What is marketing? (Rory Sutherland)

I love this philosophy on marketing from Rory Sutherland. Rory is the Vice Chairman of Ogilvy & Mather Group, and I heard this on Invest Like The Best (Source).


Begin quote:

First of all, don’t think that marketing is marcoms. Because is very many businesses the two are conflated to a point where marketing is sometimes dismissed as “the coloring-in department”.

Marketing should basically be the repository for any psychological insights, particularly those which are counterintuitive, or which run against normal economic assumptions in the space of human perception and behavior. That would apply also, by the way, to internal relations, shareholder relations, and obviously customer relations.

So marketing, in a sense needs to step back up to the plate and not allow itself to be trapped in the marcoms/communication “ghetto”. The second thing is, by being trapped in that ghetto, [marketing] tends to get framed as a cost; as a necessary evil, and not a source of value creation. Once you’re in that marcoms ghetto, you’re essentially a cost center. As a result, what you get judged by is the efficiency with which you perform the necessary. This turns the whole thing into this sort of marcoms insanity, where everybody is trying to target 2% more efficiently every month.

And I would argue that a large part of marketing needs to be both experimental and probabilistic. We shouldn’t turn it into this efficiency/optimization game.

…Maybe, actually you should do very inefficient mass advertising, simply because it increases your odds of getting lucky. You can’t predict how in advance, and you can’t attribute the success in retrospect, but yet nonetheless, fame is valuable, on balance. More lucky things happen to famous people simply because more people have heard of them.

…that’s the first thing [for new clients]: don’t turn into an efficiency/optimization game. Accept that it’s probabilistic. And accept that there need’s to be a trade-off between exploit and explore.

End quote.


Why I’m excited about data science

What’s all the rage with data science? And what exactly is data science? In this post, I can’t promise to answer either of those questions in a way that will be satisfactory to you… ha! What I will share is what excites me about the field’s study and why I think its applications will become an increasingly important part of my work from here on out.

Data science: it’s a thing.

As you can see in the above trend, the term “data science” as a topic has exploded in interest over the past decade. In an oft-quoted 2012 HBR article by academic Thomas H. Davenport and DJ Patil, the former Chief Data Scientist of the United States Office of Science and Technology Policy, the role of the data scientist will be the “sexiest job of the 21st century“.

Data science as the ego-killer

On a personal level, and regardless of my own data science skills (at present I’d consider myself an amateur and aspiring promoter of the discipline at best), I’m excited about data science because I believe that it is far more than just a “sexy job”. I believe that data science is becoming an increasingly pervasive force that is shaping the way we as individuals and groups make sense of the world around us. I’ve witnessed this force begin to emerge in business and capital markets, and I hope that it will also positively impact governments and societies on a much larger scale in the future.

I see data science as the pursuit of truth; its practice being to seek out and pay attention to the voices of empirical evidence — not those voices which scream loudest. To leverage data science is to seek influence, not power. Data science is changing the way we persuade, and are persuaded by others. I believe that “data-driven decision making” should be as much of an educational requirement as algebra is today, perhaps more so (reference: Annie Duke’s Alliance for Decision Education!).

Ben Jaffe, host of the Linear Digressions podcast wonderfully articulates a worldview and approach that relies on data science, explaining, “…we [all] have cognitive biases… logical weaknesses… pride… and that’s why we have things like the scientific method and statistical rigor: to compensate for our deficiencies as humans. And one of those deficiencies is that we see ourselves as neutral, even though just like the algorithms we build, we reflect the world that we are steeped in.” (Note: I expanded Jaffe’s quote beyond its original context, which was about proactively lessening racial discrimination through the use of data.)

Data informs how (not what) you build

Please don’t get me wrong. I believe that data and experimentation support vision, but do not replace vision. Data science itself requires creativity, but does not replace creativity. And specific to business, I’d argue that qualitative based vision is absolutely a requirement of any successful venture, but that there are many ways to build towards vision. And that holds true especially when an entrepreneur is working to achieve something that hasn’t been done before.

I started my career in construction management — working on teams that oversaw complex multi-million dollar projects, where architects, designers and their clients dreamed up details of how a physical space should look, feel, and be built. We’d be handed huge stacks of architectural drawings, specification binders, and material samples dictating nearly every detail of a project. At times it felt like a real life, more challenging version of building a Lego set (with the primary challenge being the ever complex variable of people management).

But outside of construction, manufacturing, and a select few other fields, life and progress do not have the luxury of moving forward with so many knowns. In company-building, for example, the chief executive may have a very clear vision of what they want to accomplish (in terms of business milestones, for example), and perhaps of certain aspects of which path they want to take (in terms of organizational structure and culture, for example), but will not have detailed plans and specs on how to go about building towards that vision.

