5 thoughts going into 2023
NB: These opinions are my personal ones & do not reflect my employer
As usual, December is full of predictions of what will be the state of digital advertising after another year of acquisitions, lawsuits & technology evolution. This year, I’m going to double down on a few areas that I have got my interest.
Don’t Underestimate Meta
Meta by all accounts has had a turbulent 2022. From the various fines & impending lawsuits from Cambridge Analytica to more recent changes, they have also had to provide an advertising product that could evolve from the hard hitting changes of iOS14.5’s ATT.
And this has for sure implicated how much media budgets particularly for performance campaigns is flowing through Meta. TikTok in particular has been noisy in evolving its ads product, taking marketshare of this performance budget alongside the easy option of just putting it into Paid Search.
Don’t get me wrong, Meta have bigger priorities than attempting to fix its approach to advertising & things like the Metaverse are just buzzwords which will not mean much for the standard advertiser for a little while yet. The latest EDPB ruling has the potential to completely redefine how European ad campaigns work as another large update.
However I am fairly convinced that there is no other company in the social space & arguably beyond, that is trying to innovate for what the future state of digital advertising looks likes. This is both from a technical perspective from some of the server side capabilities on Meta but also to areas like measurement. It is clear that conversion modeling in tandem with automation (Advantage+) has given Meta a strong Q4. But the future with PET (Privacy Enhancing Technologies) & products planned is far ahead of its competition. As a result, I don’t expect Meta’s role on Paid Social to change at all but it will come down to data regulation as to what the final impact may end up being.
CLeanrooms get dirty
A buzzword that no doubt is key to the evolving privacy first landscape is data clean rooms, but at this point, the definition is getting further & further from what it originally set out to be. As Eric Seufert is famously quoted as saying “Everything is an Ad Network” (very true by the way), it seems like “Everything is a Data Cleanroom” nowadays.
It is therefore important to read between the lines, understand the noise and position your use cases as to where a DCR can provide value. Which for the majority of cases, there is an use case but is a struggle to put the whole thing together.
I expect more vendors to say they have clean room capabilities but in reality they don’t & are a glorified DMP. The walled garden solutions will have to evolve, as we have already seen with what Google has done to ADH and Amazon’s intend to make AWS a viable host for anyone to build a cleanroom on. The use of a DCR for a small to medium size business is a tricky model yet to really be solved but at this moment, it is imperative to be getting something into your tech stack.
SERVER SIDE CAUTION
Server side tagging has become compulsory to have in place, as we enter an ecosystem with less signals & the likely demise of legacy advertising identifiers like the 3rd party cookie / mobile advertising id. If you have not invested in this yet, here is your wake up call.
However as mentioned previously, its not all smooth sailing when it comes to the legal implications & the general more technical approach required to implement this. This is something that should be done properly and not taking the easy route which may not necessarily futureproof your business.
But what I want to concentrate on is what server side tagging cannot solve for & in reality, nothing outside of machine learning / modeling can (which is open to interpretation). That is, the so-called anonymous user. This user who you have no information to go off as they are not a known customer or logged in, remains out of touch with a server side approach. Yes for those more technically enabled teams, you may try do some fancy workarounds with first party cookies / local storage but in reality, this is not going to viable in the long run.
And this is what in particular the 3rd party cookie was so good at, taking the anonymous user and targeting / measuring them cross-site. Without this, there are key implications on retargeting & measurement. I mean lets be honest, if you are spending paid media budgets on retargeting at scale believing that this is the best performance strategy, please seek help. Server side tagging has many other benefits which are useful but it is always worth remembering that this is not a direct replacement for what we are used to in a privacy centric world.
UNIversal IDentifiers have a ceiling
The universal identifier is a commonly mentioned solution to the future state of a 3rd party cookie-less ecosystem on the open web. Every year, we see a lot more maturity in this space as well as newer attempts to unify amongst the demand & supply side.
The premise is simple, replace that unethical cross-site legacy approach with a more compliant / controlled alternative. This often sits behind some form of hashed PII (email being popular), a SSO approach or effectively just piecing together other data points.
But ultimately as the landscape gets more complex, more walled gardens appears & fragmented media plans are bought against, there are going to be open questions over addressability as well as scale. It is already well underway with a probabilistic approach starting to show to increase coverage beyond the known deterministic base. And then there is the spanner of data regulation, particularly in Europe, alongside the intention to cutdown on workarounds that leverage fingerprinting. Make no mistake, certain vendors are doing it ethically and others more questionable.
This solution no doubt will gather pace but given there is still yet to be a truly universal / global solution that is happy on all parties, this is more like a cog within an overarching machine that spans many universes. UID 2.0 is in a very tricky place for example with nobody wanting to take the pressure job of ultimately owning it. Don’t put all your eggs in 1 basket here.
Customise your automation
2022 saw more automation in campaign build than ever before, as Google (Performance Max), Meta (Advantage+) & even TikTok (Smart Performance) looked to build broad targeting with machine learning powering buying so that all advertisers need to do is add creatives + a budget.
To some extent, their hands have been tied particularly in the mobile app space where the reality of something like the SKAdNetwork means that this is really the only way to get enough learnings for campaigns to optimse. We may see iOS shift with SKAN4.0 next year but in reality the hands off approach is here to stay.
And this opens up more questions than answers. Why cannot you do certain things in these campaigns like exclusions? Why can you not report on a granular level if you are opted in to targeting everything? Are these blackboxes actually a good thing. This has definitely been in the press recently for Google & Meta.
I do think there is a middle ground however, which I’m sure is not popular amongst vendor reps. It makes sense why advertisers are sold this route, based on the ML behind certain vertical tagging / event matching within these platforms. Think about it: if every ecommerce business implemented a “Purchase” event on their website / app consistently, then the ML these media platforms have is much more powerful & accurate.
But there is no USP nor custom optimisation / measurement at a point. So there is a viable need to be smart about this, whether that is being more technical with API led approaches for bidding / creative or even working with 3rd party tech. At the end of the day, you want to optimise to what you care about, which in some cases is not achievable in these campaign types.
FINAL THOUGHTS
Albeit these are my areas of interest but there is so much more going on in the digital ecosystem to which this only scratches the surface. The setting for the next few years will completely shift how advertisers plan, buy & measure digital as well as exactly how tech stacks are built.
It is safe to say that this is the most difficult but exciting time for digital advertising. But what nobody should be doing is waiting for the answer to come to them. Expanding your knowledge and understanding what is truly at risk / not possible in this future state is key to addressing how to solve it for your own business. Data regulation may have the final launch with Google / Meta / Amazon / Apple but these 4 continue to hold the weight for digital no matter what happens.