
Your Conversions Will Start Using Data-Driven Attribution
Posted on 26.07.2022
Yesterday, we (our agency and most of our clients) received an email from Google with the subject “Your conversions will start using data-driven attribution on or after 8/24/22”. We’ve had a number of questions about this from our clients. So, we just wanted to write a quick article to explain what this means.
This is a change in how Google has decided to record conversions. The traditional method is for advertisers to decide how they want that conversion to be counted (aka. attributed). Most campaigns we manage use “last click” which means that the conversion gets counted for the ad/keyword/campaign that the user clicked on last. Any clicks that same user made on other ads/keywords/campaigns will not get a conversion even though they may have contributed to that user’s buying decision. Here is a great article from WordStream that explains “data-driven attribution”.
Most of the campaigns we manage are for small businesses. The ad budgets are relatively small. The ads are local (city/province). Consequently, the ad the user clicks on is the ONLY ad that the user will have interacted with up to that point. So, the attribution is irrelevant. That one ad is the “first click”. It is also the “last click”. It is also the “most clicked on ad”. Thus, this change will have little to no impact for most small business and local campaigns.
However, this change may have an impact for ecommerce websites that use a “last click” attribution model. It is causing some grief for people that manage large (millions of dollars) in ecommerce data. For example, if a user clicks on multiple ads and visits a website multiple times before making a purchase, Google’s machine learning will now decide which keyword/ad/placement/etc gets the conversion. And these campaigns may now start seeing partial conversions of 1.5 vs a single conversion of 1. Traditionally, the advertiser (or their marketing supplier) would decide how they want that conversion to be attributed (last click, most clicked, more weighting on the ads that were more recently clicked, etc). This control allows the advertiser to decide how they want to review their data which in turn allows them to decide how budgets and spending should be allocated.
Last click may actually makes sense for some highly transactional businesses such as personal injury law or plumbers where there is an emergency and the customer needs help immediately. However, last click may be less beneficial for businesses such as financial advisors where a customer may (over a number of weeks/months) gradually click on some ads, view the website, read a blog, subscribe to a newsletter, follow on social media, click on another ad, view a landing page and then click on another ad and finally convert. So, in these cases, data driven attribution may be better. But, even then, we’re limited to attribution within Google’s network and not across Meta, LinkedIn, Twitter, Google Ads, etc. So, true attribution would actually require data tracking by a 3rd party application across all platforms and not just Google’s internal machine learning.
Ultimately, this is another move by Google to take away certain controls from advertisers which erodes transparency and trust. Google is essentially an auction house. So, the controversy is really an ethical question. Should the auction house also be in charge of deciding how the auction data is reported (and from that how campaign/ad budgets should be allocated)? Or should the customers be able to decide how they want to see their data and make decisions based on their preferred view of that data? As a for-profit and publicly-traded company, it’s hard to believe that Google’s machine learning would do anything that might decrease Google’s revenue (ie. save the advertisers money). And we essentially don’t know how that machine learning works. It’s proprietary AI technology that we can’t see or manage. So, it is a legitimate cause for concern. Here is the dissenting opinion posted on LinkedIn.
But, again, time will tell. And for certain businesses, the data-driven attribution model may be a welcome change. You may also be able to opt-out of this change if needed.
If you have any questions about your own Google Ads account, please don’t hesitate to contact us.