I think this experience of transitioning from construction management (an environment with instructions) into technology startups (an environment without instructions), has significantly contributed towards my attraction to data. Not knowing what to do or what to build can be disorienting, and as organizations continue to move away from top-down hierarchies, it’s becoming increasingly important to find better ways to lead, manage and build. Whether in the discipline of online customer acquisition, my specialty, or some other aspect of company-building, learning to ask smart questions and looking for answers in the Data provides the closest thing to instructions.

What’s the difference between data science and other “data stuff”?

Over the past six years as a growth marketing practitioner, I’ve definitely come a long way in understanding how to use data in my every day work, in campaign and funnel analysis, shaping marketing experiments, crafting data-infused educational content, and in making discoveries of data patterns that help inform strategy. Has any of that been data science?

In the Davenport and Patil article I mentioned above (the one about data science being sexy), here’s how those guys define the real stuff: “More than anything, what data scientists do is make discoveries while swimming in data. It’s their preferred method of navigating the world around them. At ease in the digital realm, they are able to bring structure to large quantities of formless data and make analysis possible. They identify rich data sources, join them with other, potentially incomplete data sources, and clean the resulting set. In a competitive landscape where challenges keep changing and data never stop flowing, data scientists help decision makers shift from ad hoc analysis to an ongoing conversation with data.” (Emphasis my own)

When referencing that definition of data science and asking myself if my data-related work in the past as a marketing professional has been “real” data science, I ask myself, was my fourth grade science fair project “real” science? Hell yes. I’m proud of what I accomplished in fourth grade science. So too am I proud of the “data science light” work that I’ve done over the past several years, even if the work led to relatively minor value-adding insights. That said, from here on out I expect to that I will require far more intention on a personal level, surrounding myself with a high caliber of talented people well-versed in data science, increasing my fluency in various aspects of the field, in order to deliver insights of substance and significant value. And to me, progressing along this path and knowing that I’ll be able to add more value is exciting.

Let’s get philosophical (again)

Back to my point about data science as the pursuit of truth… I believe that to succeed in data science in any capacity, even if (or especially if!) it’s just working in an organization that pays attention to its data, a deep level of self awareness and humility is required. When seeking out answers to a question, so many of us think we know, want to be one that knows, and convince ourselves of some supposed truth. These motives (be it for money, status, or other) and unconscious biases lead to a place of overconfidence. My definition of overconfidence is any degree of confidence that comes from thinking we know something that we don’t, and that can be a drug. It feels good, and it might win material reward and popularity contests, but it can also lead to systematic failure. Daniel Kahneman discusses this at length in his Nobel-prize winning work, Thinking, Fast and Slow. While this is not a self-help book, an implicit takeaway from it might be that by training ourselves to be more analytical and deliberate in judgement, we can make better decisions and lead happier lives. The study and application of data science, I believe, lies at the very core of this training.

I think it’s also important to avoid analysis paralysis. Data science isn’t for every situation, and an important part of what can make the applied science great is knowing where to apply it. Additionally, there’s a cost to collecting more data and running analyses— oftentimes it’s better to make a decision with inaccurate data, partial information, or no data at all. But even then, wouldn’t it be awesome to have the ability to know when to walk away from the data, or the potential answers the data might provide? How many of us are confident in saying “I don’t know; let’s look at the data”? How many of us build confidence through inquisitiveness; through not knowing an answer? As someone who’s spent a lot of time drumming and studying the drumming greats as a teenager, there’s an analogy I can’t resist here, which is that the best drummers know how to play softly and when to leave empty space.

The next big thing: doing data science in your underpants

I don’t believe that data science will ever provide all the answers we seek, but I predict that in the coming years it will help us unlock puzzles and challenges in mind-blowing ways that are currently unimaginable. I believe that data science will help clarify realities that may have been previously concealed. And in terms of data science being a gold rush of sorts, I think that it will actually be part of a new software revolution, in which companies won’t have to adapt their problem-solving needs to a particular program, but rather the program will adapt to the needs of the company. I believe that data-driven companies will far outperform their non data-driven counterparts.

Jose Quesada, founder and CEO of Data Science Retreat, on the SDS podcast, talks about why data science is going to be the next big thing: “The next [technology] wave… machine learning, makes websites and apps look ridiculously underpowered by comparison. We can build machines that can see the world. That can identify objects. That can understand your language as you interact with things like Alexa. This amount power is actually in the hands of anybody with a computer. So you can literally be at your kitchen table with a laptop, running a deep-learning model detecting something that is needed for you to solve a problem, because of open source libraries and pretrained models. You can literally be in your underpants or at your kitchen table, solving really important problems that were impossible to solve just five years ago.”

Here here to making data more useful, and to the pursuit of our refined individual and collective character through the practice of data science! As enthused as I am to publish this post and tell the world why I’m excited about data science, I’m even more excited to actually do the work. Surely I was overconfident about one or more things in this piece — something which I thought I knew but will soon be kicking myself for being too sure of. I guess that’s what subsequent posts are for.