Several months ago, in this same column, we touched upon data analytics and what makes data so important for creating a business advantage in broadcasting. You may recall the references to Moneyball with an emphasis on using data to predict the future and essentially gain a competitive advantage. We’ve heard about predictive analytics and how imperative the field is becoming to all professions, including broadcasting.
Now what? What can we do to implement analytics to improve our profits, ratings, and our company overall?
The first step is realizing that analytics apply to all departments, not just sales, and ultimately work for all formats of information, not just numerical data. The next step is searching within your department and understanding exactly what kind of data is already at your disposal. A wise man (Lindsay Wood Davis) once said, “Look for the magic beans…what are the magic beans that will lead you to answers?”
Relevant information is all around us, we simply have to seek it and capture it.
A sales department has access to billing, pricing, revenue, and budgets. Programming has data on cume and time spent listening/watching broken down into basic demographics. The promotions department has information on budgets and inventory. Engineering may have web and streaming statistics, along with social media data. We can also collect textual data from listener comments and emails, or even company goals.
We already know how fairly easy it is to predict revenue or ratings based on trends of past data. What about something less tangible? For decades, we have relied on instinct in a multitude of situations in our industry. This includes decisions made on station content, promotions, even hiring and employee retention. But instinct is not based on facts, and data analytics is. The other issue with instinct is that human beings come with emotional baggage and natural biases, and these attributes ultimately influence decisions, consciously or subconsciously, for better or worse.
Now let’s look at a non-traditional example of using analytics in the workplace, a process frequently used more seriously by industries outside broadcasting. Say we want a more successful retention rate for the sellers we hire. Our current system is a simple, yet not always effective, method of hiring based on the applicant’s references, personality, and history. These are all imperative themes in hiring a salesperson, but instead of saying, “She seems like she has a knack for sales,” let’s measure these qualities empirically.
This predictive system would work much like models used by financial institutions that choose whether or not to grant an applicant a loan based on predictive analytics of former applicants. A bank would take into account factors such as income, homeownership, age, and education. For our example, we’d find all the data we have on our salespeople, past and current, and create a model that will predict how likely a new hire’s success will be with the company. You can add as many factors into the mix as you’d like—education level, prior broadcast sales experience in years, how long they stayed with the company, number of new clients added, amount billed per year or month, etc. Typically, the more factors you consider, the more accurate the outcome.
Now the mathematical part of this: How do we execute this? You don’t need to have a statistics degree to solve this. While it does take some time to learn, there are several programs available. Some are free and some may cost you after a free trial. In the end, however, these costs in time, training, and software are dwarfed by the savings in making more analytically informed hiring decisions of our sellers (again, one of infinite examples).
We can keep guessing and hope our instincts choose the best, or we can let the data do the work and ultimately affect the bottom line.