Digital marketing: what roles should be given to employees, agencies and robots?

Internalization has been a very popular trend in North America for the past few years. Large advertisers have fired their agencies in order to fill roles internally, from creation to social media management, even some media buys. In an era of transparency and performance, today’s agencies must demonstrate their value more than ever.

I’m not convinced, though, that we’re asking exactly the right questions before internalizing. Internalization in itself is not an end, and marketers’ decisions should be informed by the following questions:

•   What needs to be done internally with the brand’s teams?
•   How can the agency contribute from now on?
•   In what ways can a robot amplify efforts and results?

My friend Gaétan Namouric introduced me to the now classic phrase “never delegate understanding,” and it’s truer now than ever. I’ve seen several big brands try to internalize tasks related to implementation (for example, media buys), while leaving the strategy to their agencies. So, the grunt work is being done internally, while the strategic elements, the thinking part, is contracted out. If you ask me, it’s better to do the reverse, at least as a first step. Managing an implementation team can be very difficult when you haven’t mastered strategy. Strategy and implementation need to be perfectly aligned, especially in a world of continuous optimization.

What other tasks should be internalized? Those which require a high level of coordination with other internal teams and which must be carried out on a regular basis. For example, monitoring performance in collaboration with the teams that manage your other channels, deploying content strategy, and community management, the goal of which is often to tell the story of the brand, its offers and its attributes, all of which is subject to approvals from multiple teams, etc. These usually simple tasks become very cumbersome when they require you to bend over backwards to coordinate with an external partner.

Producing and deploying information in your PIM (Product Information Management), internal systems such as POS and consumer interfaces (catalogues, online sales, etc.) should also be managed internally.

Finally, the ability to manage your automation processes should ultimately sit with your internal team, in particular because it involves relationships with suppliers, and requires detailed knowledge of your products as well as the different customer profiles the brand comes into direct contact with. To achieve this, the existing relationship between marketing and IT needs to be a good one, which can sometimes be a challenge in itself.

There is still a lot of room in the market for consulting or managed service companies that offer cutting-edge, strategic and transparent services.

To fully understand the market and their position within it, brands need the perspective that only external partners can provide. Moreover, they need to bring in expertise on a variety of topics in order to perform on channels in a state of constant evolution. Consultants and external partners with the ability to innovate can help meet these growing needs.

The ability to transform remains a major challenge for brands, and modern partners are developing new ways to support them in their management of change. The adoption of a new technology or the internalization of a new process can be a success or a failure depending on the human factor and the ability to adapt. Between external execution and training, there are several formats that can help a brand integrate new ways of doing things. For example, I strongly believe in active consultation formats, such as regular workshops or co-piloting to develop the autonomy of the brands with which Adviso works.

An agency in 2020 must be able to adapt to this desire by brands to be more autonomous, and offer adapted solutions, formats and approaches. Brands, on the other hand, must accept and invest in hindsight, perspective, as well as in the autonomy of their teams and processes before they can claim to no longer need external assistance.

Many fear we are headed for an AI apocalypse; the idea that machines will one day replace us and steal our place in the workforce or, worse still, turn against humans and enslave us. This idea is greatly exaggerated; the truth is that machines have extremely different skills than humans.

For example, a machine can easily and quickly list extremely complex terms in a very specific order based on certain criteria or filters. It can also receive information from hundreds of sensors and make a risk-minimizing decision, such as the semi-autonomous cars available today, for example. On the other hand, a three-year-old child watching his sister go outside can answer the question “Why did your sister put on her shoes?” But a machine cannot. These so-called comprehension tasks are very difficult, even impossible, for machines because they involve unspoken or missing information. A machine can therefore make real-time connections within huge amounts of data but can completely miss an important notion that would be obvious for a young child.

In addition, scientists simply cannot program machines to replicate everyday activities like clearing the table or bringing the right items to the dishwasher, pantry or sink. The number of intuitive decisions that need to be made is simply too high.

The magic happens when a robot works alongside a human on tasks that are weighty, methodical, extensive and, shall we say, boring. Humans can technically perform quality control on a 20,000-line media classification, but we’ll never do it as well as a machine, and we will certainly never derive any pleasure or satisfaction from the job. And a human is highly unlikely to be able to carry it out every 5 minutes the way a machine can, with no breaks and without complaining. On the other hand, humans need to be the ones configuring the machine, supplying it with training data, making adjustments and asking questions, to enable the machine to refine its work and increase its value. Intuition and the ability to go beyond the available data is crucial in making marketing decisions.

It is this human-machine collaboration and its constant course-correction that creates lasting competitive advantages for a modern brand. There are many opportunities this presents, here are just a few: 

•   Using data feeds to automate advertising and indexing tasks that are constantly being refined and completed by a human
•   Automating auctions and announcements in a changing context (limited, perishable or constantly renewed inventory such as in the travel, tourism or event sectors, for example)
•   Automating quality control tasks, overseen by a human who will question, adapt and direct the process
•   Prioritization of service calls taken by humans based on texts received or words heard with natural language recognition and machine learning
•   Combining active data sources (POS, transactions, traffic, etc.) to identify insights and act on them, manually or automatically

The opportunity is not only in cost savings, but in agility and performance. Betting on the right user who will maximize the conversion in real time or offering the right product to the right customer using machine learning increases revenue and gives a significant competitive advantage to companies that master this human-machine collaboration.

For this collaboration to exist, it is essential to develop an agile working culture between programmers and advertisers. Future revenues will be found in the ability to use machine learning, natural language processing and robotics to leverage the data captured and generated every day. All this cannot be learned overnight.

The ability of organizations and brands to create and optimize collaboration between employees, partners/experts and robots will be an important dividing line between companies that fail, survive and excel.