Skill shifts naturally accompany the introduction of new technologies in the workplace. According to a 2018 McKinsey Global Institute discussion paper, changes resulting from automation and artificial intelligence (AI) are projected to accelerate between 2016 to 2030. Automation and machine learning technology adoption has and will continue to transform the workplace as people increasingly interact with ever-smarter machines. This includes novel ways of working, role requirement changes, and even the rise of jobs which did not previously exist.
As programmatic advertising is a literal result of marketing automation, it’s reasonable to assume roles and responsibilities for practitioners will evolve as well. Below are some predictions of what I believe will take place for those who work in programmatic media in the coming years, based on current and expected trends in both advertising and technology overall. This includes shifts in existing responsibilities as well as new responsibilities a practitioner will need to develop.
More algorithms, less time in the platform
As marketers grow more programmatically savvy and technology advances, algorithms (bidders) will shift from standard and black box, to brand (or even product) specific and custom-built. In fact, this shift has already begun. For example, Essence has programmed client-tailored machine learning technology into our media management platform, allowing us to automatically bid towards impressions we deem valuable across multiple variables simultaneously. We’ve seen these bespoke algorithms significantly outperform standard single variable auto-bidders against client objectives on the impression level. As algorithms grow more custom, they will become “smarter” relative to discovering impressions that matter for the deploying marketer. As this happens, campaigns will require fewer optimisations, thus reducing the time programmatic operators must spend in the buying platform.
Increased machine learning literacy
As algorithms transition from single and standard to many and custom, practitioners will require a baseline understanding of machine learning principles, including how models are built, trained and optimised. Just like how programmatic operators today select targeting levers (e.g. context, geography, time of day), operators of tomorrow may be tasked with selecting from a collection of custom-tailored algorithms optimised to differing signals depending on brand, product or objective.
While this provides potential for exponentially increased efficiency and efficacy in communications, great care must be exercised. As stated by Essence’s Co-Founder Andrew Shebbeare, the collision of AI and advertising is the most complex problem our industry has ever faced. While machine learning can drive great advances in advertising value, irresponsible deployment can lead to annoyance (e.g. irrelevance, filter bubbles) at best, and harm (e.g. brand damage, harmful ads) at worst. As such, it’s doubly important that those who deploy these algorithms are well-versed in not just their capabilities, but also aware of potential pitfalls.
Greater emphasis on strategic and tactical planning
As practitioners spend less time in the buying platform, time spent on standard campaign management tasks will see a requisite decrease. This frees up practitioners to spend more time on tactical and strategic planning tasks, including inventory sourcing, data strategy and custom partnerships. In addition, practitioners will have access to campaign signals that, once identified, can be built into predictive models for business outcomes. This will further enable marketers to derive data-driven insights and opportunities for growth during business and campaign cycle planning.
Based on this, I predict practitioners will increasingly interface upstream with strategists to inform what is (and is not) technologically feasible and favourable at the execution level. There are often times today when a communications approach calls for tactics that are technically impractical or even infeasible. Having folks traditionally on the campaign “floor” involved upstream could help close the gap between strategy and tactics, leading to less operational wastage, more seamless planning and execution, and improved business outcomes.
Deeper creative synergy
Programmatic dialogue typically revolves around data and inventory, and how the two are transacted between buy and sell-side parties. More often than not, creative remains an afterthought. However, studies have time and again demonstrated creative quality to be the single most important driver for advertising effectiveness, as illustrated in a 2017 Nielsen Catalina Solutions report. If data and inventory targeting are vehicles through which a brand’s message is delivered to an audience, then creative is the message itself. As such, the most accurate and precise targeting would be largely for naught without proper creative. It’s not hyperbole to say many in media are “majoring in the minors” by not giving creative its due.
That said, I believe the gap has and will continue to shrink as connective tissue between creative platforms and adjacent advertising technology components grow stronger or even fuse under a single umbrella. As a result, I predict targeting traditionally housed in buying platforms (e.g. geography, audience, context) will increasingly migrate into the creative build via platform integrations, dynamic feeds and/or APIs. This would in turn, necessitate deeper collaboration between programmatic practitioners and creative developers bringing us closer to data-driven campaigns that match creative and media across the entire messaging funnel. In this way, we are better placed to move beyond siloed creative messaging and closer to valuable creative conversations across both screens and channels.
Closing the online and offline divide
As traditionally offline channels become ever more digitised (and inevitably cognified via AI), boundaries separating online and offline will continue to blur. And it’s not just via offline channels becoming available for programmatic purchase (e.g. digital out-of-home, advanced TV), but also with digital and analog elements commingling in less explicit ways. For example, it’s not uncommon today for digital and offline elements to supplement one another during campaign planning, execution, and measurement processes.
Examples here at Essence include but are certainly not limited to programmatic out-of-home, connected TV and digitally-derived data informing offline planning and vice versa, online to offline in-store footfall analysis, and measuring offline behaviour for brand lift using synchronised digital measurement. This is just the tip of the iceberg for the immense opportunities we will continue to see as we bridge the digital and analog worlds.
Given the above, I believe programmatic practitioners in the future will require at minimum a foundational understanding of traditionally offline channels. This includes channel roles, planning and deployment processes, transaction methods and how each works in concert with traditionally digital channels. We could very well be entering a world where sequential and parallel messaging is tracked across both digital and offline media. Those planning and executing campaigns (whether they be digital folks who’ve learned offline, or offline folks who’ve learned digital) must possess the skills necessary to plan and execute accordingly.
Vincent Niou is senior programmatic and partnerships director for APAC at Essence, GroupM's data and measurement-driven media agency.