How to build a data-driven

culture through Collective

Data Empowerment
Awaken your hidden data workforce.
Collective Data
Empowerment is the Collective Data Empowerment is the missing ingredient from most companies’

only way to build a truly

data strategies. Sure, data executives must govern and secure their data, but if
it’s only useful to elite data practitioners, it’s not living up to its promise: to
make your organization smarter, faster, and more effective.

data-driven culture In this resource, you’ll learn:

across the enterprise. • Why you should pursue Collective Data Empowerment
over Selective Data Empowerment
• The stakes and opportunities of data-driven cultures
for data executives
• How to identify and activate your hidden data
workforce

2
You can’t build a
data-driven culture
We’re on the edge of what could be the biggest multiplier in data productivity
yet: the shift from Selective Data Empowerment to Collective Data
Empowerment. If your company navigates the shift, it will activate a hidden

without your hidden data workforce, refocus your advanced data practitioners on the highest-
impact work, and achieve a data-driven culture.

data workforce. So, what makes data-driven cultures different?

• Less-technical employees work with and benefit from data.

• Anyone can contribute domain knowledge to an analysis.
• The output of one data project benefits many other projects.
• More human talent and brainpower are available for any data objective.
• People discover, reuse, and adapt prior data work.
• Exposure to data improves data literacy.
• Network effects compound data’s impact.

Only 33% of full-time employees in the U.S. are confident in their data
literacy, according to Qlik1. The rest are members of what we at data.world
call the “hidden data workforce.”

Your hidden data workforce holds the lion’s

share of available brainpower and talent.

1. www.qlik.com/us/company/press-room/press-releases/0129-qlik-reveals-data-
3
literacy-skills-shortage-in-us
You already have
Selective Data Empowerment Collective Data Empowerment
makes advanced users more makes everyone more productive
productive with data. with data.

the people. Now

you need the plan.
Lift all boats Here, “empowerment” means the lift in data productivity resulting from a
combination of tools, practices, and strategies.
Companies that only pursue Selective Data Empowerment will never realize more
than a fraction of data’s potential because Selective Data Empowerment leaves
most people out.
Example A
In data and analytics, cross-functional teams If you buy better statistical modeling software, that’s Selective Data Empowerment
with a broad skill set as well as different work because only those who understand statistical concepts benefit directly. But if you
pair it with training that helps more employees use statistics to make decisions,
and behavior styles will lead to more innovative the statistical models become useful to more people. Now you’re in Collective Data
thinking and better outcomes. Empowerment territory.
—Rita Sallam, Distinguished VP Analyst, Gartner2

Companies that evolve into Collective Data Empowerment quickly enjoy the Example B
profound business impact of inclusion and connectivity. These companies will
become data-driven cultures faster than their competitors. Most companies know they need an inclusive data catalog, built for broad use across
teams, salary bands, and a wide spectrum of data literacy. But if they evaluate data
This is a shift that builds on the progress you’ve already made. It’s the rising tide catalogs without including those who have the business problems that need solving,
that lifts all boats. they’re veering toward Selective Data Empowerment. To help data.world prospects
find the path to Collective Data Empowerment, we ask them to bring into the sales
This is a shift led by those who aren’t satisfied with a tiny slice of data’s process not just the people who will help implement and integrate the data catalog,
potential. Is that you? but also business stakeholders who can be the authoritative voice on the business
problems.

2. https://www.gartner.com/smarterwithgartner/boost-innovation-with-diverse-data-
and-analytics-teams/ 4
The high stakes and
staggering opportunity of
data-driven culture
The world looks different from an airplane window. You notice things you can’t see from
the ground: the way power lines cut across remote terrain, how rivers intersect and shape
communities, the tidy grids that characterize newer cities. Features become patterns. People are
invisible from 33,000 feet.

Data helps you soar above the limitations of subjective experience. Up here, you see farther. You
spot connections and anomalies among millions of rows. Data is altitude. Altitude gives you room
to observe the trends, to measure the effects, and to explain the spikes. Data makes sense of the
world through abstraction. The people it represents disappear into aggregations and averages.
Have you ever met a statistically average family of 3.14 people? Neither have we.

