Emerging from the Data Dark Ages

with Jennifer Belissent, Principal Data Strategist, Snowflake

Jennifer Belissent
Principal Data Strategist, Snowflake

Jennifer is a strategy and insights expert whose research focuses on data-for-good, particularly as it applies to climate science, humanitarian aid, and government. She spent 12 years as a principal analyst at Forrester Research, where her work focused on data leadership and literacy.

Satyen Sangani
Co-founder & CEO of Alation

As the Co-founder and CEO of Alation, Satyen lives his passion of empowering a curious and rational world by fundamentally improving the way data consumers, creators, and stewards find, understand, and trust data. Industry insiders call him a visionary entrepreneur. Those who meet him call him warm and down-to-earth. His kids call him “Dad.”

Jennifer Belissent
Principal Data Strategist, Snowflake

Jennifer is a strategy and insights expert whose research focuses on data-for-good, particularly as it applies to climate science, humanitarian aid, and government. She spent 12 years as a principal analyst at Forrester Research, where her work focused on data leadership and literacy.

Satyen Sangani
Co-founder & CEO of Alation

 

As the Co-founder and CEO of Alation, Satyen lives his passion of empowering a curious and rational world by fundamentally improving the way data consumers, creators, and stewards find, understand, and trust data. Industry insiders call him a visionary entrepreneur. Those who meet him call him warm and down-to-earth. His kids call him “Dad.”

Jennifer Belissent
Principal Data Strategist, Snowflake

Jennifer is a strategy and insights expert whose research focuses on data-for-good, particularly as it applies to climate science, humanitarian aid, and government. She spent 12 years as a principal analyst at Forrester Research, where her work focused on data leadership and literacy.

Satyen Sangani
Co-founder & CEO of Alation

 

As the Co-founder and CEO of Alation, Satyen lives his passion of empowering a curious and rational world by fundamentally improving the way data consumers, creators, and stewards find, understand, and trust data. Industry insiders call him a visionary entrepreneur. Those who meet him call him warm and down-to-earth. His kids call him “Dad.”

The World Economic Forum estimates there are at least 44 zettabytes of data in the world.

That means that we have 40 times more bytes of data than there are stars in the sky.

We’re surrounded by entire universes of data — and like the universe, this data keeps expanding.

Political campaigns overcome odds by micro-targeting voters, by using data.

Most recently, public health officials allocate resources to the COVID-19 response. Again, by using data.

It sounds like we’re living in the Renaissance of Data. Where we have the tools, insights, and knowledge to analyze data and create better outcomes for all.

But I don’t think that’s true. As Aaron, one of my co-founders, put it: We think we’re still in the Dark Ages of Data.

There’s nothing wrong with the data.

The issue is us.


 

How We Are Struggling With Data Literacy

Even though we have vast amounts of information at our fingertips, we struggle with data literacy.

According to The Data Literacy Project, a staggering 76 percent of key business decision-makers do not believe they are data-literate.

Research from McKinsey estimates that less than 1 percent of the U.S. population is data literate.

It’s like we’ve invented the printing press, but we’ve forgotten to teach people how to read.

We’re never going to be able to tap into the power of data without making significant progress on data literacy.

On this episode, I’m talking with Jennifer Belissent, Principal Data Strategist at Snowflake.

Throughout her career, Jennifer has been at the forefront of the data revolution — whether that’s leading teams at Forrester or studying housing policies in the Soviet Union while getting her PhD at Stanford.

Talking with Jennifer gave me confidence that we can enter the Data Renaissance.

And, that’s an incredible opportunity in front of us.

Think about the innovations we can make if just five percent more of the population is data-literate.

Now, imagine if that number was thirty.

In a data-literate world… The possibilities are endless.


Producer Read: Welcome to Data Radicals, a show about the people who use data to see things that nobody else can. This episode features an interview with Jennifer Belissent, Principal Data Strategist at Snowflake.

On this episode, she and Satyen discuss misconceptions around data, organizational tactics for improving data literacy, and the evolving role of the Chief Data Officer in enterprise organizations.

Data Radicals is brought to you by the generous support of Alation, the data catalog and data governance platform that combines data intelligence with human brilliance. Learn more at Alation.com.

 

Data Literacy and the Role of the Chief Data Officer

Satyen Sangani: Jennifer began thinking about data literacy while working on smart city projects at Forrester.

But she noticed that there were very serious gaps in how people work with data.

Jennifer Belissent: The first report I wrote on smart cities happened around the time of the [2008 Summer] Olympics in Beijing. The Beijing government was investing a lot in technology — not just in the stadium, but also across industry. They were focused on transportation within the city, and improving public health, and seeding the clouds (so there’d be rain at the right time and not at the times they didn’t want it).

So, there was a big investment in technology. My first report on smart cities was called Smart Cities Think Outside the Stadium (or something like that).

From there, I took a step back and said, “What else is happening within this space?” And I started looking at cities and the different initiatives. And at the time a lot of people were focused on infrastructure. They were focused on IoT [the “Internet of things”], on sensors, building out that infrastructure, and making garbage cans and streetlights and parking places intelligent.

But these were the shiny objects. Everybody was focused on that, but they weren’t really focused on helping the cities use data to become smarter, to become more intelligent. So over the years I focused on how to structure an organization, how to improve a government’s ability to use that data. What was the role of the chief data officer (CDO)? What kinds of things needed investment — across people, process data, and technology?

