Entering the Digital Age:

Step One Develop a Data First Strategy

If you are not in the business of data, you may think that data driven Artificial Intelligence and Machine Learning are just another technology fad, over hyped by tech consultants whose marketing departments have an awesome record of over promising and under delivering. However, today you need to think of yourself in the business of data and develop a ‘Data First’ business strategy.

Over the last couple of years I have been spending a lot of time on the issue of how to bring ‘non-digitally driven’ companies into the digital age. Even though I have built and been involved in several ‘digitally driven’ companies I have struggled to communicate the importance of and exactly what is a ‘digitally driven’ company. However, I have come to understand that the first step toward being a ‘digitally driven’ company is having a Data First Strategy.

Most successful businesses are already run with a variation of this in that they operate with an ‘Information First’ strategy. We do not make strategic decisions without first gathering information and analysing it. This is Business School 101 stuff; Porter’s Five Forces, Kaizen, SWOT analysis, etc… In fact, we do not even call it a strategy because it is how we are hardwired. This framework is at the core of how we have successfully evolved as a species to the top of the food chain. We have the ability to gather new information and change our behaviour based on this new information, while all other species rely on hard-wired, slow-to-change instincts to make decisions.

HOWEVER, we are on the verge of a tectonic shift in the way humans make decisions. In fact tectonic, might be the wrong metaphor, because even though this is a change at the scale of the earth’s surface moving about, it is going to happen much, much quicker.

So what has changed? Computers have been around for many decades and have become an increasing integral part of our decision making process. However, to paraphrase Dr. Richard Sutton1, renowned reinforcement researcher, “Until this year it was cheaper for the human brain to do computations than computers.” Sutton explains that many of the AI algorithms that are in use today were actually written over a decade ago, but we didn’t have the computation power to execute them. (The comparison of the computation pros and cons of the human brain and computers is a complicated and deep subject and well beyond the scope of this article, I would recommend reading Ray Kurzweil’s book “The Age of Spiritual Machines: When Computers Exceed Human Intelligence”2 to understand it in more detail.)

So what has changed is that until now the human brain has been by far our most effective decision making tool. We are only at the ‘Tip of the Spear’3, the human brain is still the most popular decision making tool, however, this is changing. It is no coincidence that there has already been a fairly dramatic changing of the guard of the world’s most powerful companies. In the last five years we have officially made a change from industrial and energy companies to tech. The current top 5 companies in North America all have been leading the way with digital strategies. And to bring this back to the thesis of this article, an important step to take advantage of new digital technology is to develop a ‘Data First Strategy’ within your firm.

So What is a Data First Strategy?

A Data First Strategy is first about changing the way you think about and gather information for decision making. The human brain is awesome at taking in a jumble of information from different sources, different senses and in a very unstructured way. The value of our big brain is that it can take this mish-mash of info and generally make successful decisions from it.? The human brain is like Microsoft Excel, you can take data from all different types of sources, enter and/or import it into excel spreadsheets, build formulas and some relationships… basically reorganize the data in a way to come to some conclusion. However, computers and specifically AI and Machine Learning do not work this way. AI needs data in a structured, organized, electronic format (Excel, although in ‘electronic’ format, usually has to be restructured for AI to read it, so usually does not count as electronic).? A good rule of thumb is that your data needs to be in some kind of database.

Start with some easy wins and easy business process improvements

For established businesses start with what are your key business metrics and work backwards on how you collect the information to get to those metrics. Do a flowchart of all of the processes that are required to get the data for those metrics and highlight all the places that any kind of human interpretation or human intervention is required and look to reduce or remove them. Look for processes where information needs to be re-entered from one system to another. Second make sure that the data from all the required processes is available in a structured format.

This is called ‘Data First’ because you need to be relentless when looking at existing or new processes and make sure you getting your business data into a structured format. Even though at this stage you are not using any AI or Machine Learning technology it still will have a significant impact on your decisions making. You will have more accurate and timely data on which to make better decisions.

A second outcome of architecting your business with sound data driven processes is that it will be much more scalable. Processes that require human intervention usually mean increased costs and errors as you grow.

Build data analysis into your culture

Early I talked about how the human brain is good at gathering information and using that information to make decisions, HOWEVER, there is a very well established dark side to this.? Quite often our brains will see ‘information’ that supports a decision we have already made and filter out information that does not support this decisions.? By putting more importance across the whole organization with the gathering and valuing of clean, structured data you will greatly reduce cognitive bias.

This cognitive bias has been extensively written about a couple of my favorite reads are Moneyball: The Art of Winning an Unfair Game by Michael Lewis? and Fooled by Randomness: The Hidden Role of Chance in Life and in the Markets by Nassim Taleb.?

Lastly, getting ready for AI and ML

The above steps will prepare your organization for both getting your data ready for AI, but more importantly building a culture that can look at different ways of analysing data and thus new ways of making decisions. A ‘Data First Strategy’ is not the end game, but rather a framework for building an organization that can not only handle, but embrace what AI is going to throw at us in the future.

Notes and Further Reading

1 Richard Sutton: https://www.ualberta.ca/science/about-us/contact-us/faculty-directory/rich-sutton

2 Kurzweil, Ray, The Age of Spiritual Machines: When Computers Exceed Human Intelligence, NY: Penguin Books (1999) https://en.wikipedia.org/wiki/The_Age_of_Spiritual_Machines

3 Jim Gibson has written an important book “Tip of the Spear” on the social impact that technology change is having on our society. 

Gibson, Jim, Tip of the Spear (2018) https://jimagibson.com/

? Although, sometimes I tend to agree with Kurt Vonnegut Jr that our big brains are what get us into trouble. https://articles.latimes.com/1985-10-23/news/vw-14060_1_brains

? Data Scientist spend over 80% of their time getting data into a readable format. I suspect this is going to get worse before it gets better.

? Cognitive Bias: https://www.verywellmind.com/what-is-a-cognitive-bias-2794963

? Lewis, Michael, Moneyball: The Art of Winning an Unfair Game, WW Norton, (2003) https://en.wikipedia.org/wiki/Moneyball

? Taleb, Nassim, Fooled by Randomness: The Hidden Role of Chance in Life and in the Markets, Random House (2005) https://en.wikipedia.org/wiki/Fooled_by_Randomness