Building a Data Literacy Program From the Ground Up

Building a Data Literacy Program From the Ground Up

In an age of on-demand customer expectations, constant market vacillation, and rapidly evolving competition, the clarity and reassurance data provides is critical to running a successful business.

But as pressures to be data-driven build, line of business teams find themselves ill-equipped to uncover business insights without relying on their company’s data experts. A whopping 39% of business domain experts admit they’re “not totally sure” what being data-driven means, according to a recent survey report.

While business teams grow increasingly frustrated and self-conscious about their lack of data skills, data experts become overloaded and annoyed with never-ending ad hoc reporting request queues. More than three quarters (76%) of data experts say between 20 — 49% of their time is spent preparing these reports for line of business needs. This inhibits their ability to put their expertise to use to surface truly transformative business insights.

Nearly 40% of teams say it takes upwards of 2 weeks to fill an average data request.

The data skills gap does more than stoke frustration between teams — it restricts an organization’s ability to compete. Nearly 40% of teams say it takes upwards of 2 weeks to fill an average data request. Delays like these have caused 1 in 4 (25%) business experts to give up on getting an answer they needed because it simply took too long. These companies are quickly losing market share to organizations that make generating and taking action on data insights a key priority — insight-driven companies are growing 8x faster than global GDP.

Data literacy is no longer just a nice to have skill for business teams, it’s key to growth. And bridging the gap between data and business teams to improve data literacy across the organization starts at the top. Leaders of both of these groups must band together and lead the charge to break down the data language barrier. A comprehensive data literacy program prepares everyone to participate in the data conversation, surface impactful insights, and drive exponential business growth.


What Is Data Literacy?

Data literacy is the ability to read, understand, analyze, manage, and act on data. An effective data literacy program must focus on developing a latticework of technical, communication, and analytical skills and the confidence to present new discoveries with others.

Dr. Francisco Javier Calzada Prado and Dr. Miguel Angela Marzal Garcia-Quismondo, two professors of Information Science in Spain, developed a useful, universal framework for teaching data and information literacy. They divide the instruction of data literacy into five core competencies that build on each other and culminate in data competence, as shown in the pyramid below.

3 Types of Barriers to Data Literacy

Despite the growing need to make data-driven decisions, companies trying to cultivate data literacy can encounter significant resistance. Roadblocks to data literacy can be categorized in three ways:

Technical

Harnessing the full value of data requires infrastructure and tools that are accessible to all. Organizations need the capability to collect, store, analyze, and act on on data efficiently. However, many data analytics tools require specialized coding knowledge that’s prohibitive to non-technical users. Additionally, growing data compliance and security risks and requirements cause many companies to keep data assets siloed or restricted entirely.

Organizational

In most companies, the language of data is spoken only by the data team, forcing them to take on the responsibility and burden of acting as “keepers of data.” This leaves all other employees totally reliant on their time and expertise, stirring feelings of inadequacy and frustration that ultimately lead to a tense culture of division. What’s more, line of business teams often work in silos, with no single source of data truth and no mechanisms for sharing insights or building on each other’s work.

Personal

Making data-driven decisions and using data insights to influence strategy is relatively new for many non-technical employees. Introducing new tools and methods too quickly can be overwhelming and cause significant fear and anxiety. Organizations must work hard to combat any misconceptions around what data can or can’t do and build confidence in data-driven decision making for people to engage.


The 5-Step Data Literacy Program

This ebook examines each of these barriers at every step of the data competency pyramid. It provides leaders with proven tactics, tools, and techniques to effectively overcome these barriers and move their organizations forward on the path to data literacy.

Step 1: Understanding data

For people to use data effectively in their roles, they must first understand it. This includes making it clear why they should care enough to understand. Business users need to fully grasp the tangible benefits of incorporating data into their work for the program to have any chance for success.

Getting a read on what your employees already know and understand is a critical first step to building an effective data literacy program. At the same time, leadership must secure data experts’ support in educating their colleagues by conveying the positive impact true, organization-wide data literacy will have on their day-to-day jobs.

Barriers at this stage:

Technical

Better understanding data means learning the basic definitions and functionalities of data types, tools, and processes like JSON, SQL, data modeling, ETL/ELT, and data warehouses.

Organizational

Assumptions about the level of data understanding that exists in the larger org may be perpetuating knowledge gaps and fueling a lack of empathy or patience on the data team’s side.

