Big Data is a buzzword that has become part of our business lexicon in recent years. What exactly “Big Data” means,however, remains amorphous. For most private businesses, (such as ours) it denotes something useful to only the largest corporations, who have the financial resources to delve into the murky world of web generated data bases, with the possible objective of producing synthesized strategic forecasts from complex algorithms. It’s certainly not something that would yield the kind of real world, street level information that will improve smaller business performance. We have reports that satisfy our more basic management needs, and they largely seem to be sufficient. Nonetheless, there is a sense that at some point Big Data will impact us, like a large cloud approaching that we know will eventually rain on us.
Big Data is not the only perturbing term on the business horizon. Internet of things, data analytics, data visualization, machine learning, artificial intelligence and a host of others that portend change on a global scale are out there. The purpose of this document is to see if, and how, any of this is applicable to private businesses like ours.
The data in a business
“Data” is at the root of all these developing analytical systems. It’s simply a record of something that happened, and has been around since writing was invented some 5,000 years ago. Today’s businesses generate increasingly large amounts of data due to the computerized systems we all use. Almost all the data our business generates is never accessed or used. Literally, every time a computer does something it’s recorded in an “event log” which can amount to thousands of records each day. All this data is usually deleted and overwritten automatically within a few days. It’s typically only useful to technicians in diagnosing software or configuration problems.
The kind of data we find useful in running our business is usually transactional in nature, and results from the interaction of three basic databases, the Customer, Product and Vendor database. These databases and the transactions between them, make up the figures we use in our everyday business operations management.
Sales management, forecasting and prospect development is often handled by a separate program such as goldmine or salesforce.com that generates its own data. And frequently, sales people keep additional notes on their laptop or mobile devices. Email, text, social media and voice connections are another source of data that can be close to our business operations but in disconnected databases. Accounting and finance is yet another business area frequently handled by a program separated from operations and has its own databases.
As we can see, there is significantly more data generated by our businesses than we use. And most of that data is not correlated. In many ways, we are not “connecting the dots” and gaining useful business insight from the “breadcrumbs” of data resulting from our business activities. As the saying goes, “data is not information”. Part of the goal of data science is to turn data into information regardless of business size.
Beyond reports
Today’s computing power and analytics technologies allow for a far more interactive relationship with our data than the flat reports we’re used to using. We’re no longer restricted to reviewing data in the form of strictly defined historic numbers. Today we can view and manipulate the flow of data in multiple dimensions. We can now gain insight from the volume and rate that data is generated in addition to its absolute numeric value. We all know that our businesses are living systems, and when we make a policy or procedure change in one area, it can often have ripple effects in other areas. We can now see all these effects in a single multi-dimensional interactive display.
It’s the difference between looking at a paper report of sales by region for last month, and looking at a report of sales, margin, customer, item, resources, shipping, time, success and failure, and location, on a timescale as small as minutes, and as large as years, in one display. And one that allows us to “drill” on any variable, while showing the resulting impact on all the other dimensions, instantly and intuitively from our own computer. We not only get to see the numbers, we see what causes the numbers to change and why. In other words, we get to see what’s behind the old paper report number. This kind of information power is transformational. And yet it goes further than that. All this can be appended with data from other sources. Public records of all kinds, our industry specific data, internet of things data and search results from online programs such as google, maps, social media, LinkedIn, Facebook, email, text and others can all add context to our business information. And this is all possible for us today.
Data science and the smaller business
Albert Einstein said, “If I had only an hour to solve a problem, I’d spend 55 minutes thinking about the problem and 5 minutes thinking about the solutions”. There’s nothing more wasteful than getting the right answer to the wrong question. The power of the technologies we have been exploring here is entirely dependent on us asking the right questions.
Data science is a relatively new, and potentially powerful discipline to business that focuses on modeling large amounts of disparate data and applying algorithms to produce statistical analysis results. Larger companies are beginning to employ data scientists as part of their management team in the hope that these new wizards will produce enlightenment and clarity in the conference room. Unfortunately, unless they have a firm grasp on the questions, their results can be misleading at best. The phenomenon of corporate “group think” can exacerbate this problem.
As owners and executives, we have an instinct about where the problems and opportunities lay in our businesses. We, more than anyone, know how our business works from beginning to end, and usually in greater detail than our fortune 500 brethren. That’s partly what makes us so agile and gives us our competitive edge. We make policy and procedure changes based on our experience. When the knowledge and experience of an owner/executive is paired with the expertise of a good data scientist, the results can be formidable. As we guide the data scientist with the right questions, they can bring clarity to our instinctive understanding of a problem, or they can point us in a new direction by revealing previously unsuspected facts. Applying good data science to smaller businesses produces reliable solutions more rapidly with greater clarity that enhance our agile competitive edge, and significantly improve the odds of success.
Roger Noakes, owner of Teamlogic IT, Setauket, NY, is a strategic partner of the Bardess Group. Roger and his team augment the Bardess offering by providing managed IT services for our SMB clients.
