BI Leadership

How to Avoid Analytics Failure—A Tale of Two Strategies

Did you know that Gartner estimates that 50% of all Analytics projects fail?  

Have you taken part in one or more of these projects?  

I have, and I have also seen my fair share of these projects from afar over the last 17 years in the Analytics space.  Each project had its reasons for failure.  The focus of this post is to share common concepts that I have seen in the most successful Analytics projects over the years.  When you are done reading this post and series, you will be better prepared for your own successful Analytic Project, and be more likely to drive results for your business.

A common theme that I have seen across successful Analytics projects is separating projects into two distinct paths, one focused on the process of making impactful decisions and one focused on data.  Let’s call the decision making process the Analytics Strategy and the data gathering process the Data Strategy.    

Analytics Strategy

I’ve seen four best practices drive successful Analytics Strategies for organizations. 

The first is to find high impact areas where analytics can make a difference in a organization.  This is typically much harder than people think (I have seen external consulting help in this process) but is very important to get focus and measure results.  

The second best practice involves the decision making process itself.  Train your decision makers how to use key inputs that are inclusive of data (not exclusive) to make decisions and mitigate risk. 

After understanding your focus areas and after participants are trained in the decision making process, drive decisions in the impact areas as soon as possible.  

The final and one of the most important elements of an Analytics Strategy is to educate the organization how to respond, not react to decisions, both good and bad.  Many people use react and respond synonymously.  In the article “React vs. Respond, What is the difference?” Dr. Matt James discusses the difference and provides insight on how to best control reactions and help develop response mechanisms that can be helpful in business.   

Data Strategy

Like the Analytics Strategy, I break the Data Strategy into three common best practices.  As you explore these, keep in mind that information is often lost in translation between decision makers during the data curation process.  

First, determine what data is high impact for the decision process.  For example, if an organization's focus is to win at the margin level, supply chain decisions make a huge impact on performance.  The data for these decisions could come from sourcing, inventory management, distribution and logistics, quality management or supply chain efficiency.  

What metrics in these areas will drive the most return?  Most people would certainly think that unit cost is important—and it is—but metrics like responsiveness ratings, late deliveries, change in inventory costs, damaged units, defects, or time to market may drive quicker decisions on the manufacturing process or supplier and distribution partners that could maximize margins over the year.  

Many times, I have seen the lack of ability to define high impact areas and the data to support the decision process be the downfall of Analytics Projects.  

The second best practice for a Data Strategy is to assess and document the risk associated with the data for the decision process.   The data will often not be perfect, but if the risks are understood the data can still help an organization make impactful decisions.  

The final practice is to create the fastest method to get data into the decision process.  This may introduce risk into the decision process, but can be mitigated with other inputs.  Successful organizations do not inhibit the ability to make decisions in high impact areas by creating unnecessary technology requirements.  

Speed is the key to success.  

I’m certainly not saying a technology platform is not important but its primary requirement should be to provide the ability to quickly get data into the decision process.   

How important is data quality?

“At this very moment, there’s an odds-on chance that someone in your organization is making a poor decision on the basis of information that was enormously expensive to collect.” 
—Harvard Business Review

All of this is dependent on quality data, of course.

But what is quality data?  Who is looking at it?  And how do they use it?

The quote above is from an article that anyone who must leverage data to make decisions in business should read.  

In their research of 5000 employees at 28 global companies, the Harvard Business Review categorized decision makers into 3 groups: unquestioning empiricists (numbers driven), visceral decision Makers (gut driven), and informed skeptics (many inputs.)  They found that the informed skeptics outperformed the others in effectiveness, productivity, employee engagement and business results.  

There was one major problem—there were just not that many informed skeptics.  

I’ve seen this validated in the companies I have worked with.  Successful data driven companies focus on the process of developing informed skeptics and the use of data in the decision-making process—not on the data itself.  Obviously you want the best available data, but in this research they identified that in companies with a data driven culture, “If given the option of good-enough data now or perfect data later, most executives choose the former, confident that they can apply judgment to bridge the gaps.”   

The question isn’t really whether the data you have is good enough or perfect.  The right question is, “what data do you have today?”  Experience can bridge the gap today, and data quality can catch up to the pace of your business.

The role of technology

I have seen many analytics projects that are driven by a specific data technology.

This is a mistake.  The most successful Data Strategies do not focus on a single warehouse methodology, technology stack, or standard.  Successful strategies focus instead on using an array of technologies that organization have at their disposal to deliver data with speed to the decision-making process.  

I’ve seen technologies that range from simple data extracts to “Big Data Projects”—a topic which deserves its own post.. The common requirement for these Data Strategies was the speed of data to the decision-making process, not boiling the proverbial data ocean.  

A very good article that was recently published in the Harvard Business Review named “What is your data strategy?” outlines a strong framework based upon how some very successful data driven organizations created data strategies around offensive (flexibility to business groups) and defensive (control and consistency) data concepts.  For those of you struggling with how to start or revamp a Data Strategy in your organization, this article offers some very good insight.  

It is very rewarding to be a part of a cultural shift in an organization that includes analytics as a driving force in its business.  Focus your Analytics Strategy on the decision-making process and focus your Data Strategy on providing speed of data to the decision-making process and see the improvements that are common among the most successful data driven organizations.