Personal

Learning new skills often feels intimidating, overwhelming, and confusing. Without proper context and understanding, some may wrongly conclude data is irrelevant in their roles. Others may consider themselves too advanced to need training.

Key action items:

1. Conduct an assessment

Determine existing and missing knowledge and establish a baseline through an employee survey. Use this information to figure out what people want to learn and how they envision being able to use data so you can customize your program to match their needs and expectations.

2. Set clear and measurable goals

How will the company know if the program is successful? Since the purpose is to bring everyone together for the good of the business by enabling them to create and use data, try setting a “BHAG” (or big, hairy, audacious goal). This could range anywhere from each participant eventually creating and maintaining their own data dashboard to reducing the time the data team spends generating high-level reports for business teams by a certain percentage.

3. Segment your audience

Organize your employees into groups based on their role, job level, or current data understanding (as determined by your assessment). This is key to avoiding a “one size fits all” program and ensuring everyone gets the level and/or type of training they need.

4. Use the “buddy system”

Depending on the size of your audience segments, it may be necessary to break them into smaller groups of 3-5 people who help one another and discuss key topics throughout the program. Be sure each group contains at least one data team member to help provide guidance and answer questions. This also encourages data experts to feel more invested and involved in the program.

5. Build tailored training

As you begin to formulate training modules for each segment, clearly focus on:

  • Terms or phrases they will need to understand and use. A dictionary of common data terminology and acronyms helps reduce confusion.
  • A brief overview of the types of data being collected on a regular basis.
  • An explanation of how data fits into their workflow and will help them improve and contribute to the business.

Step 2: Finding and obtaining
data

With a basic understanding of data terminology, the next step is to teach teams where data lives and how to access it. As de facto “data gatekeepers, ” it’s easy to understand why 95% of data experts maintain fears and concerns over making data more accessible to the larger org — especially when business teams are lacking data literacy.

However, 52% of people say they’ve been unable to access the data they need to perform their job, and 38% have given up entirely on asking for a piece of data they needed. Companies must prioritize putting the right training, tools, and protocols in place to get everyone governed access to the real-time, highly accurate data they need to make decisions.

52%

of people say they’ve been unable to access the data they need to perform their job

Barriers at this stage:

Technical

According to IDG, the average company collects data from more than 400 sources, which means data lives across multiple systems. While many business folks have access to pre-built dashboards inside individual applications, it’s impossible for them to get the holistic picture they need to make truly impactful decisions without the help of the data team.

Organizational

Data teams are reluctant to enable access to data analytics tools due to privacy, legal, and security concerns. Data governance is approached as a roadblock rather than an enabler.

Personal

More than a quarter (29%) of domain experts say “fear of messing it up” keeps them from exploring data and using it more freely. Intimidation and lack of confidence prevent business teams from seeking the data they need.

Key action items:

1. Map out your company’s data ecosystem

Create a visual diagram of how data flows through your company’s technology stack. Be sure to list out what each of the tools does, which team owns what, and how everything is connected.

2. Store your data in the cloud

Cloud data warehouses and cloud data platforms like Snowflake make it easy to aggregate data across hundreds of sources and house it all in a centralized, secure, and governed repository, powering faster and easier discovery and queries. They scale elastically, don’t require an upfront infrastructure investment, and can manage both structured and semi-structured data.

3. Modernize your approach to data governance

As real-time data access grows increasingly critical, data teams must retrain themselves to think of data governance as a way to enable teams to obtain the data they need, rather than a means to block them. The latest BI tools support this approach by making governance work for teams, rather than making teams work around governance. For example, Sigma gives teams direct, real-time, yet governed access to data inside the cloud data warehouse by providing robust role-based permissions, generating query logs, allowing data teams to endorse particular data sets, keeping data safe inside the warehouse, and more.

4. Build a sandbox for people to explore

Assuage teams’ fears by giving them a safe space to practice accessing, identifying, and exploring data. On-demand manufacturing company Fictiv built a sandbox for business users to begin exploring data by pre-modeling joins between data sources, giving non-technical users a guided, endorsed path for exploration.

Step 3: Reading, interpreting
and evaluating data

Being able to identify and access relevant data doesn’t equate to successfully drawing actionable insights from it. Sounding out letters and recognizing words isn’t the same thing as comprehending their meaning.

