Last week, Benn Stancil wrote a post asking a basic question — “Do data driven companies win?”. Within, he gave 5 fictitious fashion companies that looked the same operationally but had different hypothetical competitive advantages, for example, having decades of deep knowledge working in the industry, or using data to identify and respond to the market. Ultimately, Benn goes on to ask whether “being data-driven” is actually a competitive advantage — then we should be more willing to invest in a company employing that advantage over other advantages.
It’s a sharp question because the general consensus of the business zeitgeist is that being “data-driven” (whatever that actually means in practice) is not only a competitive advantage, but something that existing “not data driven” companies need to get on board with or otherwise perish. Not to mention all of our collective data practitioner careers depends on this near-universal belief to fuel continued salary growth. This backdrop of near-universal acceptance really does warrant questioning what are the boundaries and limits of its applicability.
What are the most obvious “easy” wins for data?
Perhaps the most obvious place where data is put to use is in “optimization”. Before the term “data science” was ever uttered, entire fields for accounting and scientific/operations management had been developed to use numbers and data to help make companies better decisions. For example, accounting, the literal giving an accurate account of where money in a company is coming from and going to, has been used for centuries to make sure resources aren’t stolen or wasted. There are tons of important decisions that are made thanks to the information that accounting practices generate.
While data scientists are rarely asked to encroach on the territory that accounting and finance are doing their work, we very often are asked to give accurate an accounting of where users are coming from, where are they going, and what are the users doing. This information is ostensibly used to help people make better decisions on how to design the pages that users are touching to optimize some metric or other.
Similarly, people can employ basic data science techniques to collect data and make decisions on how to best allocate their ad spend, or see if a new version of a web site is better than the old one. Tons of data scientists, myself included, do this sort of work in their day-to-day activities.
I’m sure that there are many places that would love to have even this level of insight into their data and don’t have it for whatever reason. There are going to be plenty of places who make use of such insight in their daily work and call themselves “data-driven”.
But I wonder to what extent is this form of data usage a competitive advantage? Because this level of data work is relatively accessible with even CSV files and spreadsheets, more and more competitors will be able to adopt it and having it stops being an advantage, while not having it becomes a disadvantage. Like with accounting, I’m sure only a very few places can say “our way of doing accounting let’s us earn more money and make better decisions than our competitors”.
Perhaps all the fearmongering of “become data driven or DIE!!!” is because people see the landscape shifting to everyone doing these basic level data tasks, and those left behind will be at a constant disadvantage.
Value beyond “accounting” quickly leaves the “data-driven” space
If we leave behind the “simple stuff”, where else can data generate value?
Oddly enough, we also start leaving the realm that many people would call “data driven”. More advanced uses of data typically involves using data as a product feature of some sort. For example, any recommendation system uses copious amounts of data to create a recommendation “product” that is used to grow the business.
Productized data like recommendation systems and ad platforms are definitely valuable things that count as competitive advantages even if they don’t drive decision making. Massively profitable empires like the ad systems that underpin the majority of the revenue for Google or Meta are also what helps those companies fend off competitors. Somehow we’ve stopped talking about what data activities are helping drive decisions, and instead talking about using massive troves of data to squeeze more ad spend out of the internet. They’re products that need to be designed, tested, and constantly tuned themselves.
About the only category I can think of that fits a “decision driving advanced data process” is using ML techniques to scale out filtering and decision making. For example. YouTube’s ability to detect porn and stop it from getting on the site allows them to handle the hundreds of hours of video that gets uploaded to it every second of the day. It’s almost assuredly unprofitable to hire enough human reviewers for such a monumental task, so the filter acts as a competitive advantage. But even then, the filter doesn’t guild company policy or actions, but instead reacts to changing conditions of the world. At best it’s a tenuous link to “decision driven”.
Data provides visibility, but not vision
So, to circle back to the original question of what advantages that being data driven confer relative to other competitive advantages… the answer seems to be that data that isn’t turned into a product that can be used or sold is best suited to being a force multiplier.
A smart team can take data and combine it with their other strengths to make good interpretations of the data and make decisions that work out “better than chance”. It increases the team’s hit rate and lowers their risk of making decisions. What’s important here is that the team has the domain knowledge and experience to to make decisions under uncertainty, using data to help their odds. Data here acts as a windshield, providing the visibility to how the business is working and changing. The vision, the guiding principles as to where to go and what to pursue or not does not come from the team and won’t come from the data itself.
For a team that doesn’t have the vision and domain expertise to take calculated risks based on data, data can act as a negative force multiplier. It’s like designing by committee or creating TV shows based strictly on well-worn formulae. Bad things will eventually happen because there’s no additional vision to guide things along.
This is where I think many can get confused about being data driven. Data provides a critical reference point to start making better decisions with. It does NOT mean ceding the important decision making authority to the data itself. Imagine how quickly a company would fall apart if people merely followed the results of whatever data analysis happens to come out every day. It would be like keeping the same homepage design for 5 years because every little aspect of the page has been optimized in that time — no fresh redesign can ever hope to match the performance. The data all leads to a local maxima. The only way to get out is to have someone dare to look beyond the local data and take a bet that doing things differently.
That’s why of the five example startups in Benn’s post, the data-driven example looks so much less appealing than a startup that has access to other competitive advantages. A team that has stronger internal vision and talent can make use of data much more effectively than clueless folk who “will go where the data takes them” without some opinions of their own. We all know that different people can analyze the same data and come to conflicting conclusions, so the bigger determining factor to success would necessarily lie outside of the data — within the decisionmakers themselves.
This is why, for many years now, I’d think about starting a business but quickly decide against it. As a consummate data nerd, I know that my skills are best used to optimize processes and make existing things work better. Creating things wholesale from scratch is a different skill set. But on top of that, I also lack the optimistic vision to believe in something to the point of being willing to bet a bunch of time and money on an idea.
Maybe someday I’ll find that conviction somewhere, but until then I will continue to find opportunities to make people around me smarter.
Standing offer: If you created something and would like me to review or share it w/ the data community — my mailbox and Twitter DMs are open.
About this newsletter
I’m Randy Au, Quantitative UX researcher, former data analyst, and general-purpose data and tech nerd. Counting Stuff is a weekly newsletter about the less-than-sexy aspects of data science, UX research and tech. With excursions into other fun topics.
Curated archive of evergreen posts can be found at randyau.com.
Join the Approaching Significance Discord, where data folk hang out and can talk a bit about data, and a bit about everything else.
All photos/drawings used are taken/created by Randy unless otherwise noted.
Supporting this newsletter:
This newsletter is free, share it with your friends without guilt! But if you like the content and want to send some love, here’s some options:
Tweet me - Comments and questions are always welcome, they often inspire new posts
A small one-time donation at Ko-fi - Thanks to everyone who’s sent a small donation! I read every single note!
If shirts and swag are more your style there’s some here - There’s a plane w/ dots shirt available!
Interesting conclusion, and I think I agree with you that being 'data driven' purely with a purpose of optimizing a specific process is not necessarily a competitive advantage in commerce.
I'd like to hear your perspective on it from a public sector perspective though. In my mind being 'data driven' in the public sector is primarily about transparency. In essence showing with data the consequences/results of certain policies.