The Goldilocks Problem and Business Analytics

JUNE 28, 2017

Technologies to acquire and store data are becoming cheaper every day, which means that enterprise retailers have an unprecedented ability to capture and process terabytes of data. To harness this firehose, most large retailers have a sophisticated business analytics unit, responsible for maintaining key performance indicators (KPIs) and distributing the reports summarizing them throughout the organization. 

It starts with the best of intentions. As a data-driven retailer, you want everyone in your business to have the right information at the right time to drive the right decision. Yet in the quest to become data-driven, you may have achieved the opposite. When reports become hard to find, people stop looking. When messaging around new business processes is opaque, people fall back on old habits. When KPIs are confusing, people make wrong decisions.

“Having too much data is often as counterproductive as having too little data.”

This is the Goldilocks Problem: If there is too little data available, or if data is too hard to access, people make decisions based on their instincts. If there is too much data available, people are either overwhelmed or are subject to “analysis paralysis”, and end up making decisions based on their instincts, or (and this is worse by far) not making any decision.

“In order to make decisions based on data you need just the right amount.”

At Square Root we understand this problem in large retailers deeply. CoEFFICIENT Metrics is designed to streamline the workflows in your retail enterprise by organizing information and leveraging artificial intelligence to support the decision-making process.

Consider a Store Manager, tasked with day-to-day strategy in her store. She has 5-10 KPIs that she likely focuses on daily. There are another 5-10 that become important monthly, quarterly, or yearly. This same dynamic is true for a District Manager—they’re generally tasked with the same KPIs as the Store Manager, but have to keep an eye on all of the (10-15) stores in their district.

But what about the other 90% of the KPIs you’re tracking? Why bother tracking them if they’re not generally useful? Of course the KPIs are being maintained for a reason, but if no one ever pays attention to them, what good are they doing? It’s not just about having access to the data—most humans simply don’t have the cognitive facilities to deal with such a large body of data (this is “analysis paralysis”).

The challenge, and the opportunity, is to figure out how to get your field and store management teams to “level up”. How can you help them base their decisions on more than just the 15 KPIs that loosely affect their bonuses? How can you help to give them more context, and keep them more aligned, to what you’re trying to accomplish? How can you reduce their cognitive load so that they’re spending more time giving your customers a first rate in-store experience?

When we built CoEFFICIENT Metrics, we designed it to accommodate five types of KPIs:

  • Focus Metrics are the 3-5 metrics that dictate 80% of your day-to-day decision making.
  • Related Metrics help offer context to the Focus Metrics, providing simple root cause analysis.
  • Index Metrics serve as a way to aggregate large amounts of information into a simple health check.
  • Forecast Metrics help you put today’s performance into context based on historical data
  • Recommended Metrics leverage artificial intelligence to learn patterns in your data and workflow in your employees, proactively finding hidden issues before they’re problems.

Everyday we work to help organizations avoid analysis paralysis with these five types of KPIs. They are they to help focus work for your team and provide direction across a large number of stores and people. Interested in how these KPIs could help your team? Learn more about CoEFFICIENT Metrics now. 


As an astrophysicist, John used to look for black holes and dark matter in nearby galaxies. At Square Root, he gets to put that knowledge to work on our Data Science team. He’s a real data power lifter with an interesting perspective who makes our data science team even stronger. With a true passion for problem solving, John is an expert in breaking down complex ideas and data into information we can all understand. He holds a Ph.D. from the University of Texas and a B.S. from Rutgers University in astrophysics.