Data must be read, interpreted, and evaluated within the context of specific business goals or opportunities for it to have value. Individuals must learn how to draw insights from data and then make those insights actionable in their work.

This is one of the toughest levels in the pyramid, but the juice is well worth the squeeze: Nearly half (49%) of business domain experts think they’d be 50-100% more effective at their jobs if they had a clearer understanding of data and how to use it.

49%

of business domain experts think they’d be 50-100% more effective at their jobs if they had a clearer understanding of data and how to use it.

Barriers at this stage:

Technical

Learning how to organize and interpret data isn’t easy. Analytics and BI tools with complex, unfamiliar user interfaces add to the confusion and sharpen the learning curve for business teams.

Organizational

Although they’ve traditionally been excluded from the data conversation, if data analysis and interpretation does not include domain experts, insights will be minimal. Both data and business teams must work together and combine their expertise to get the most out of data.

Personal

Without proper training and practice, people can easily misunderstand, oversimplify, or draw wrong conclusions from data. In fact, “people doing incorrect analysis and making decisions based on inaccurate insights” is data experts’ number one concern when it comes to making data more accessible to business teams.

Key action items:

1. Teach people to ask the right questions

Insightful data analysis begins with asking the right questions, and more thought needs to go into this first step than people typically assume. Take some time to remind folks what a hypothesis is, the difference between quantitative vs. qualitative data, how to avoid common logical fallacies, etc.

2. Put data in a familiar format

Seek out analytics and BI tools that have taken the consumerization of BI trend to heart and present data in a familiar and user-friendly way. For example, 85% of people use a spreadsheet in their work — even 88% of those who understand and use SQL still turn to a spreadsheet when exploring data. Analytics tools with UIs that resemble spreadsheets are a great choice for helping business teams immediately feel more comfortable doing data analysis.

3. Add business context to data models

Excluding business teams from the data conversation means leaving out those who are often closest to business processes and the data they create. This makes it harder for data teams to build analyses business experts need, and nearly impossible for business teams to do their own analyses. Bake a step into your data modeling process where data and business experts sit down together to add business context to tables and data sets with definitions, column descriptions, etc.

4. Build dashboards to visualize data

Dashboards summarize complex data visually in charts and graphs so the audience can absorb the information quickly. This enables business leaders across the enterprise to more easily interpret the data and reach accurate conclusions. Visualization helps make data actionable. Build dashboards for each team or department with relevant metrics and KPIs so they can quickly begin getting value from data.

Step 4: Managing data

As the Peter Parker Principle states, with great power comes great responsibility. As business teams become more knowledgeable about and involved in data analysis, their awareness of data management risks and responsibilities must also increase. Mismanaged data can lead to all sorts of problems ranging from inaccurate insights to security breaches, so it’s critical everyone in the organization feels comfortable doing their part.

Cloud data platforms like Snowflake eliminate most of the administrative and management demands of traditional on-premises solutions by automatically handling infrastructure, optimization, availability, data protection, upgrades, and more. This allows teams to focus on remaining DBA tasks like making data accessible and accurate. And while the vast majority of this remaining responsibility still falls on data and engineering teams, business experts must make a concerted effort to uphold data management best practices.

Barriers at this stage:

Technical

Many analytics solutions require data prep extracts prior to analysis. Furthermore, because they have limited access to and understanding of existing BI tools, many business experts export data to Excel or Google Spreadsheets in an attempt to get the answers they need. Data extracts not only pose major security and compliance threats, but they also result in stale, siloed insights.

Organizational

Most business teams’ data access is limited to department-specific tools like Salesforce, Zendesk, or HubSpot. If these tools don’t “talk to” one another, teams will have a very fragmented view of the business. Data extracts and the inability to effectively share analyses with one another further exacerbate this issue.

Personal

Forced to find workarounds to get their hands on the data they need in a timely fashion, many business experts have developed poor data hygiene habits. It’s likely many people in your organization don’t understand the true long-term impact of seemingly innocent mistakes such as duplicating data.

