Google has never been a company to shy away from setting the pace of change. They tend to do this in two distinct ways, either by force to expedite an inevitable outcome or by enabling easier access emerging technologies that would otherwise be difficult to realise (for both consumers and businesses). The introduction of ‘Enhanced’ campaigns from 2014 remains a perfect example of the former, pushing businesses to rightfully accept the impending smartphone takeover, even if no one at the time was ready or convinced by it. Initiatives and services like ‘Feed Rules’ and ‘Campaign Translations’ are a great example of the latter, unlocking more opportunities for businesses to grow reach, which in turn keeps their own ad revenues steadily growing over time.
Google has taken this approach since the very beginning, and although the underlying motivations remain the same, there is a distinct theme driving the ‘automation’ elements of products and features it has rolled out over the past 12-18 months. Some will say (including Google themselves) that the drive to automation is nothing new, and in the general sense they are right but look closer and you’ll see a very particular difference. You’ll see a business that is not only developing but embracing machine learning and AI faster and more openly than any other business on the planet. This results in a very significant difference between the automation of yesteryear and the inevitable destination we are headed towards, the era of automation through AI.
AI everything
You only need to look at their recent hardware product launches to see how machine learning AI is being weaved in to be the differentiating centrepiece of all their services. Take Google ‘Clips’ for example, a piece of camera hardware that’s named like a piece of software, powered by AI, it is able to ‘recognize expressions, lighting and framing and capture beautiful, spontaneous images’. The physical camera continuously records footage and is able to understand what photos and videos would be desirable for the user, enabling a completely new (albeit slightly creepy) way of capturing candid special moments with your loved ones.
Of course, a consumer-facing video/photography product may seem quite irrelevant to online marketing, but it’s a perfect example of Google’s confidence to add value to a relatively stagnant product segment through its’ machine learning wizardry. Their approach has been no different to their ad business over the last 12 months, initiatives centred around machine learning going well beyond typical automated bidding strategies they have been pushing for many years.
It’s not perfect, but it’s here so deal with it
App Promotion – Pre and post manual mobile campaign targets
This may seem like a small change in performance but not an insignificant one. Others may have seen a good return on their investments relative to manual targeting approaches, if managed correctly however, UACs most definitely fall behind on the performance scale. The cynic in me says the expedited transition to UAC only app campaigns is Google’s play to push remnant mobile display inventory that would otherwise go unused, hidden under the umbrella of machine learning AI (creating an incremental stream of ad revenue growth). This is reinforced by the lack of detail available in any reports when running UACs, the conscious lack of transparency adding to the conspiracy. Detailed performance data would still be invaluable to advertisers, allowing them to understand what is working well and why, which in turn can help inform marketing strategy across other channels, even if they are unable to use it to directly influence UAC performance itself.
That being said, some of Google’s initiatives to increase advertiser reach aren’t necessarily performing poorly, even if advertisers are oblivious to them. The accuracy of keywords matching to search terms that you’ve not asked for, regardless of match types, functions surprisingly well. In case it isn’t clear, match types and how they are understood to work are dead. Much like an ageing night watchman armed with a fading flashlight, the practical use of them is diminishing by the day; they provide a false sense of control and security for those who normally live and die by meticulous account structures.
Given the direction of travel, it’s not difficult to imagine in the next 18-24 months eCPC as being the only manual bidding option, search text ads only being available via DSAs or shopping feed based campaigns being run like UACs, ‘campaign management’ in turn being limited to setting monthly budgets and defining campaign objectives. This phenomenon of automated ‘machine learning’ campaign execution (whilst removing existing control) will eventually come to other publishers, even if they are poorly executed in comparison to Google’s efforts, other media players will follow.
So where does this leave the value of digital marketers operating in markets dominated by the publishers managing large parts of campaign execution? What value can agencies add? Digital marketing agencies will need to adapt their proposition to either focus on elements that clients will always find difficult to manage in-house, due to technological and business constraints, or extend the value-add elements to their service that clients have always craved more of. There will need to be a focus on digital strategy and consultation with a view to guide brands.
Data management and broader audience planning along with extracting additional value and insights from existing first and third party datasets are some of the obvious ways in which this can be done. This along with distilling actionable insights from activity being run that has application beyond channel silos is what should be the overarching goal. Many agencies will struggle with this transition, where the norm has become to repackage off the shelf tech and follow a ‘me too’ approach to execution. How well they differentiate themselves with the impending changing landscape will dictate if they keep their seats at the proverbial tables.