And a big part of my work in data literacy is helping people understand what data is today — which isn’t necessarily obvious: not only in smart cities, but across companies today.

 

Everyone Works with Data — Even If They Don’t Realize It

Satyen Sangani: If you look at the thread of a lot of your work, you’re looking at what people do and their behaviors and information about the choices they make:

  • At the macro level: the governors who get massive subsidies from the central government
  • At the micro level: how people should choose housing and how housing should be subsidized

Data, fundamentally, is a lens into people’s behavior. A lot of understanding data is mapping it to that behavior in the world. And that’s a challenge, right?

How do you see that in your work at Forrester and where did you see people struggling — and succeeding?

Jennifer Belissent: One of the ways this became glaringly obvious to me happened a couple of years ago. I was talking to the chief digital officer at Sodexo. (It’s a food service management company that runs cafeterias around the world.) She told me they first started looking at their data at one of their cafeterias [in France] and saw an increase in the sale of breakfast sausage.

And that was pretty curious because breakfast sausage isn’t among the culinary habits of the French. When they dug deeper, they actually found the increase in the sale of breakfast sausage started happening at about the time they had changed their traditional cash registers for point-of-sale machines with individual buttons for each item.

The cashiers — maybe in the interest of time, or because the breakfast sausage button was the easiest to press — kept pressing that button. If managers of their site used that data for ordering supplies for the next month, they wouldn’t have ordered anything but breakfast sausage. And clearly, that wouldn’t have resulted in a positive breakfast experience for their customers.

That really highlighted the fact that these cashiers probably didn’t know they were working with data. If they’d been asked, they would have said, “No, we don’t work with data within our roles.”

I ended up doing qualitative survey with an online panel and we asked three questions:

  1. Do you work with data?
  2. Are you comfortable working with data?
  3. What kind of training would you like?

This was when I was doing the data literacy work, so I thought the third question would be the most interesting, to find out how people wanted to become more data literate.

But what really fascinated me were the answers to the first question:

  • “I don’t do anything with numbers.”
  • “My team’s not responsible for calculations.”
  • “We don’t deal with statistics.”

They were clearly associating data with numbers and they didn’t recognize what data is today. And like the cashier, they probably didn’t know:

  • How their companies were using that data
  • The value data potentially brought to their company
  • Their particular role with data, whether it’s capturing, protecting, or using it

That brought home the fact that there are huge gaps in data literacy. Within the industry, we’re all in data; tech is now very data focused. We think everybody knows about data. And often, the vendors promoting data literacy focus on those who are already data literate — the business analysts, the data scientists, those who are already experts — and it’s important.

But a fundamental part of my research on data literacy was to focus not just on the top of the pyramid. I want to work with companies to create a very strong base, with more awareness of what data is within the organization.

 

How to Develop a Data Literacy Program

Satyen Sangani: How does that bridge into this sort of formalized definition of data literacy? When you go into an organization and say, “You need to be more data literate,” what do you do?

Jennifer Belissent: There are a couple of ways of approaching it:

  • Through formal data literacy programs that teach people about data
  • Demonstrating what you can do with data and getting people excited about how you can use data to solve problems — as if to solve a mystery

There are a couple of stories I uncovered during my smart city research. Several years ago, Buenos Aires undertook a massive streetlight modernization project that replaced traditional lighting with LED lighting, which is supposed to be brighter and can improve public safety.

But when the lights were in place, they started getting complaints that the streets were dark! For this mystery, as was the case with the Sodexo cafeteria, they dug into the data.

By overlaying different data sets, they found out the complaints were happening during the summer months, when the leaves were still on the trees. LED lighting doesn’t diffuse — that’s one of the advantages; it remains concentrated and doesn’t get lost in the ether — and that’s supposed to be a good thing. But when leaves are on the branches, the LED lighting doesn’t “go around” them.

The data indicated the answer: The city changed its tree-pruning schedule to ensure there were no leaves blocking the lights in those summer months.

That shows how we can figure something out by bringing in data — data from a variety of different sources. It’s not just one data set that we’re going to use. With data diversity we’re going to be able to solve the problem. And people tend to get more excited about and can understand data in a more concrete way when they see it applied to a specific problem.

 

Making Data Personal — And a Way to Simplify Employee Tasks

Satyen Sangani: Particularly a specific problem they’re either experiencing themselves or are trying to solve. What’s interesting about data stories is that in the abstract or when it’s somebody else’s problem, people don’t quite understand it. Data tends to be pretty inscrutable and hard to access, but when you make it personal and germane and relevant to that individual, it can be really powerful.

Jennifer Belissent: The same goes with an employees thinking they’re going to be replaced by an algorithm and concerned about their job.

Another element of data literacy is helping people understand how to use and engage with data, and understand new roles and how data is going to change their existing job.

Another smart city example is how Chicago addresses its health inspections. There’s a group of health inspectors within the city doing their job the same way for years. Then they’re told they’re going to have to do it differently. Traditionally, they rip a couple pages out of the Yellow Pages and say to some inspectors, “You’re going to do A through L” and to others, “You’re going to go through M through Z.”

But that’s not very efficient. Inspectors tend to run all over the city. You might try to organize it by neighborhood, but that’s not necessarily efficient, either.

