One of the most important lessons to keep in mind when it comes to the digital marketing ecosystem is that “no platform or publisher is a charity”. They are, at the end of the day, businesses here to make money, just like the rest of us! I’m not saying this to disparage these companies; after all, any B2B or service-based business has their foundations in mutual beneficiality. The more money we make through them, the more we’ll spend with them, and the more money they make. Win-win. Except as time goes on, and growth pressures mount platform-side, we know that they’ll happily rest their hand on the scale to tip the balance a little bit more in their favour. This is business, not charity.
I’d like to put on my Macquarie Dictionary hat and coin a term: the Janus Algorithm. As the true nerds out there will know, Janus is the two-faced Roman deity, and this duplicity is inherent in the platform algorithms (I couldn't name it after Gemini for obvious reasons). The premise of the Janus algorithm is thus: The optimisation algorithms that get you the “highest” return per dollar spent also must maximise the return for the parent platform. This means you will necessarily have to sacrifice some proportion of ROI for the platform to profit, but this must be balanced to ensure you are profitable as well.
Seems pretty obvious when it’s typed out, right? When you extrapolate into the human realm, the job of a marketer/agency is to tip the scales in favour of the account, while the job of the publisher rep is to tip the scales in favour of the platform.
Given this, there’s an alarming lack of credulity in the industry. People at all levels will happily take a publisher or platform’s word as gospel without applying a lens of skepticism or even realism. I spoke to a friend recently, who was struggling with some of her senior staff leaving low-performing ads running when higher-performing ads were available but spending less. For sake of example, let’s say these are the two ads in the ad set:
- Ad A: $2000, 5 leads, $400 Cost per lead
- Ad B: $1000, 4 leads, $250 Cost per lead
The account manager said that they’d spoken to the platform rep, and that Ad A needed to be left live because “it was playing an integral part in storytelling and sequencing, which is why the platform shows Ad A so much more!”
Sorry, but that’s bullshit.
If a platform cannot produce a verifiable path analysis that proves such a sequence exists, I’d bet my house on the fact that the platform will show as many poor ads as possible while it can still hit the target CPL. Or, to put it another way;
- Ad A: 2% chance of conversion post-click
- Ad B: 4% chance of conversion post-click
The Janus Algorithm is going to show Ad A significantly more than Ad B; in fact, the platform will show Ad A as much as it can get away with without raising the average CPL beyond the acceptable limit. Will pausing Ad A get you a better CPL? Most probably. Will the platform make less money if you pause Ad A? Almost definitely. So what do you think the recommendation from the platform is going to be?
This is only going to become more and more prevalent as we move towards the black-box, AI-generated, 20-creatives-per-ad-set “best practice”. The more obfuscation within a platform, the less a user can delineate between the helpful and harmful arms of automated optimisation, the more effective the Janus Algorithm will become.
This isn’t “wastage”; this is the optimisation algorithm doing exactly what it is designed to do. As I said before, these companies are not charities and we shouldn’t expect them to act like one. In the case of Meta and TikTok, their revenue is based almost entirely on ads - they need to milk every account for whatever they can take. In the case of Google, we’ve already seen that they’ll happily tweak the algorithm for returns rather than efficiency. This isn’t direct criticism of these platforms, because I’d never expect any business to sacrifice revenue that’s on the table - this is more of a critique on my colleagues in the industry who parrot every talking point and follow any automation down the rabbit hole without regard to how it actually affects their clients or company.
The inner workings of platforms are only going to become more opaque as time goes on. Again, I must labour the point that this isn’t out of pure malice - black box automation is actually a very good method of scaling, particularly when the input data is strong! However, the black box also provides much more opportunity for the Janus Algorithm to sway the balance towards the publisher, and it’s our job as marketers to battle against it for the sake of our company or clients.
Marketers have been scared of AI “taking their jobs” for a long time, but the reality is that we’re safe, because only humans can recognise and account for these types of situations. There are manual interventions required to get the best results for your client: everything from pausing creative, using left-field messaging, opting out of placements or networks, omitting oneself from unnecessary auctions, etc. Do you truly think one day you’ll trust a platform enough to give it carte blanche to spend on whatever it wants with no oversight?
It’s clear what platforms really want. To completely misquote the great comedian Steve Hughes: “Don’t spend your time targeting a precise audience, making an effort to remove bad creative from the pool, and developing a strong messaging rapport with your key audience. Just… give all your budget, signals, and creative to us and uh… we’ll make sure they see them.” Marketers are here to hold them accountable, and we’re not going anywhere.
Luke is a Co-Founder and Head of Growth at Hard Refresh. He's been wrangling algorithms for over a decade now. You can connect with him on LinkedIn here and see if he ever says anything nice about tech publishers. Hero image credit: Hassan Sherif.