An Inquisitive Citizen Data Analyst

Every day, people around the world generate a collective 2.5 quintillion bytes of data. To put that in perspective, one “quintillion” is a 1 followed by 18 zeros — it's a number frequently used to quantify the mass of the Earth in tons.

istock 1096811078
iStock

Every day, people around the world generate a collective 2.5 quintillion bytes of data. To put that in perspective, one “quintillion” is a 1 followed by 18 zeros — it's a number frequently used to quantify the mass of the Earth in tons.

One aspect of my job is to make sense of sometimes staggering volumes of information. Which data sets are crucial? Which ones can we simply ignore? Where is the actionable signal amidst all of the noise?

I'm far from the only business person asking these sorts of questions. In today's digital economy, every business must make use of data to outsmart its competitors and please its customers. In short, successful brands need all of their people to think like data analysts. So how do you do that?

The rise of the citizen data analyst

First, businesses should strive to make every employee a “citizen data analyst” — someone who lets data drive their decision-making process. Citizen analysts are analytical thinkers—not trained analysts—who use data to advance specific goals and objectives of their job function, such as boosting revenue or reducing costs. These objectives need to be concrete and quantifiable.

Citizen analysts must keep an open mind. They should be cautious not to develop a hypothesis and go out in search of data that supports it. Instead, they analyze data first and figure out what it's telling them.

Consider what happened when UPS analyzed their operational data with the goal of discovering waste. After crunching the numbers, analysts at UPS found that shaving a mile off each driver's daily route could save $50 million each year. They also found that trimming idling time by one minute per driver could save $515,000 in fuel costs and over $14 million in operational costs.

Key to this data-driven approach was beginning with the right question. Consider the difference between these two questions in the context of the UPS example: 1) Where can we gain greater efficiencies in our delivery operations? versus; 2) I think we are paying our delivery drivers too much, is that true? The second question contains a hypothesis in the form of a question, creating bias that would ultimately never allow the analysis to look at factors in a broader scope, such as delivery vehicle idle time.

The biggest mistake I see in analytics is when questions are posed that look for data that supports a hypothesis, rather than looking for the strongest signal in the data. This happens all the time—in fact, I made this mistake with an analyst on our team just the other day.

Show, don't tell

Identifying insights is only half the battle. Too often, employees don't pair their factual findings with actionable recommendations. Successful data scientists don't just tell the business “this metric is off” or even “this metric is off in this way” — but they add “this is what we should do about it.” This is often referred to as progressing beyond “what,” “why” and “how” to “so what?”

UPS did just that. The brand began using algorithms to determine the shortest routes for its drivers, installed sensors to ensure that drivers were driving safely and efficiently. Due to the changes, UPS drivers drove 85 million fewer miles per year, saved 8.5 million gallons of fuel, and prevented 85,000 cubic tons of carbon emissions.

When the haystack becomes a needle stack

Today, there's no longer just one needle — or data point — in the haystack. There are countless potential useful data points. So the challenge is to decide which ones matter. Sometimes, the most important data sets aren't always the most apparent. That's why data scientists constantly push boundaries and let their curiosity take them down unexpected paths.

For instance, imagine that a hotel chain wants its data scientists to identify areas to save money or grow revenue. One obvious choice would be to analyze guests’ preferences to determine the optimal room temperature or experiment with different shampoos and linens.

That’s all well and good. But Red Roof Inn went in a different direction — and it paid off.

An analysis by Red Roof Inn found that on any given day, 90,000 passengers are stranded at airports around the world.

So the company used weather and air-travel analysis to determine which areas were hotspots for flight cancellations. It then sent targeted mobile ads to stranded travelers. Potential customers received texts like “Stranded at Chicago’s O’Hare? Check out Red Roof Inn.” The result was 10 percent growth year over year.

Did you find everything you were looking for?

Sometimes figuring out which data to pay attention to is as simple as asking your customers. When you check out at a store, cashiers are often trained to ask, “Did you find everything you were looking for?” In the world of data, the same question can be extremely valuable. Your customers want to point you toward an aspect of your business you should examine or improve – all you have to do is ask.

When it comes to data, it’s important to be purposeful. Always work with a goal in mind. But remember to be open-minded in your approach.

The era of big data is here. Businesses who embrace it are prospering.
Jeff Allen is senior director of product marketing for Adobe Analytics.

Related:

Copyright © 2019 IDG Communications, Inc.