So Chicago installed an inspection optimization system to predict the likelihood of a health violation. They brought in a wide variety of data types for restaurant neighborhoods such as:

  • Were there noise complaints?
  • What time of year were complaints filed?
  • What was the weather?
  • Were there past violations?

You can imagine all of the data they can record. Then a prioritized list is given to the inspectors to investigate. Rather than visiting 30 restaurants that didn’t need an inspection, they now have a list of priorities where a violation is much more likely.

The inspectors feel they’re doing their job better — it’s easier to do their job. Showing them how data and these new techniques are going to make their job much easier is also part of data literacy.

 

Data Evangelism from Within the Organization

Jennifer Belissent: I mentioned how I see a data literacy program as a pyramid, starting with awareness at the bottom, and then the next layer is comprehension, which is understanding what data is. This is more for the decision-makers, regarding the kinds of questions they need to ask about data:

  • Where does it come from?
  • Is it representative of the population?
  • What’s the underlying analytic logic of the model?

They’re understanding the origins of the insights of the data they’re given.

At the top of the pyramid are the experts. Of course, you need to focus on the experts, making sure they have access to the newest tools and latest data sets. But I also advocate using those experts to scale, to give back across the organization.

The model here is ACES:

  • Awareness
  • Comprehension
  • Expertise
  • Scale

You’re building data ACES within your organization, but scale is really about communicating to the rest of the organization what data is and how it’s being used.

 


 

How to Build Data ACES Within Organizations

To do this, we’re seeing things like lunch-and-learns and showcase sessions. Maybe it’s having a specific lunch-and-learn around a particular topic or setting up a booth in the cafeteria and providing incentives for people to come by and see what the data team is doing.

It could be upward mentoring: assigning someone from the data team to an executive to help understand how to:

  • Navigate the dashboards they’re getting
  • Ask the right questions about where the data came from
  • Define a data project that will solve some of their business problems

 

Data Literacy Starts with a Communications Strategy

Satyen Sangani: You can also provide resources that help employees connect data to their everyday lives.

Jennifer Belissent: I’ve seen data literacy programs create cheat sheets. At Lurie Children’s Hospital in Chicago, they had cheat sheets for the clinicians:

  • On one side, medical concepts and related questions
  • On the flip side, possible techniques, available datasets, and tools to help address problems

What they found was that sometimes those clinicians had had some sort of data training in graduate school, but they’d never really put it into practice on the job. So for them, they needed that push. They needed that connection to their actual job to see where to apply some of the tools they already had in their toolbox. These are tools they hadn’t necessarily learned how to use on the job.

I’ve been talking to a number of CDOs who have been talking about their comms strategy. I’ve always thought of it as a data literacy program, but they really think of it as communicating:

  • How do we communicate what we do?
  • How do we get people excited about what we do?
  • How do we build that broader community?

 

Overcoming Emotional Obstacles to Data Literacy

Satyen Sangani: I would even take it a step further and think of it as marketing. These devices you described sound eerily similar to the things we give our salesforce in the enterprise software industry to enable and train them.

The lunch-and-learns sound and feel so much like events where we put up booths to teach people new to our products about those products and give them a demo. That marketing and sales problem strikes me as so interesting because many data radicals — whether they’re CDOs or people implementing self-service enablement programs — don’t always realize their job isn’t just to stand up some infrastructure and make some data available.
Do you see that as an emotional challenge for people?

Jennifer Belissent: Most of the successful CDOs I talked to see that as a big part of their role. When I ask them, “What are the skills required to be a chief data officer?,” what often comes up is business and technology skills, but also communication skills.

One of my favorite CDOs called himself the chief diplomatic officer, because a big part of what he felt he did was diplomacy — being the bridge across different stakeholders. It was negotiating to get access to data or prioritizing specific projects.

 


 

The Three Pillars of Data Governance

Satyen Sangani: Do you feel that’s prevalent with CDOs today? The first version of the CDO job was about getting data governance to prove compliance and making sure you’ve got the appropriate data retention policies. It was about ensuring the boxes are checked and certain policies are satisfied. Do you feel the transition of the role — to include the evangelism side — is complete?

Jennifer Belissent: I really do. About five years ago there was all this talk about CDOs shifting from being defensive to offensive. That metaphor came from American football. However, the better metaphor of what a CDO should be — because it’s not really a shift — comes from European football (soccer). The same players are always on the field. They play defense and offense and some of the best offensive players — the strikers — are actually defenders. Look at Sergio Ramos from Spain. He’s a defender, but he has one of the highest scoring rates in the league.

So it’s not a question of being defensive or offensive. CDOs have a mandate across the data value chain, across that whole life cycle of data. Data governance also extends across that life cycle. It’s not just about security or privacy or ensuring data quality; it’s also ensuring the right people can access it and use it to deliver value to the organization.

At Snowflake, we talk about three pillars of data governance. It’s knowing your data, protecting your data, but also unlocking your data to ensure people can collaborate securely using that data.

The CDO reflects that holistic, end-to-end view of data governance across that whole spectrum. One of the things that we’ve seen, particularly in the U.S., is that the CDO is increasingly reporting to the CEO and the chief executive in an organization. That does reflect more of the strategic view of data — and of data as a strategic asset — and has the CDO really driving that effort within the organization.

 

Building a Data Culture with Quick Wins

Satyen Sangani: And that’s kind of the essence of this entire podcast: How do you build a data culture? If you’re talking to a technical CDO who maybe hasn’t done this work, where would you tell that individual to start?

