If using Last-Touch Attribution, the Advertiser will only see “traffic source” in GA; regardless as to whether that User has visited the site before, or whether that User may have been exposed to a marketing campaign off-site previously.

What is Multi-Touch Attribution (“MTA”)? How does it work?

Multi-Touch Attribution takes into account all of the Touchpoints involved in a user’s journey to purchase, and assigns credit to each Touchpoint based on various factors.

The process involves using a consistent system for click-tracking and impression-tracking, in order to cookie Users at every Touchpoint across all marketing campaigns, and upon each visit to the Advertiser’s site. This way, an Advertiser can “recognize” when a User has already been exposed to the Advertiser’s marketing efforts, and then “map” the most common path to purchase, identifying how frequently (and in what order, how exclusively, etc.) different Touchpoints are involved.

Now that an Advertiser can track every Touchpoint with Users exposed to its various marketing campaigns, the next step is to understand and “model” the value of each Touchpoint.

There are several different “types” of prescriptive Multi-Touch Attribution models; the most common are “Linear” and “Position Based.”

Linear models will divide the value of a conversion equally, among each touchpoint involved in the consumer’s journey to purchase.

Position Based models will typically credit 1/3 of the conversion value to the First Touchpoint, another 1/3 of the conversion value to the Last Touchpoint, and then divide the remaining 1/3 of the conversion value equally among all Touchpoints taking place between First and Last.

Various Attribution Models Illustrated

Alternatively, an Advertiser can use sophisticated, statistical models to develop their own attribution model. This is less common, and is typically performed as more of an audit once every quarter or even less frequently.

For an Advertiser to responsibly make changes to its marketing mix (i.e. add, remove and/or reallocate budget across different channels), the Advertiser should really conduct what is called a “Hold Out” test before doing so; which means that on a subset of traffic (for a specific, and single marketing channel at a time) the Advertiser displays a Public Service Announcement (or anything unbranded) rather than its actual Ads. If this test impacts conversions negatively, that marketing channel and its value according to the Advertiser’s attribution model has been verified.

How might an Advertiser be frustrated by Attribution?

If an Advertiser is only measuring conversions via Last-Touch Attribution, then they’re not able to see the whole picture.

If an Advertiser is to depend on reporting from its various partners/channels (e.g. Taboola, Google, Facebook, etc.), then it’s very likely that the Advertiser will see a greater number of conversions being reported from all of its partners, compared to its total number of conversions actually generated in a given timeframe. The reason for this is simple… the average path to purchase probably involves multiple Touchpoints, so multiple partners are reporting the same conversions.

If an Advertiser sets a $30 CPA goal with Taboola, then sees it is paying a $30 CPA on Taboola, plus a $30 CPA via Retargeting on Facebook, it’s likely the Advertiser is paying somewhere between $30 — $60 per conversion… but they won’t know for sure.

Without a third-party Attribution System (such as Atlas or DoubleClick), or without building an attribution system in-house, it’s impossible for an Advertiser to distinguish which channels/partners interacted with the same Users, in which order, and how frequently, etc.

The most important aspect, in terms of Attribution, is following each individual User along his or her journey to purchase. Insight and strategy is then gleaned by aggregating this information.

Like what you read? Give M.Gifis a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.