From Mad Men to Math Men (m/f)
What every CMO needs to know to win with AI.
I have been in marketing roles at various companies and agencies for approximately 20 years now. Starting 2018 I’ve been upskilling myself in data science and artificial intelligence. My conclusion: these are great times to be a marketeer! Thanks to data science and advanced analytics we are finally moving from a mainly (subjective) gut-feel-based to a data-driven marketing profession.
This will definitely help every CMO to become a valued sparring partner in the board room, as AI driven marketing is proven to have a significant impact on both the top- and bottom-line. Recent McKinsey research shows that artificial Intelligence has the highest potential business value in Marketing & Sales. Microsoft expects that by 2025 95% of all customer interactions will be driven by AI.
But many companies still have a hard time turning AI initiatives into business gains. In my new role as Marketing AI Lead at DEUS my main objective is to help companies overcome the barriers that are keeping them from winning with AI.
Marketing AI technology is well developed
The main challenge is not in the technology or the models behind the AI. There is an enormous range of off-the-shelf apps and services available nowadays such as Clarifai, GumGum, LeadSpace, Ideas, Albert, Octane AI (and the list goes on and on and on…). Also, AutoML applications such as DataIku and DataRobot reduce the need for advanced coding skills when building your own bespoke models.
Inadequate amounts of the right (clean) data are also often mentioned as a barrier to start with AI. But given the number of pre-trained tools available for marketing the requirements in this area have also been drastically lowered.
From Mad Men to Math Men
The key to creating value with AI in marketing is understanding the business problem and translating it into a mathematical challenge. So, being successful requires significant marketing expertise first and enough data- and AI-expertise, second. In that order.
The key to creating value with AI in marketing is understanding the business problem and translating it into a mathematical challenge.
This will require senior marketeers with a skill-set that includes data science and AI. They should know how common modeling techniques work, what form of data they require and what kind of questions these models can (and cannot) answer. At the same time they should understand that AI is not a substitute for decision-making, rigorous hypothesis development and testing. But where do you find such a five-legged sheep?! Hard to find indeed. But they do exist, and they carry the humble title of ‘analytics translator’.
Analytics translators are invaluable for proving the financial contribution when you start piloting with AI in marketing. Adding one or more analytics translators to your team will help you gain traction with AI because they share a common language with both the business and the data scientists.
This type of five-legged sheep does exist. And they carry the humble title of ‘analytics translator’.
You can train your own marketing talent with proven strong analytical skills and/or background. But while this type of upskilling is taking place, marketers should also consider hiring an external analytics translator with proven marketing domain expertise, to support their initiatives and help create a marketing AI DNA. They can be deployed as the product owner of an AI pilot project, ideally working with a team that consists of at least a marketer, marketing analyst, developer and a UX designer. In case of a larger marketing AI program, analytics translators are well suited to also take the program manager role. They have the perfect skillset to align with senior business stakeholders and connect the business strategy and objectives with AI initiatives.
It’s important to realize that these external analytics translators are still hard to come by at the moment. Many companies are struggling to find the right talent. Data science and analysis will remain the most in-demand technical skills for the coming years.
Only a bias to action will get AI off the ground
Another important challenge while establishing a successful Marketing AI program at your company is to avoid ‘aiming too high’. Given the potential impact of AI, the temptation is to start with big transformation initiatives with matching game changing expectations. Such a project can potentially deliver exceptional value, but it will require exceptional effort, budget and time. Resources you probably don’t have available yet for your first steps into AI. So try to gain traction with Marketing AI with smaller ‘pilot’ use cases that can prove success in the short term but are not too small to be considered trivial by your stakeholders.
Running a marketing AI pilot project consists of 4 steps:
- Start with a granular look at the value drivers of your business (see figure 2.)
- Select the ones where AI can be used as a lever with relatively limited effort
- Screen for available ‘ready-made’ AI solutions or build your own prototype
- Test, learn, optimise and (when successful), scale
Once the value of Marketing AI has been proven with these pilot projects, it needs to be sold to the senior business leaders to get an OK and the necessary funding to start scaling. Most marketeers (still equipped with some of the old ‘Mad Men skills’ ) are naturally strong at storytelling so they should be well equipped to take business decision makers by the hand and communicate in the language and KPI’s that resonate best.
Final note
If you would like to further discuss how to upskill marketing professionals into analytics translators or my personal learning journey, please send me a message. I’m always happy to follow-up!