Jennifer Belissent: The most important thing for a new CDO coming in is to get a quick win under the belt. I’ll refer back to this “diplomacy” effort. It requires a “listening tour” where you talk to your peers to understand their issues and priorities to find something that’s going to demonstrate the value of data.

That’s your starting point. You need something to showcase as the way data delivers value into the organization. Demonstrate that to others to get them excited, and then more projects will come in, and it snowballs.

With the issue of prioritization, you go back to your stakeholders and assemble a data governance council (or whatever you want to call it) that sets your priorities.

At the same time, as you consider the project trajectory, you also need to consider the cultural aspects:

  • You’re evangelizing to your peers, getting them excited about current projects and proposing future projects.
  • You’re also segmenting your employee base — as you would a customer base for a marketing campaign — and pitch the idea of the value of data.

You’re showcasing with lunch-and-learns, training sessions, office hours for people who want to drop in to discuss ideas. Over time, you’re creating the cheat sheets, maybe launching a blog, and delivering content to address the base of the pyramid, at the awareness level. At the same time, you’re working with decision makers.

It’s not something that can happen all at once; it needs to happen over time. But in the first instance, it’s creating some of these quick wins so you can demonstrate value to people.

 


 

Build the Data Infrastructure as You Execute Data Projects

Satyen Sangani: I love the idea of a quick win, and I think it applies not only to the CDO or senior roles, but on some level to every data professional. Organizing and governing data — and making it available and describing it — can just seem like a daunting, painful, inaccessible task. And often you just want to lock down and focus on a data glossary for the next 12 months, and you can get so lost in those efforts.
I find that the people who are really successful at almost any level just say, “Look, I’m going to focus on one or two things and get that right.” And then do that cycle you’ve described.

Jennifer Belissent: One of the CDOs I spoke to described how he got his infrastructure in place. He started with a few projects, taking the data required for those projects and populating his data warehouse and data catalog. Then he added new projects, for which there was more data to put into the warehouse and the catalog.

Eventually, an enterprise data warehouse and catalog without saying to everyone, “Stop! We’re going to not do anything until we build out the data warehouse and data catalog.”

His big reveal was that over the course of the time spent working on those projects, they were building an infrastructure that’s now available for everything else.

EIt was a valuable lesson in how to approach this in a more agile way, not stopping the clock and trying to build out something before delivering any value.


 

Level-setting for Data Literacy

Satyen Sangani: Let’s take a step back for a moment.
If you want to build a data-literate organization, then you need to meet your employees where they are.
Jennifer believes that inclusion is a critical component of achieving this.

Jennifer Belissent: Although data literacy is often defined as the ability to read, write, and communicate with data, I like to add to recognize because there are people who don’t recognize what data is within the organization. They don’t realize they work with it. Therefore, a data literacy program needs to start at “level zero” to address “What is data?”

This helps people understand that everything within the organization is data-generating. Then you identify different personas within the organization, assessing where people are in their understanding of and ability to use data.

One of the things I found in the qualitative research I did was that many people are not comfortable with data. We’ve created haves and have-nots within an organization. The answers I received included “I judge myself” or “I feel bad about myself because I don’t understand when people start to talk about data.” These were questions of self-worth — and it really disturbed me that people felt like that — so we need to make sure data literacy programs are inclusive.

We also need to focus very heavily on the decision-makers and those whose jobs are really impacted by the use of data — those who are given a dashboard or some sort of data or insight and expected to make decisions on it.

At Forrester my research found that fewer than half of decisions are made based on quantitative information, as opposed to experience, opinion, or gut feeling, because we don’t really teach people how to make those decisions:

  • Evaluating data
  • Understand where the data came from
  • Knowing the underlying analytic logic

We need to focus on those decision-makers and anybody else whose job might have been disrupted by data, because they might not be out of a job, but their job is transformed and they need to understand how data plays into their new role.

And with a data literacy program, you can’t forget the experts, so give them access to Coursera or other online training courses, if available. Asking them to scale and contribute to building the data literacy program internally, and that’s a big part of their professional development: giving them speaking opportunities and helping them enhance their own skills. What’s a better way to help you understand something than having to teach it to somebody else?

So look at the different personas across your organization, understand their data maturity through some sort of assessment, and then create a program that delivers content to these various personas in the ways most suitable to them.

 


 

Building a Data Community

Satyen Sangani: I love this notion of an assessment because it sort of takes data in order to figure out where you should strategize with your data programs. And so you start by learning, which sounds obvious but it’s maybe something people don’t usually do.

Jennifer Belissent: An additional piece of advice I’d give is to build that data community. Talk about data, communicate about projects and successes, and incenting those data teams to help scale.

It also starts with onboarding. We offer sessions on business ethics and sexual harassment — often the things not to do — but we need to ensure they’re aware of data. If you want your company or city to be data-driven, everybody has to have that as a baseline for the way that they approach their job.

 

Data Literacy Requires Change Management

Satyen Sangani: And it has to be fairly real in the sense that you have these dual and competing notions of fear and change and then understanding data. You can understand data, but there are certainly cases where you can understand data and it adversely affects your role or your job, or maybe your standing politically because you’re doing and saying something a leader in a particular part of the organization doesn’t want to hear.
That’s where it can get quite complicated because then the cultural change gets tested and leaders have to put their money where their mouth is and actually drive change within the organization.
Do you find that leaders understand that this can even affect how they make decisions and maybe make them uncomfortable, too?

