Tuesday, September 6, 2016
Using Science to Make Business Decisions
Why Use Trial and Error?
Even in our current modern age of computers and mobile technology, too many businesses use trial and error or other decision-making techniques without any evidence of potential results before they get started. Keeping in mind the adage of what happens to people who make assumptions, there are better ways to decide on business matters.
Using the Scientific Method
If you think back to your high school days, you may remember learning about the scientific method. Like many high school students, if you didn't pursue a career in science, it is likely that you have not thought about the scientific method in recent years. However, business is more like science than you might expect. You can prove and disprove many theories with factual evidence before risking time and money on a new project or campaign. Why should you risk your company income and employees' paychecks when you can test theories before you take the plunge? If you could predict behavior, you would be able to achieve much more reliable results.
Let's see how much you remember about the scientific method. The basic method is to create a theory and then set up a scientific experiment to test your theory. You need a test group equally divided into control and experimental subjects.
Setting Up the Experiment
Google is a prime example of a company that tests its theories on a regular basis. They are constantly running tests to see how people react to various changes in their search engine. When they find a particular change that nets the results they want, they then implement the successful change over a larger group of search parameters.
Tests have been run by various companies to answer questions such as these:
Do lobster tanks increase lobster sales at Food Lion supermarkets?
Do eBay users bid higher in auctions when they can pay by credit card?
Do Subway promotions on low-fat sandwiches increase sandwich sales?
Does a Toronto-Dominion branch get significantly more deposits when open 60 hours a week compared with 40? (from https://hbr.org/2009/02/how-to-design-smart-business-experiments)
When Tests Do Not Work
Testing does not work in all situations. You have to have a large enough collection of data to learn anything significant. However, when you do have enough data to create a test, testing will give you measurable and repeatable results. According to the Harvard Business Review,
"Whether in marketing, store or branch location analysis, or website design, the most reliable insights relate to the potential impact and value of tactical changes: a new store format, for example, or marketing promotion or service process."
If you have a situation with specific, measurable results, instead of guessing the outcome and taking the risk, create a test that will give you a valid answer and confidence in your investment.