Analytics is meant to improve decision-making, but it can also do harm. One of the most common harms is “analysis paralysis” where the organization is so busy analyzing the options that it doesn’t make a timely decision. Business leaders need to act to mitigate this risk, especially if they are investing a lot of money in an analytics capability.
There are two categories of causes that lead to analysis paralysis:
- Technical – the organization is not skilled at using analytics to improve decision making
- Political – the organization doesn’t want to make decisions
Let’s consider these two causes and possible solutions.
Technical Causes of Analysis Paralysis
Here are three common ‘technical’ causes of analysis paralysis:
- Analysis rarely gives a clear answer as to what decision is best. This may not be apparent to analysts or their managers because popular stories about analytics always pretend there was a clear-cut right answer. Managers may ask analysts for more and more studies in the hope that an unambiguous right answer will ultimately emerge. It never does.
- The purpose of the analysis is not sufficiently clear, so analysts produce data that doesn’t shed any light on what action to take.
- Analysis is fun and the analyst is more interested in doing that than in making a decision.
Solutions to the Technical Causes of Analysis Paralysis
The heart of the solution to the technical causes of analysis paralysis is recognizing that using analytics in the organization takes managerial skill, not just skill in mathematics. Managers need to be trained how to use analytics in decision-making. This training (or more broadly ‘learning’) should cover awareness of when to use analytics, how to ask a clear answerable question, and how to deal with the resulting ambiguity.
Related: What is HR analytics?
After managers develop skills in how to use analytics (note it’s “how to use analytics” not “how to do analytics”) then they’ll be comfortable working closely with analysts. The close working relationship will prevent many problems, including analysts having so little direction that they spend long hours playing with the data just for fun.
Political Causes of Analysis Paralysis
The political cause of analysis paralysis is usually that managers fear that if they make a mistake they will be punished. In this kind of culture managers will inevitably put off making decisions and ensure they prepare a thick raft of defenses when they are forced to make one. It’s tempting to say this is not an analytics problem it is a culture problem. Fixing all the technical causes of analysis paralysis will not have much impact if managers, in effect, want paralysis.
Related: How to create an HR dashboard
I wrote “it’s tempting to say this is not an analytics problem it is a culture problem”. Why did I use the word “tempting” rather than come right out and absolve analytics of being part of the problem? The reason is a fundamental aspect of analytics is this: Analytics = Politics.
Experienced managers know that organizational politics pervade everything that happens in a business. That’s okay; politics is the mechanism for aligning interests across a group of people—it’s inevitable, it’s necessary.
Related: Read more about HR metrics
The reason I harp on about the rather obvious point that “Analytics = Politics” is that people attracted to analytics are often hoping they can escape the messy human world for the pure world of numbers. We can’t escape, and we need to be reminded of that often.
So the political cause of analysis paralysis is quite clear, quite common, and we should be willing to face up to it directly and recognize we need a political or cultural solution, not a technical one.
Solutions to the Political Causes of Analysis Paralysis
An individual manager trying to both make a good decision and get that decision accepted has to see these as two distinct tasks. Analytics will help them make the best decision based on the available evidence, but it won’t help get the decision accepted. To get the decision accepted involves understanding stakeholders’ interests and personalities. It involves taking the time to build relationships and credibility. It involves investing thought, time and effort in influencing those stakeholders. This is all classic change management stuff; the fact that we may have used machine learning or some other fancy technique to make our decision doesn’t reduce the need for change management in any way.
As a leader in an organization that is routinely affected by political analysis paralysis then there are two things you can do. The first is to teach managers to work within the system (as described in the previous paragraph). Yes, the system may resist action, but you can teach managers how to get around that resistance. The second approach is to work on the system. You don’t want a culture of fear where decision-making is punished—so you need to change that culture.
Culture change often goes wrong because it swings between the poles of two extremes. Fear is bad, but so is lack of accountability. The trick is to articulate the good and bad of both extremes (“we don’t make mistakes, but we are very slow” vs. “we are fast, but make foolish mistakes”). There are many books on culture change, but this particular issue is explained in Barry Johnson’s The Polarity Principle. It’s a must-read. We do want to reduce analysis paralysis, but not to the point where we become careless.
- Train your managers how to use analytics, not how to do analytics. This training should ensure that they don’t overlook the core fact that Analytics = Politics.
- If managers determine that the shortcomings of analytics are not technical, but arise from a culture of fear, then they need to tackle that broader issue head-on. One key is to recognize that a culture of fear is not all bad, it’s actually the opposite pole of a culture of no accountability. The goal isn’t to drive out fear to the point that people become sloppy and unaccountable. The goal is to navigate that space where you are adequately careful and disciplined while being adequately fast and innovative.
Check David’s latest post on the biggest mistake that is made when producing an HR dashboard.