Jennifer Belissent: I think they do. Embracing change, as well as potential conflict, reflects the maturity of the leader. Having support from a chief data officer or other executive support, makes it easier to make internal changes.

Likewise, cultivating support from below is another way to address potential conflicts. Again, data literacy can help. I mentioned assigning an upward mentor to executives in the organization to help them understand data — and that’s as important as fostering data literacy at the grassroots level as well.


 

Translating Data Into Information

If we want to get out of the Dark Ages, the answer is simple — we need to translate data into information.

So let’s make sure we surround information with context.

Let’s work to lower the costs of using data.

Let’s reward the people and systems that provide transparency and usability.

And then — just maybe — our data Renaissance might be upon us.

This is Satyen Sangani, Co-Founder and CEO of Alation. Thank you for listening.

The World Economic Forum estimates there are at least 44 zettabytes of data in the world.

That means that we have 40 times more bytes of data than there are stars in the sky.

We’re surrounded by entire universes of data — and like the universe, this data keeps expanding.

Political campaigns overcome odds by micro-targeting voters, by using data.

Most recently, public health officials allocate resources to the COVID-19 response. Again, by using data.

It sounds like we’re living in the Renaissance of Data. Where we have the tools, insights, and knowledge to analyze data and create better outcomes for all.

But I don’t think that’s true. As Aaron, one of my co-founders, put it: We think we’re still in the Dark Ages of Data.

There’s nothing wrong with the data.

The issue is us.


 

How We Are Struggling With Data Literacy

Even though we have vast amounts of information at our fingertips, we struggle with data literacy.

According to The Data Literacy Project, a staggering 76 percent of key business decision-makers do not believe they are data-literate.

Research from McKinsey estimates that less than 1 percent of the U.S. population is data literate.

It’s like we’ve invented the printing press, but we’ve forgotten to teach people how to read.

We’re never going to be able to tap into the power of data without making significant progress on data literacy.

On this episode, I’m talking with Jennifer Belissent, Principal Data Strategist at Snowflake.

Throughout her career, Jennifer has been at the forefront of the data revolution — whether that’s leading teams at Forrester or studying housing policies in the Soviet Union while getting her PhD at Stanford.

Talking with Jennifer gave me confidence that we can enter the Data Renaissance.

And, that’s an incredible opportunity in front of us.

Think about the innovations we can make if just five percent more of the population is data-literate.

Now, imagine if that number was thirty.

In a data-literate world… The possibilities are endless.


Producer Read: Welcome to Data Radicals, a show about the people who use data to see things that nobody else can. This episode features an interview with Jennifer Belissent, Principal Data Strategist at Snowflake.

On this episode, she and Satyen discuss misconceptions around data, organizational tactics for improving data literacy, and the evolving role of the Chief Data Officer in enterprise organizations.

Data Radicals is brought to you by the generous support of Alation, the data catalog and data governance platform that combines data intelligence with human brilliance. Learn more at Alation.com.

 

Data Literacy and the Role of the Chief Data Officer

Satyen Sangani: Jennifer began thinking about data literacy while working on smart city projects at Forrester.

But she noticed that there were very serious gaps in how people work with data.

Jennifer Belissent: The first report I wrote on smart cities happened around the time of the [2008 Summer] Olympics in Beijing. The Beijing government was investing a lot in technology — not just in the stadium, but also across industry. They were focused on transportation within the city, and improving public health, and seeding the clouds (so there’d be rain at the right time and not at the times they didn’t want it).

So, there was a big investment in technology. My first report on smart cities was called Smart Cities Think Outside the Stadium (or something like that).

From there, I took a step back and said, “What else is happening within this space?” And I started looking at cities and the different initiatives. And at the time a lot of people were focused on infrastructure. They were focused on IoT [the “Internet of things”], on sensors, building out that infrastructure, and making garbage cans and streetlights and parking places intelligent.

But these were the shiny objects. Everybody was focused on that, but they weren’t really focused on helping the cities use data to become smarter, to become more intelligent. So over the years I focused on how to structure an organization, how to improve a government’s ability to use that data. What was the role of the chief data officer (CDO)? What kinds of things needed investment — across people, process data, and technology?

And a big part of my work in data literacy is helping people understand what data is today — which isn’t necessarily obvious: not only in smart cities, but across companies today.

 

Everyone Works with Data — Even If They Don’t Realize It

Satyen Sangani: If you look at the thread of a lot of your work, you’re looking at what people do and their behaviors and information about the choices they make:

  • At the macro level: the governors who get massive subsidies from the central government
  • At the micro level: how people should choose housing and how housing should be subsidized

Data, fundamentally, is a lens into people’s behavior. A lot of understanding data is mapping it to that behavior in the world. And that’s a challenge, right?

How do you see that in your work at Forrester and where did you see people struggling — and succeeding?

Jennifer Belissent: One of the ways this became glaringly obvious to me happened a couple of years ago. I was talking to the chief digital officer at Sodexo. (It’s a food service management company that runs cafeterias around the world.) She told me they first started looking at their data at one of their cafeterias [in France] and saw an increase in the sale of breakfast sausage.

And that was pretty curious because breakfast sausage isn’t among the culinary habits of the French. When they dug deeper, they actually found the increase in the sale of breakfast sausage started happening at about the time they had changed their traditional cash registers for point-of-sale machines with individual buttons for each item.