Key action items:

1. Minimize data extracts

Modern cloud analytics tools sit on top of your existing cloud data warehouse, using a secure connection to directly query your data warehouse. Data is never extracted, copied, or stored. Combined with permissions and restrictions set on databases directly, this single point of access method allows you to establish robust data governance, eliminate dangerous Excel or Google Sheets extracts, and keep data off local PCs. Additionally, performance is in no way sacrificed. An unlimited amount of concurrent users can access the same source of accurate data.

2. Build a data governance framework

Develop standards, definitions, and policies for how data is to be handled in your organization, and ensure they align to current best practices and regulations. A strong data governance program helps ensure your company remains compliant, information is being handled in a safe and secure way, and data stays clean and accurate.

3. Integrate applications and pre-join data sources

Many of the most beloved business applications have APIs or pre-built integrations that allow them to pass data back and forth to each other, giving teams a more complete picture of what’s happening in the business. Modern analytics and BI tools take this a step further by allowing data experts to pre-join data sources in the cloud data warehouse so business experts can easily “link” them together as needed and explore them as one data set.

4. Encourage secure sharing

Eliminating data silos and avoiding duplicate analyses requires teams actively and consistently share their data efforts and insights with one another. But this must be done in a secure and governed fashion. One way to achieve this is to schedule a weekly standing meeting where business leaders meet and discuss current data initiatives and recent findings. Another strategy is to create team workspaces within your company’s BI tool where users can securely share relevant reports and visualizations with one another.

Step 5: Creating and using data

The final step in the data literacy pyramid is teaching people how to create and leverage their analyses. This is where the rubber meets the road as business teams must leverage the skills they’ve learned in the previous 4 steps to generate actionable insights.

Despite the cumulative nature of this step, it still presents its own unique set of challenges that company leaders, data experts, and business teams must work together to overcome. The good news? Four-fifths (79%) of data experts want to collaborate more closely with their lines of business colleagues, and 64% of domain experts feel the same way about data experts.

79%

of data experts want to collaborate more closely with their lines of business colleagues

Barriers at this stage:

Technical

Despite 62% of companies recognizing self-service BI is essential in 2020, most “self-service” analytics tools require users to know SQL or other proprietary languages to conduct analyses. Without these specialized coding skills, business leaders are unable to even access the underlying data in pre-built dashboards to ask follow-up questions, let alone do independent data exploration.

Organizational

Because teams have grown accustomed to working around data roadblocks and silos in their organization, they often fail to think of data insights as shared assets. Value is lost and opportunities are overlooked when companies fail to empower teams to reuse, repurpose, and improve one another’s analyses.

Personal

With so much data at their fingertips, it’s easy for business users to fall prey to analysis paralysis — literally. Not knowing where to start is overwhelming, damages confidence, and can kill an analysis before it even begins.

Key action items:

1. Invest in visual data analysis tools

Although many BI tools have failed to deliver on their promises of data democratization, newer, true self-service solutions (like Sigma) are emerging. These tools enable business teams to visually analyze data by translating familiar Excel-like formulas into SQL on the back-end. The best part? They still allow SQL lovers to code if they prefer, uniting both business and data teams in a single BI platform.

2. Go beyond dashboard views

Dashboards can’t answer every question — in fact, they usually raise more questions. That’s why it’s important to provide teams with interactive dashboards that act as starting points for analysis, not stop signs. These dashboards enable teams to easily click into the underlying data analysis and visually drill down to answer follow-up questions in real-time.

3. Give teams a place to start

Providing business teams with clear and relevant jumping off points to start their analyses helps them feel more comfortable and accelerates time to insight for the entire org. This can be accomplished through report templates for the most common business use cases, as well as datasets with linked data sources, column descriptions, and definitions, predefined calculations etc.

4. Embrace community-driven analytics

The concept of community-driven analytics is an increasingly critical approach to business intelligence focused on empowering teams to share, amplify, and accelerate one another’s data insights. Repurposing and building on reports across teams, embedding interactive dashboards into relevant business workflows, and reusing analyses across applications are just a few examples of community-driven analytics in action. It’s important to note this requires a BI solution that takes a truly modernized approach to data governance, as discussed in step 2.

A Step-by-Step Framework

The following framework factors in the barriers and action items discussed at each level of the data literacy pyramid. Having a clear, actionable plan to share with key stakeholders and champions is critical for building organizational consensus and turning the dream of company-wide data literacy into a reality. The framework may be used as is, or leveraged as a guide and adapted to your business’ unique needs.

