If you’ve read any research report centered around key trends in business lately, it’s likely “big data”, “analytics”, or some related but not quite synonymous term was part of the marquee findings. The conclusions are often that if an organization doesn’t embrace analytics1 today, it risks becoming uncompetitive and eventually commercially irrelevant.\nWhile business professionals may have ambitions around analytics, the taxonomy of these terms is confusing enough in itself. If my business uses elaborate executive dashboards that relay by-the-minute corporate information are we “doing analytics”? Or to make that claim must we have a team of data scientists applying intense statistical analysis within every corner of the organization? \nThe business case for embracing analytics is hard to refute; a quick search for success stories shows the dividends successful execution can pay. That said, simplifying the approach and cutting through the semantics is a necessary step in deriving real benefits from analytics. \nBusiness intelligence or analytics?\nBusiness intelligence and analytics are terms that can be considered closely related in the business context, though with notable differences. The path many organizations take in developing analytics programs is adding additional yet complementary capabilities beyond their already established business intelligence programs. \nHere is a useful oversimplification: when you cross the boundary from analysis based on the standard order of operations (addition, subtraction, and so on) into statistical analysis, you’ve moved from business intelligence into analytics. \nThough statistics is a vast field, regression is one of its most useful tools. In fact, it is one that represents a massive portion of statistical analysis taking place in business today. \nSimply put, a regression is an attempt to measure the strength of the relationship between two variables. In a business setting, let’s say you wanted to determine what impact an increase in advertising expenditures would have on sales. A regression assessing the historical relationship between advertising and sales could help you in that effort (with some additional nuance likely required). \nSimple linear regression (like our two-variable example above) may be the most rudimentary and easy-to-implement node on the analytics spectrum, with artificial intelligence, machine learning, and deep learning residing at its far, more specialized end. \nReading headlines, it would seem that these more specialized approaches beyond regression are some form of arcane magic, with infinite potential to improve business performance. And to be sure, analytics professionals consider them more robust and capable in many contexts. That said, there are a lot of similarities between these approaches and regression: the association between variables holds a key influence within each.\nBusiness intelligence is commonplace in organizations today. Using accounting system outputs to complete variance analysis would be considered business intelligence. Using a database to populate executive dashboards used to make business decisions would also fall under the umbrella. Again, business intelligence is analysis, but it is generally not statistical. Business intelligence tools can include enterprise resource planning (ERP) systems, databases (SQL as an example), and data visualization tools (Tableau, Power BI).\nAnalytics relies on different sets of processes and tools. From widely used programming languages such as Python and R to data processing software like Hadoop, these tools are used by programming and statistics specialists. If you haven’t encountered these tools in your professional life you are likely not alone, as we’ll see in the next section.\nCanadian businesses slow to take the plunge\nStatistics Canada surveys the use of advanced technologies by Canadian businesses as part of its Survey of Innovation and Business Strategy. Included in advanced technologies are “business intelligence technologies”, which include technologies like executive dashboards and data processing software like Hadoop. Though the survey uses business intelligence as a catch-all that includes technologies both in the business intelligence and analytics realms, the category is a reasonable proxy for the types of technology an organization would seek out when looking to increase analytics capabilities. \nIn 2017, 36.9% of large organizations (250+ employees) employed some use of these technologies, compared with 29.2% for medium-sized organizations (100-249 employees) and 20.7% for small organizations (20-99 employees). These figures indicate that the application of these technologies is by no means ubiquitous in Canadian business. \nThe bottom line\nBecoming an organization that embraces analytics can seem daunting. Cultural resistance and lack of alignment across departments can cause headaches throughout implementation; these are both likely driving forces behind the tepid pace of adoption we’ve seen among Canadian businesses to this point. Despite that, the benefits of analytics are tangible and successful execution is within grasp. Cutting through the buzzwords to uncover the set of solutions appropriate for your business is a good starting point. \n1I’ll use “analytics” throughout this piece primarily, but generally speaking it’s interchangeable with big data and statistical analysis, and also highly related to terms like machine learning and artificial intelligence.