The cashiers — maybe in the interest of time, or because the breakfast sausage button was the easiest to press — kept pressing that button. If managers of their site used that data for ordering supplies for the next month, they wouldn’t have ordered anything but breakfast sausage. And clearly, that wouldn’t have resulted in a positive breakfast experience for their customers.

That really highlighted the fact that these cashiers probably didn’t know they were working with data. If they’d been asked, they would have said, “No, we don’t work with data within our roles.”

I ended up doing qualitative survey with an online panel and we asked three questions:

  1. Do you work with data?
  2. Are you comfortable working with data?
  3. What kind of training would you like?

This was when I was doing the data literacy work, so I thought the third question would be the most interesting, to find out how people wanted to become more data literate.

But what really fascinated me were the answers to the first question:

  • “I don’t do anything with numbers.”
  • “My team’s not responsible for calculations.”
  • “We don’t deal with statistics.”

They were clearly associating data with numbers and they didn’t recognize what data is today. And like the cashier, they probably didn’t know:

  • How their companies were using that data
  • The value data potentially brought to their company
  • Their particular role with data, whether it’s capturing, protecting, or using it

That brought home the fact that there are huge gaps in data literacy. Within the industry, we’re all in data; tech is now very data focused. We think everybody knows about data. And often, the vendors promoting data literacy focus on those who are already data literate — the business analysts, the data scientists, those who are already experts — and it’s important.

But a fundamental part of my research on data literacy was to focus not just on the top of the pyramid. I want to work with companies to create a very strong base, with more awareness of what data is within the organization.

 

How to Develop a Data Literacy Program

Satyen Sangani: How does that bridge into this sort of formalized definition of data literacy? When you go into an organization and say, “You need to be more data literate,” what do you do?

Jennifer Belissent: There are a couple of ways of approaching it:

  • Through formal data literacy programs that teach people about data
  • Demonstrating what you can do with data and getting people excited about how you can use data to solve problems — as if to solve a mystery

There are a couple of stories I uncovered during my smart city research. Several years ago, Buenos Aires undertook a massive streetlight modernization project that replaced traditional lighting with LED lighting, which is supposed to be brighter and can improve public safety.

But when the lights were in place, they started getting complaints that the streets were dark! For this mystery, as was the case with the Sodexo cafeteria, they dug into the data.

By overlaying different data sets, they found out the complaints were happening during the summer months, when the leaves were still on the trees. LED lighting doesn’t diffuse — that’s one of the advantages; it remains concentrated and doesn’t get lost in the ether — and that’s supposed to be a good thing. But when leaves are on the branches, the LED lighting doesn’t “go around” them.

The data indicated the answer: The city changed its tree-pruning schedule to ensure there were no leaves blocking the lights in those summer months.

That shows how we can figure something out by bringing in data — data from a variety of different sources. It’s not just one data set that we’re going to use. With data diversity we’re going to be able to solve the problem. And people tend to get more excited about and can understand data in a more concrete way when they see it applied to a specific problem.

 

Making Data Personal — And a Way to Simplify Employee Tasks

Satyen Sangani: Particularly a specific problem they’re either experiencing themselves or are trying to solve. What’s interesting about data stories is that in the abstract or when it’s somebody else’s problem, people don’t quite understand it. Data tends to be pretty inscrutable and hard to access, but when you make it personal and germane and relevant to that individual, it can be really powerful.

Jennifer Belissent: The same goes with an employees thinking they’re going to be replaced by an algorithm and concerned about their job.

Another element of data literacy is helping people understand how to use and engage with data, and understand new roles and how data is going to change their existing job.

Another smart city example is how Chicago addresses its health inspections. There’s a group of health inspectors within the city doing their job the same way for years. Then they’re told they’re going to have to do it differently. Traditionally, they rip a couple pages out of the Yellow Pages and say to some inspectors, “You’re going to do A through L” and to others, “You’re going to go through M through Z.”

But that’s not very efficient. Inspectors tend to run all over the city. You might try to organize it by neighborhood, but that’s not necessarily efficient, either.

So Chicago installed an inspection optimization system to predict the likelihood of a health violation. They brought in a wide variety of data types for restaurant neighborhoods such as:

  • Were there noise complaints?
  • What time of year were complaints filed?
  • What was the weather?
  • Were there past violations?

You can imagine all of the data they can record. Then a prioritized list is given to the inspectors to investigate. Rather than visiting 30 restaurants that didn’t need an inspection, they now have a list of priorities where a violation is much more likely.

The inspectors feel they’re doing their job better — it’s easier to do their job. Showing them how data and these new techniques are going to make their job much easier is also part of data literacy.

 

Data Evangelism from Within the Organization

Jennifer Belissent: I mentioned how I see a data literacy program as a pyramid, starting with awareness at the bottom, and then the next layer is comprehension, which is understanding what data is. This is more for the decision-makers, regarding the kinds of questions they need to ask about data:

  • Where does it come from?
  • Is it representative of the population?
  • What’s the underlying analytic logic of the model?

They’re understanding the origins of the insights of the data they’re given.

At the top of the pyramid are the experts. Of course, you need to focus on the experts, making sure they have access to the newest tools and latest data sets. But I also advocate using those experts to scale, to give back across the organization.

