Over the past eight years, working in Digital Marketing, I saw the industry obsessing over many different metrics when talking about performance campaigns: CPM, CPC, CTR, and then finally focusing on more meaningful metrics such as CPA, ROAS, and profitability.
The reason why DR online marketing is a no-brainer nowadays is that, after years of debating on which metrics were meaningful, the right tools came along to take the guesswork and the heavy-lifting away from the people buying the media, and give it to machine-learning algorithms. Products like AdWords can automatically manage bids and placements to hit the desired cost per action, return on ad spend, or profitability.
This level of automation has come with incredible advantages to the industry as a whole. What used to be a very manual, slow, and tedious process, is now done by machines in a fraction of a second. Don’t get me wrong, true DR optimization is a very involved process, and that is why it is so hard to find real talent in this space: in my experience, I’ve seen only a handful of people doing it right and every expert you talk to has a slightly different way to go about it. Also, even when we reduce the complexity from the media buying process, there is still much to be done to optimize the creative and the conversion funnel. When deciding who owns the landing page or the creative development process, the debate often becomes political rather than tactical.
All this is possible in the Direct Response world because the relevant metrics are clear, the industry has reached consensus on what matters and machines can be leveraged to optimize the buying process.
Branding is not there. The industry talks about viewability, and there is not even agreement on what constitutes a viewable impression (e.g. for videos it could be a muted play for 3 seconds, or a full in view play with audio and video). But that’s just scratching the surface: if we were to draw a parallel for the DR world, it would be like asking that each ad should be clickable. It is not a measure of success, but the minimum requirement for even having a conversation.
I get it; it’s a vast and complex problem: publishers have been building their revenue models on different metrics, and asking them to switch suddenly is not sustainable for the industry. The work that many are doing in the space of “native advertising” is exciting, although there are still many technical and moral gray areas (e.g. advertorial).
Testing for the sake of understanding what matters has high opportunity costs, and you can run for weeks or months only to disprove a hypothesis. I recently concluded a 3-month-long test that was proven inconclusive, and it wasn’t an easy conversation to have with my stakeholders. But getting the right brand metrics in the world of programmatic buying is paramount to unlock the real power of digital advertising. Through testing, we may prove that time-in-view is the metric that influences brand-lift in a banner campaign: if so, you may want to optimize your programmatic buys towards sites where content is consumed by slowly scrolling through the page, or pay more for banners appearing next to a video player. At that point frequency, rather than reach, will become key: advertisers will pay more to serve the third or fourth impression to the same user, rather than serving the first impression to a new user.
The industry will be finally able to move forward only after identifying the right metrics: the DSPs can fine-tune their products to optimize towards brand-lift in the same way they now optimize towards profitability, the advertisers can confidently run digital brand campaigns programmatically, and publishers can get paid appropriately for quality inventory. In this context, the industry will also have more tools to combat fraud.
Unfortunately, as I’m writing this article, we are still in an era where tools for conducting programmatic brand testing are only available to the big-budget advertisers and agencies: major DSPs still require ad verification technologies to be piggy-backed into their tags that are then trafficked through an ad-serving platform into different ad exchanges that are too siloed to optimize for ideal frequency. And since we are in the wild West of testing for the sake of understanding what matters, advertisers still need to rely on the strong analytic team, often provided by an agency that may be using the advertiser’s data to build their own technology to then acquire new business.
In conclusion, I believe that only by removing the guesswork from programmatic buying we will be able to rely safely on online advertising for brand marketing. The DR-world has taught us that the industry first needs to reach a consensus, and then the technology providers need to build solutions that remove the human element and commodify the buying process. With this outlook, the future becomes very exciting since the same principles will be applied to the emerging sectors of Programmatic TV and Digital Out of Home. In an on-demand world where screens multiply and a single identity becomes the real currency, the next challenge will be providing the necessary tools to generate value to the online advertising industry, while still respecting the user’s privacy.