Technical Barriers

Organizational Barriers

Personal Barriers

Key Action Items

Step 1: Understanding Data

Technical Barriers

Poor understanding around basic definitions and functionalities of data types, tools, and processes.

Organizational Barriers

Knowledge gaps and lack of empathy or patience due to assumptions about data understanding.

Personal Barriers

Fear and confusion over learning new skills; resistance to learning; lack of context around data relevance.

Key Action Items

  1. Conduct an assessment
  2. Set clear and measurable goals
  3. Segment your audience
  4. Use the buddy system
  5. Build tailored training

Step 2: Finding and Obtaining Data

Technical Barriers

Hundreds of disconnected data sources and systems.

Organizational Barriers

Reluctance to provide data access to business teams due to governance concerns.

Personal Barriers

Fear of “messing data up” prevents business teams from seeking data out.

Key Action Items

  1. Map out your company’s data ecosystem
  2. Store your data in the cloud
  3. Modernize your approach to data governance
  4. Build a sandbox for people to explore

Step 3: Reading, Interpreting, and Evaluating Data

Technical Barriers

BI tools with complex, unfamiliar UIs create confusion and sharpen the learning curve for business users.

Organizational Barriers

The need for business and data teams to work together and combine their expertise to get the most out of data.

Personal Barriers

Risk of business teams misunderstanding, oversimplifying, or drawing incorrect conclusions from data.

Key Action Items

  1. Teach people to ask the right questions
  2. Put data in a familiar format
  3. Add business context to data models
  4. Build dashboards to visualize data

Step 4: Managing Data

Technical Barriers

Governance issues caused by required data prep extracts; business experts exporting data to spreadsheets.

Organizational Barriers

Data silos across business teams and applications cause a fragmented business view.

Personal Barriers

Poor data hygiene habits developed among business users.

Key Action Items

  1. Minimize data extracts
  2. Build a data governance framework
  3. Integrate applications and pre-join data sources
  4. Encourage secure sharing

Step 5: Creating and Using Data

Technical Barriers

Coding knowledge prohibits business users from freely exploring data and digging into dashboards.

Organizational Barriers

Failure to empower teams to reuse, repurpose, and improve each other’s analyses.

Personal Barriers

Analysis paralysis and uncertainties around where to start among business users.

Key Action Items

  1. Invest in visual data analysis tools
  2. Go beyond dashboard views
  3. Give teams a place to start
  4. Embrace community-driven analytics

Business Transformation Through Data Literacy

The massive economic growth some nations experienced throughout the 20th century is directly linked to rapid increases in literacy. Growing data literacy will have the same level of impact in the 21st century.

Data is the new language of business, and organizations that fail to make cultivating data literacy a priority are putting themselves at a competitive disadvantage. Over the next decade and beyond, the ability to make good decisions in rapidly evolving environments will determine which companies succeed or fail.

Sigma envisions an insight-driven world where people, organizations, and ultimately humanity are transformed for the better by data. As the only analytics and business intelligence platform built for the cloud data warehouse, Sigma works closely with Snowflake to empower all employees with the ability to access, analyze, and action data insights.

To learn more about how Sigma and Snowflake work together, sign up for a free 14-day Sigma trial, or request a demo.

To learn more about how Sigma and Snowflake work together

Request a Quote

About Sigma

Sigma is the first enterprise-ready cloud business intelligence and analytics (A& BI) solution designed to run natively inside cloud data warehouses (CDWs). Providing live, guided access to CDWs, Sigma maximizes their value, eliminates the need to change data models as new questions arise, and transforms A& BI into an iterative process. The Sigma Spreadsheet empowers anyone to analyze data — without code or extracts — and make insight-driven decisions quickly, freeing data experts to focus on more innovative, fulfilling projects. Sigma powers a collaborative, community-driven approach to A& BI and delivers on the self-service promise.

About Snowflake

Snowflake’s cloud data platform shatters the barriers that have prevented organizations of all sizes from unleashing the true value from their data. Thousands of customers deploy Snowflake to advance their businesses beyond what was once possible by deriving all the insights from all their data by all their business users. Snowflake equips organizations with a single, integrated platform that offers the data warehouse built for the cloud; instant, secure and governed access to their entire network of data; and a core architecture to enable many types of data workloads, including a single platform for developing modern data applications. Snowflake: Data without limits.