The model here is ACES:

  • Awareness
  • Comprehension
  • Expertise
  • Scale

You’re building data ACES within your organization, but scale is really about communicating to the rest of the organization what data is and how it’s being used.

 


 

How to Build Data ACES Within Organizations

To do this, we’re seeing things like lunch-and-learns and showcase sessions. Maybe it’s having a specific lunch-and-learn around a particular topic or setting up a booth in the cafeteria and providing incentives for people to come by and see what the data team is doing.

It could be upward mentoring: assigning someone from the data team to an executive to help understand how to:

  • Navigate the dashboards they’re getting
  • Ask the right questions about where the data came from
  • Define a data project that will solve some of their business problems

 

Data Literacy Starts with a Communications Strategy

Satyen Sangani: You can also provide resources that help employees connect data to their everyday lives.

Jennifer Belissent: I’ve seen data literacy programs create cheat sheets. At Lurie Children’s Hospital in Chicago, they had cheat sheets for the clinicians:

  • On one side, medical concepts and related questions
  • On the flip side, possible techniques, available datasets, and tools to help address problems

What they found was that sometimes those clinicians had had some sort of data training in graduate school, but they’d never really put it into practice on the job. So for them, they needed that push. They needed that connection to their actual job to see where to apply some of the tools they already had in their toolbox. These are tools they hadn’t necessarily learned how to use on the job.

I’ve been talking to a number of CDOs who have been talking about their comms strategy. I’ve always thought of it as a data literacy program, but they really think of it as communicating:

  • How do we communicate what we do?
  • How do we get people excited about what we do?
  • How do we build that broader community?

 

Overcoming Emotional Obstacles to Data Literacy

Satyen Sangani: I would even take it a step further and think of it as marketing. These devices you described sound eerily similar to the things we give our salesforce in the enterprise software industry to enable and train them.

The lunch-and-learns sound and feel so much like events where we put up booths to teach people new to our products about those products and give them a demo. That marketing and sales problem strikes me as so interesting because many data radicals — whether they’re CDOs or people implementing self-service enablement programs — don’t always realize their job isn’t just to stand up some infrastructure and make some data available.
Do you see that as an emotional challenge for people?

Jennifer Belissent: Most of the successful CDOs I talked to see that as a big part of their role. When I ask them, “What are the skills required to be a chief data officer?,” what often comes up is business and technology skills, but also communication skills.

One of my favorite CDOs called himself the chief diplomatic officer, because a big part of what he felt he did was diplomacy — being the bridge across different stakeholders. It was negotiating to get access to data or prioritizing specific projects.

 


 

The Three Pillars of Data Governance

Satyen Sangani: Do you feel that’s prevalent with CDOs today? The first version of the CDO job was about getting data governance to prove compliance and making sure you’ve got the appropriate data retention policies. It was about ensuring the boxes are checked and certain policies are satisfied. Do you feel the transition of the role — to include the evangelism side — is complete?

Jennifer Belissent: I really do. About five years ago there was all this talk about CDOs shifting from being defensive to offensive. That metaphor came from American football. However, the better metaphor of what a CDO should be — because it’s not really a shift — comes from European football (soccer). The same players are always on the field. They play defense and offense and some of the best offensive players — the strikers — are actually defenders. Look at Sergio Ramos from Spain. He’s a defender, but he has one of the highest scoring rates in the league.

So it’s not a question of being defensive or offensive. CDOs have a mandate across the data value chain, across that whole life cycle of data. Data governance also extends across that life cycle. It’s not just about security or privacy or ensuring data quality; it’s also ensuring the right people can access it and use it to deliver value to the organization.

At Snowflake, we talk about three pillars of data governance. It’s knowing your data, protecting your data, but also unlocking your data to ensure people can collaborate securely using that data.

The CDO reflects that holistic, end-to-end view of data governance across that whole spectrum. One of the things that we’ve seen, particularly in the U.S., is that the CDO is increasingly reporting to the CEO and the chief executive in an organization. That does reflect more of the strategic view of data — and of data as a strategic asset — and has the CDO really driving that effort within the organization.

 

Building a Data Culture with Quick Wins

Satyen Sangani: And that’s kind of the essence of this entire podcast: How do you build a data culture? If you’re talking to a technical CDO who maybe hasn’t done this work, where would you tell that individual to start?

Jennifer Belissent: The most important thing for a new CDO coming in is to get a quick win under the belt. I’ll refer back to this “diplomacy” effort. It requires a “listening tour” where you talk to your peers to understand their issues and priorities to find something that’s going to demonstrate the value of data.

That’s your starting point. You need something to showcase as the way data delivers value into the organization. Demonstrate that to others to get them excited, and then more projects will come in, and it snowballs.

With the issue of prioritization, you go back to your stakeholders and assemble a data governance council (or whatever you want to call it) that sets your priorities.

At the same time, as you consider the project trajectory, you also need to consider the cultural aspects:

  • You’re evangelizing to your peers, getting them excited about current projects and proposing future projects.
  • You’re also segmenting your employee base — as you would a customer base for a marketing campaign — and pitch the idea of the value of data.

You’re showcasing with lunch-and-learns, training sessions, office hours for people who want to drop in to discuss ideas. Over time, you’re creating the cheat sheets, maybe launching a blog, and delivering content to address the base of the pyramid, at the awareness level. At the same time, you’re working with decision makers.

