Predictive analytics is an advanced form of statistical data analytics that seeks to make predictions about future events based on historical data.Summary by The World of Work Project
Predictive analytics is the process of making predictions about the future that are informed by historical data. This is usually done by looking for relationships between different types of data in historic data sets and changes in historic data over time, then using these to make predictions about the future. This, of course, assumes that past actions can be used to make future predictions, which is often the case.
At it’s simplest we can think of predictive analytics as having computers look for things that usually happened in the run up to a specific event, then saying if those run-up things happen again, the same outcome event will probably happen again.
An Example to Bring it to Life
We could analyze the web searches of employees in an organization. If we did, we might find a relationship between the number of people who search for new external jobs, and the number of people that go on to resign from the organization.
As a result of our analysis, we might be able to calculate that for every 100 people that search for external jobs this month, 2 people will hand in their notice next month.
If this relationship holds true, we can use it to predict the future. For example if we notice this month that 600 people in our organization have searched for new external jobs, we can predict that 12 people will hand in their notice to leave next month.
This is, of course, a huge simplification of a process that usually considers a huge range of data points with multiple correlations and which may require huge volumes of data to be accurate.
Why do we care about predictive analytics?
As computing power becomes cheaper, as our software programs are increasingly able to learn and as our data sets grow, it becomes easier and easier for us to build and test the complex statistical models required to make accurate predictions about the future. And as it gets easier and cheaper to do so, we find more and more uses for this type of technology.
Increasingly we are seeing predictive analytics playing a larger and larger role in the world of work. Many organizations use predictive analytics in relation to their production and supply cycles, to assess consumer demand, to manage their just in time deliveries, to assess their expected server-loads and so on.
In recent years we’ve also seen predictive analytics take on a bigger role in back office functions and HR, where it’s now used for things like predicting employee turnover and retention, regulatory compliance breaches, fraud and future capability requirements.
The World of Work Project View
Predictive analytics is just statistical analysis aided by computers. We’ve been doing it for years and years in many areas, and now it’s cheaper and more effective and entering more areas of the world of work.
There’s not too much to say about it really. It’s neither good nor bad, it’s just what it is. Provided that we use the insights it provides in a good way it will help us, if we don’t it won’t.
Predictive analytics does lend itself well to the conversation about the automation and changing nature of work. With newer technologies come some suspicion and some loss of labor (for example if analytics and predict when people need to work, we may need fewer people to work in down times). That said, new technologies also bring new opportunities, for example many data analysts are needed to build and assess all of our analytical models.
Sources and further reading
Where possible we always recommend that people read up on the original sources of information and ideas.
The concepts behind this post is based on lectures we have attended, conversations with various podcast guests and general reading around the subject. There are no specific references for it.
If you see any errors on this page or have any feedback, please contact us.