How To Figure Out Attribution in an Evolving Marketing World

February 3, 2022
Jump to...

    Long ago there were two sides to the marketing coin: Brand building and direct response, and marketers had different expectations for each. For direct response they accepted its tactical constraints and embraced the accountability to business results. Meanwhile, with branding they embraced its flexibility and accepted the constraints on holding it accountable to business results.

    Then, paid social brought new branding-like formats to direct-response measurement, and streaming/OTT brought direct response-like addressability to TV and audio. And things got … confusing.

    Brand marketers started bringing their creatives and their cash to YouTube and Facebook and direct-response marketers started using addressable TV to extend their reach and scale their growth. Best practices started to merge, creating new opportunities but also tension. And a big question got a little harder to answer:

    What should be measured, when and how?

    So, what is attribution anyway?

    In short, attribution means giving credit or tracking the efficacy of all touch points on the path to purchase by giving credit for business results. Maybe that shopper on line at the big box store saw an ad while watching connected TV (CTV), then another saw the ad on social, and then yet another saw the ad on mobile—and all three made their way to the store. You might ask, is this a good thing? Is it important? Well, it is if it’s done right and everything is in order and it’s realistic.

    Attribution can be gold for marketers. It can increase ROI, help you spend more on channels that are working, eliminate ones that aren’t and enhance your personalization efforts. It can also make creative messaging more effective across all parts of the funnel.

    Can anyone really offer complete, fail-proof end-to-end attribution?

    The challenge is that no one is offering complete, fail-proof, end-to-end attribution. Many martech companies claim to offer complete attribution solutions, but it’s not that simple. The data-driven marketer needs to layer different measurement and attribution tactics to triangulate their campaigns’ impact and find ways to improve performance.

    There are a number of reasons the promise of total attribution may simply not work. For one, it’s not “portable” across walled gardens nor between mobile OS and the web. There can also be differing match rates between impressions served and transaction data. And if campaigns are not initially set up with agreed upon controls and variables across disciplines, you may be looking at apples and oranges. Additionally, cheap inventory can throw off results, as can trying to determine what’s causative versus what’s correlative when it comes to gauging success.

    What components of attribution really matter?

    Incrementality is an important attribution metric because it indicates causation and measures the “incremental” increase due to an adjacent marketing activity. For example, this could be the difference between running a solo CTV ad and then adding a supplemental but related banner ad. Knowing if marketing caused a change is more than just knowing whether a campaign impression is tied to a conversion event, and it’s also more than knowing whether additional conversion events occurred.

    Brand lift is also significant. Measuring brand lift within an experimental campaign structure is your biggest opportunity to detect incrementality and report the causal impact of your marketing. It’s also imperative to make sure you’re withholding marketing from some portion of your potential audience, and that your spending “density” (the amount of spend per audience member) is high enough to create a detectible lift in at least part of your audience. This will help make it easier to detect the difference.

    Identity-matched attribution helps tie marketing activities to certain personas, since knowing the causal impact of your marketing only matters if you can also know how to improve it. Using identity to understand conversion rates for different variables gives you that ability. Think of the variables as levers. Each variable of identity impacts what lever you can pull to improve campaign performance. Your levers might be budget allocation by channel, audience segment or creative. You can also use identity to tie in-store traffic back to a campaign for a more complete end-to-end picture.

    “Consideration” metrics are also valuable for determining attribution. Against the backdrop of today’s data-driven marketing landscape, we rightly default to holding our campaigns accountable to transactions and conversions. But research-oriented “consideration” actions can be an important part of your measurement strategy too.

    Views of deep pages, like product details or location pages, signal intent, happen more often than conversions and often happen before conversions. Because these consideration events have more scale and happen more quickly, measuring and attributing them can give you an earlier view of your marketing’s causal impact, and let you optimize your campaign faster and more completely.

    What’s the bottom line?

    Today’s marketers are under more pressure than ever to tie campaigns to revenue through measurement and attribution. While that will not be possible in all cases, there are optimization benefits to be gained. The critical element is knowing how you define and calculate attribution, and ensuring you derive value from the results.

    Originally published by AdAge on 11/2021



    At Stirista, Identity is our Identity, and we have a simple mission: help marketers generate revenue with our identity-level data. As marketers ourselves, we know that one-size-fits-all solutions and decayed data don’t work, so we built our modular and real-time OMNA Identity Graph from the ground up to arm you with marketing data that actually works.