It’s not something that can happen all at once; it needs to happen over time. But in the first instance, it’s creating some of these quick wins so you can demonstrate value to people.

 


 

Build the Data Infrastructure as You Execute Data Projects

Satyen Sangani: I love the idea of a quick win, and I think it applies not only to the CDO or senior roles, but on some level to every data professional. Organizing and governing data — and making it available and describing it — can just seem like a daunting, painful, inaccessible task. And often you just want to lock down and focus on a data glossary for the next 12 months, and you can get so lost in those efforts.
I find that the people who are really successful at almost any level just say, “Look, I’m going to focus on one or two things and get that right.” And then do that cycle you’ve described.

Jennifer Belissent: One of the CDOs I spoke to described how he got his infrastructure in place. He started with a few projects, taking the data required for those projects and populating his data warehouse and data catalog. Then he added new projects, for which there was more data to put into the warehouse and the catalog.

Eventually, an enterprise data warehouse and catalog without saying to everyone, “Stop! We’re going to not do anything until we build out the data warehouse and data catalog.”

His big reveal was that over the course of the time spent working on those projects, they were building an infrastructure that’s now available for everything else.

EIt was a valuable lesson in how to approach this in a more agile way, not stopping the clock and trying to build out something before delivering any value.


 

Level-setting for Data Literacy

Satyen Sangani: Let’s take a step back for a moment.
If you want to build a data-literate organization, then you need to meet your employees where they are.
Jennifer believes that inclusion is a critical component of achieving this.

Jennifer Belissent: Although data literacy is often defined as the ability to read, write, and communicate with data, I like to add to recognize because there are people who don’t recognize what data is within the organization. They don’t realize they work with it. Therefore, a data literacy program needs to start at “level zero” to address “What is data?”

This helps people understand that everything within the organization is data-generating. Then you identify different personas within the organization, assessing where people are in their understanding of and ability to use data.

One of the things I found in the qualitative research I did was that many people are not comfortable with data. We’ve created haves and have-nots within an organization. The answers I received included “I judge myself” or “I feel bad about myself because I don’t understand when people start to talk about data.” These were questions of self-worth — and it really disturbed me that people felt like that — so we need to make sure data literacy programs are inclusive.

We also need to focus very heavily on the decision-makers and those whose jobs are really impacted by the use of data — those who are given a dashboard or some sort of data or insight and expected to make decisions on it.

At Forrester my research found that fewer than half of decisions are made based on quantitative information, as opposed to experience, opinion, or gut feeling, because we don’t really teach people how to make those decisions:

  • Evaluating data
  • Understand where the data came from
  • Knowing the underlying analytic logic

We need to focus on those decision-makers and anybody else whose job might have been disrupted by data, because they might not be out of a job, but their job is transformed and they need to understand how data plays into their new role.

And with a data literacy program, you can’t forget the experts, so give them access to Coursera or other online training courses, if available. Asking them to scale and contribute to building the data literacy program internally, and that’s a big part of their professional development: giving them speaking opportunities and helping them enhance their own skills. What’s a better way to help you understand something than having to teach it to somebody else?

So look at the different personas across your organization, understand their data maturity through some sort of assessment, and then create a program that delivers content to these various personas in the ways most suitable to them.

 


 

Building a Data Community

Satyen Sangani: I love this notion of an assessment because it sort of takes data in order to figure out where you should strategize with your data programs. And so you start by learning, which sounds obvious but it’s maybe something people don’t usually do.

Jennifer Belissent: An additional piece of advice I’d give is to build that data community. Talk about data, communicate about projects and successes, and incenting those data teams to help scale.

It also starts with onboarding. We offer sessions on business ethics and sexual harassment — often the things not to do — but we need to ensure they’re aware of data. If you want your company or city to be data-driven, everybody has to have that as a baseline for the way that they approach their job.

 

Data Literacy Requires Change Management

Satyen Sangani: And it has to be fairly real in the sense that you have these dual and competing notions of fear and change and then understanding data. You can understand data, but there are certainly cases where you can understand data and it adversely affects your role or your job, or maybe your standing politically because you’re doing and saying something a leader in a particular part of the organization doesn’t want to hear.
That’s where it can get quite complicated because then the cultural change gets tested and leaders have to put their money where their mouth is and actually drive change within the organization.
Do you find that leaders understand that this can even affect how they make decisions and maybe make them uncomfortable, too?

Jennifer Belissent: I think they do. Embracing change, as well as potential conflict, reflects the maturity of the leader. Having support from a chief data officer or other executive support, makes it easier to make internal changes.

Likewise, cultivating support from below is another way to address potential conflicts. Again, data literacy can help. I mentioned assigning an upward mentor to executives in the organization to help them understand data — and that’s as important as fostering data literacy at the grassroots level as well.


 

Translating Data Into Information

If we want to get out of the Dark Ages, the answer is simple — we need to translate data into information.

So let’s make sure we surround information with context.

Let’s work to lower the costs of using data.

Let’s reward the people and systems that provide transparency and usability.

And then — just maybe — our data Renaissance might be upon us.

This is Satyen Sangani, Co-Founder and CEO of Alation. Thank you for listening.

Other Episodes You Might Like

Start with Story, End with Data

Ashish Thusoo

Ashish Thusoo

Founder of Qubole and Creator of Apache Hive

Subscribe to the Data Radicals

Get the latest episodes delivered right to your inbox.

Marketing by