ellipsis flag icon-blogicon-check icon-comments icon-email icon-error icon-facebook icon-follow-comment icon-googleicon-hamburger icon-imedia-blog icon-imediaicon-instagramicon-left-arrow icon-linked-in icon-linked icon-linkedin icon-multi-page-view icon-person icon-print icon-right-arrow icon-save icon-searchicon-share-arrow icon-single-page-view icon-tag icon-twitter icon-unfollow icon-upload icon-valid icon-video-play icon-views icon-website icon-youtubelogo-imedia-white logo-imedia logo-mediaWhite review-star thumbs_down thumbs_up

The problem with click-based attribution

The problem with click-based attribution Yair Halevi

A recent study by Econsultancy and Google Analytics indicates that a majority of marketers are still using last-click attribution to measure performance-oriented campaigns. But at the same time, marketers are realizing that it is ineffective at gauging the true influence campaigns have on the consumer's path to conversion. For online video campaigns, be it advertising, email, or on-site, last-click attribution suffers from similar inaccuracies.

To quote Morpheus, it seems that many marketers are still on the blue pill.

Marketers know something is wrong with last-click attribution, but they still view the world through this lens. And while ignorance is bliss for many in "The Matrix," for marketers, it can be a bane to budgets and returns. Because evaluating campaigns and understanding cross-campaign interactions are critical in assigning budget and planning the marketing mix, companies embracing video strategies employ a different approach.

Attribution modeling

Last-click attribution assigns the entire value of a conversion to the last campaign the consumer clicked prior to converting. Naturally, last-click attribution favors campaigns that influence the lower end of the funnel. In contrast, first-click attribution favors campaigns that influence earlier in the funnel. Both are simple to use, widely supported by analytics tools, and are still the most common attribution models used, as depicted in the below chart from the Econsultancy and Google Analytics study.

The problem with click-based attribution

However, both models do not accurately capture the real influence a campaign has on a purchase. February 2012 data from Adobe shows that for search, there is a 38 percent increase in assigned revenue when moving from a last-click model to a first-click model. Similar results are shown for social, hinting that these channels are more influential in generating awareness than in triggering a buy decision. This example teaches us three important lessons:

  • The choice of attribution model can change the perceived value of a campaign, and hence affect future budgets, affiliate payments, and the marketing mix.

  • Attribution models place value on influence points in the funnel. Any model that puts all value on one point, as last-click and first-click do, will under-value other influences. Both last-click and first-click completely under-value marketing efforts that influence the middle of the funnel. In any case, research implies that models that assume a single touch-point influence are not great at estimating value.

  • By comparing several models, we gain valuable insights into the type of influence each channel and campaign has on revenue.

The good news is that marketers today are more aware of the shortcomings of last-click and first-click attribution.

Modern attribution modeling tools allow marketers to explore and evaluate alternative value distribution rules such as linear, time decay, or fully customized approaches that take into account channel types, keywords, and interaction (check out Google's Attribution Playbook for more information on these models).

Video attribution

A strategy for video attribution offers new potential to execute performance-driven marketing -- in advertising, email communications, or on-site conversion optimization. This is a paradigm shift from traditional video marketing, which placed more focus on brand messaging than online, conversion-driven performance.

This shift means that online marketers need to adopt adequate performance metrics. Last-click attribution models used for video programs suffer from the same shortcomings described for other campaigns. Modern attribution modeling techniques and comparing between multiple models are effective ways to gain better insights into smarter digital advertising strategies.

For video, however, any click-based attribution model under-values the effect of merely viewing the video. Being a more personal, intimate medium that captures consumer attention and engagement for longer periods, video has more potential to influence a buy decision at distance, with or without a click.

Consider omni-channel strategies, where tactics contribute to online and offline impact. Video advertising measures incremental orders, but there's also brand impact that drives value beyond the 15 to 30 days after a viewer sees the ad.

Use control-group analysis with attribution

It is important to remember that attribution models assign value to campaigns based on the marketer's view of the weight of each point in the funnel. As such, it is subject to effects of intuition, politics, budgets, and corporate structure. But it is not a true, bias-proof measure of value.

To get a true value measurement for a campaign, enhance your toolbox by using a control-group methodology. Control-group analysis delivers a true measure of a campaign's effect, regardless of clicks, funnel position, or personal taste. Not every marketing channel supports this approach (e.g., organic search). However, many marketing channels do, and marketers would be wise to select technologies, vendors, and processes that enable it.

Use control-group analysis together with attribution modeling, not instead of it. Attribution models are still important for analyzing behavioral segments, optimizing the campaign via creative selection and smart targeting, and understanding cross-campaign interactions using control-group testing. However, using control-group testing can take much of the guesswork out of attribution modeling. If you employ both tools, you can use the results of control-group analysis to tune and validate your attribution models.

So, it's time to take the red pill and see how deep the rabbit hole goes. The result is flexible attribution modeling combined with control-group analysis for true, no-bias value measurement.

Yair (Spock) Halevi is chief architect of SundaySky.

On Twitter? Follow iMedia Connection at @iMediaTweet.

Yair Halevi (Spock) is Chief Architect of SundaySky, heading the architecture design and next-generation technology innovation for SundaySky’s SmartVideo platform. In this role, Spock is focused on pushing the envelope for the company’s...

View full biography


to leave comments.