Top 5 Macro Level Changes in Digital Advertising, 2022

In 2022, Meta released an updated playbook and Google rolled out a number of new products and initiatives. The digital advertising industry has new and evolving privacy standards, emerging channels and new campaign types. Of course, there are always new things to test and expand into but for the most part, these changes have been slowly marching out over the last ten years. Here we talk about these high-level macro changes and how to stay on the cutting edge.

1.     Rise of effective native automation.  By far the largest change stems from the improvement of native automated biddings.  This is the convergent trends of processing becoming cheaper, algorithms becoming more sophisticated, and the channel’s ever-growing cumulative data collection. These tools allow the bids and targeting to be made using criteria that are unavailable in manually bidding. It’s imperative to have the automation optimizing toward the conversions that matter by importing offline conversions.

There were once massive third-party bid technologies that consistently outperformed native automation but these companies have largely either disappeared, radically transformed their product, or linger in post-market fit obscurity. The vast majority of advertisers I speak to predominantly exclusively use bid automation but change targets based on performance, or backend metrics.

Action: Use bidding automation and make sure that you are importing backend conversions to make sure the algos have the data that matters.

If you are not not tracking offline events then the algos are blind to your true KPIs.

2.     Simplification and homogenization of account structures.  Because of this better ML programs are strongly incentivized to 1. aggregate data to allow the algos to have more data density and 2. provide more flexibility to the algorithms to make choices.  The channels also prefer less complicated structures to make advertising more accessible to the greatest amount of people.  Google is the largest perpetrator with removing match types, smart shopping, performance max and discovery display. Facebook’s number one recommendation is always to aggregate targeting.  One of Facebook’s strongest tactics to improve numbers is to make sure all ad sets have over 50 per week. 

Action: Make sure that your account structure has enough conversion data density. Too much segmentation feels like a natural way to advertise but in the vast majority of the time will harm performance.

3.   With less control there is a rise of non-incremental and low-quality conversions.  As we roll up our campaigns and move much of the bidding and targeting decisions to automation we lose the ability to control high vs. low incremental clicks and conversions, to visualize marginal vs. average performance and cleanly understand what specific targeting produced what result. Brand and Non-brand start to bleed together as well as Prospecting vs. retargeting. YouTube always includes a portion of what traditionally we would call view-through conversions.

More and more there is a demand for better insights in what levers matter because the major channels are actively reducing the levers and insight. The answer has to be an emphasis on on/off testing to measure halo and incrementality. Introducing variance in channel invested to be able to model different returns as you scale and really focusing on optimizing your media mix to maximize return.

Action: Even as you aggregate campaign structures you can still segment out targeting and audiences to test for their viability though this should be considered a test and not so much a way to drive cost-effective conversions. Check GA New vs. returning metrics to make sure that your campaigns are not pushing a ton of branded search or retargeting. Be aware of the percentage of view-throughs specifically for Meta.

4. Degradation of third-party cookie tracking and app ID sharing. Cookie tracking is now under attack necessitating a variety of new ways to measure success. Apps have been hurt the most and we now have a much more probabilistic model in how Apple attributes performance.  We need to rely on systems that are non-deterministic but that isn’t necessarily a bad thing.  We also have new tools in CAPI and old ones like offline conversion transfer.  These are essentially mapping qualities from click Ids to user information from purchase.  Lastly, we can borrow from traditional advertising and scientific method powerful measurement tools like geo-holdout tests to understand true value.

Action: We also need to rely much more on/off lift studies. Google and Facebook have native tools to run these tests but you can always set up a geo sequential A/B test to measure impact. From there you can understand the true value of the ads and optimize your media pan accordingly.

5.     Emerging channels and requirements of new media. Video and UGC are of particular note.  Tik Tok, CTV, Influencer platforms. Twitter. Google is investing in short-form media and Instagram has had the story ad place for years now. Here is the general hierarchy of channels with the big three being Google, Meta and LinkedIn but this list is changing all the time.

Action: Test the new channels and diversify your creative assets!

Big three dominate the industry but there are always challengers

So what is the takeaway?

Invest in making sure that you are maxing out CAPI/Enhanced/OCT and that your backend conversions are being ported into channels. We recommend using a combination of upper and lower funnel conversions with more weight on the lower funnel. Make sure that you are not over-segmenting your account and be conscious of audience sizes and data density. Maintain an exploratory budget meant for new channels and test these channels using an on/off methodology to account for incrementality/halo. Creative and messaging continue to be critical to any marketing success but there is more emphasis on short-form video and user-generated content.