Attribution Modeling Key to Determining Ad Effectiveness
In November, Facebook announced a new tool for marketers: the ability to track conversions after members of the social media network...
In November, Facebook announced a new tool for marketers: the ability to track conversions after members of the social media network view specific ads on Facebook. When the news broke, a Facebook product manager told Reuters that “measuring ad effectiveness and outcomes is absolutely crucial to all types of businesses and marketers.” However, when marketers rely solely on advertising outlets to measure the efficacy of their investments, they lose the more useful data that comes from media attribution.
Certainly, it’s helpful that the social media giant can now show marketers what consumers do after they click on a Facebook ad. The conversion data can be more useful than a report that notes the number of clicks, for example. However, this tool won’t tell marketers whether buying ads on Facebook is the most profitable use of their advertising budget. In order for ad buyers to know whether it makes sense to spend money advertising on Facebook, they need a multi-attribution tool that weighs all of the marketing events that contribute to conversion. After all, most users who interact with Facebook and then convert will undoubtedly have been exposed to other marketing channels such as search, display or television. Which of these is most influential in the eventual conversion? Attribution helps marketers answer that question.
Leverage attribution to optimize marketing
Facebook and other advertising outlets are in the business of making their ad spaces look as appealing as possible to marketers. That’s why ad buyers need to look elsewhere for reliable information about which marketing sources are worthy of their budgets. Businesses need a method for evaluating every element that goes into a conversion to determine which touchpoints should get the bulk of the credit for eventual sales.
The best marketing optimization tools incorporate this kind of attribution model so that marketers can make informed decisions and shift their spending away from inefficient or fraudulent sources and toward better performing ones. In order to get a complete picture of the marketing channel and determine which elements influence the consumer’s decision to convert, marketers need to collect more data than a single outlet, like Facebook, can provide. That collection should include:
- Click tracking (client site): Solve cookie deletion and multiple devices per user issues.
- Impression tracking (display): View tags certified with Facebook, Google, YouTube, Yahoo!, AOL and others. Track viewable impressions above the fold.
- Owned media tracking (third-party sites): Evaluate traffic from Twitter, Instagram, Facebook, YouTube and more.
- Earned media tracking (recommend button): Which sources and user IDs drive “recommend” buttons that lead to conversions?
- Offline tracking (TV/radio): Cross attribute offline elements with online investments.
- Marketing cost data (application programming interfaces): These include Google AdWords API, Bing API, Commission Junction API and others.
With all of that unbiased data in hand, marketers are poised to answer the four questions critical to the process of increasing return on investment and ad spend (ROI and ROAS). Those questions are:
- Which marketing sources am I missing?
- Which marketing sources are fraudulent?
- Which marketing sources are inefficient?
- Which marketing sources are efficient?
A mix of aggregated and client-specific data enables marketers to answer these questions, and attribution modeling delivers that combination.
A primer on attribution modeling and how marketers can use it
Marketers can employ a variety of model versions when they opt for attribution over outlet-specific analysis. The predominant types are rules-based heuristic models or algorithmic regression models. However, other options include predictive and media mix modeling, or concurrent use of two or more of these varieties.
Once an attribution model is in place, marketers can export raw data to a business intelligence tool or automatically convert data into any bid tool or major demand-side platform such as Marin, SearchForce, Mediamath or Kenshoo. The resulting data sets the foundation for tactical marketing actions that grow revenue and decrease cost.
The Facebook pixel is a great step forward in connecting the users who interact with Facebook’s ads and those who eventually convert. However, this single outlet’s tool does not provide marketers with the information they need to make the best media buying decisions. Holistic media optimization – which includes TV, radio, direct marketing, social media, email, pay-per-click (PPC), affiliates, comparison shopping engines (CSEs) and display and retargeting ads – is far more effective in arming marketers with the data required to boost ROI and ROAS. Such a strategy can cut the cost per lead by 25 percent and improve revenues and profits by 35 percent. These are the kind of results that can only be generated by a complete view of marketing assets, not just data from one advertising outlet.