When companies win, when companies fail, people feel it. Profit, productivity, shareholder value,
EBITDA, and every other 33,000-foot metric only matters because it affects people. The same is
true of data-driven culture.

Achieving data-driven culture can be the difference between hiring amazing people and laying
them off, between earning loyal customers and losing them forever, between getting the best from
people and burning them out, between teams that collaborate and teams that clash.

For data executives and their allies, it’s a career-defining moment.

This is your chance to grab the controls and rocket your company and career to new heights.
There’s no return capsule. This is it.

Will it be easy? Not exactly. But it’s much easier than surviving in a graveyard of failed data
initiatives, bad decisions, and small thinking.

5
Fear, uncertainty, and doubt
The data leader is no stranger to fear and failure. Only 28% of C-level execs
feel the CDO role is successful and established3. In 2020, only 38% of C-
level execs think they have created a data-driven organization, up only one
percentage point from 37% in 20173.

According to Gartner analyst Nick Heudecker, 85% of Big Data projects

fail5. The most common causes are poor integration with
business processes and apps, management resistance, internal
politics, skills shortages, and security and governance challenges.

Just 9% of C-level execs think the principal challenge to becoming data-

driven is technology. The other 91% think the principal challenges are
related to people, business process, and culture3. Fifty-six percent of CEOs
have doubts about the data that feeds their decisions4.

Source: NewVantage Partners, "Big Data and AI Executive Survey 2020."

3. http://newvantage.com/wp-content/uploads/2020/01/NewVantage-
Partners-Big-Data-and-AI-Executive-Survey-2020-1.pdf
4. https://home.kpmg.com/xx/en/home/insights/2017/08/trusted-
analytics-matter-to-ceos.html
5. https://twitter.com/nheudecker/status/928720268662530048

6
Hope and opportunity
60% of top-10 companies by market cap
With FUD like that, the upside better be irresistible — and it
is. Data-driven cultures experience several major financial
were data-driven as of April 2017
Data�
Driven Company
benefits, including a 20%-30% EBITDA improvement over
peers6 and 15% growth in revenue and operating margin7.
Rank April 2017 Q4 2011 Q4 2006
In 2006, only 1 of the top-10 companies by market 1 Apple: $741B Exxon Mobil: $406B Exxon Mobil: $447B
capitalization was data-driven. In 2017, data-driven
companies took 60% of the list8. 2 Alphabet: $585B Apple: $376B General Electric: $384B

3 Microsoft: $505B PetroChina: $277B Microsoft: $294B

If you believe, as we do…
4 Amazon: $432B Royal Dutch Shell: $237B Citigroup: $274B

…that most companies are nowhere close to realizing more 5 Facebook: $408B ICBC: $228B Gazprom: $271B
than a fraction of data’s value, and…
6 Berkshire Hathaway: $404B Microsoft: $218B ICBC: $255B

…that your hidden data workforce holds more promise 7 Exxon Mobil: $344B IBM: $217B Toyota: $241B
than any combination of elite talent and technology, and…
8 Johnson & Johnson: $330B Chevron: $212B Bank of America: $240B

…that data-driven cultures need leaders like you, then… 9 JPMorgan Chase: $303B Walmart: $205B Royal Dutch Shell: $226B

…the opportunity is staggering. 10 Alibaba Group: $278B China Mobile: $196B BP: $219B

Source: S&P Capital IQ, “Top 10 Companies with Highest Market Capitlization Worldwide.”
Inspired yet? Not so fast. You have to take stock of your
Note: Market capitalization figures have been rounded and are in $billions
company’s current situation to know what you need to do
to start awakening your hidden data workforce. Next, the
signs you’ll need to watch for to know what to address first.

6. https://www.bcg.com/d/press/23may2017-data-driven-transformation-160648
7. https://i.forbesimg.com/forbesinsights/ey_data_analytics_2017/EY_Data_Analytics_Report.pdf 7
8. https://www.bcg.com/en-us/publications/2017/digital-transformation-transformation-data-driven-transformation.aspx
16 patterns you see
in pre-data-driven
Think about a challenge your company faces. Who among your
colleagues knows the most about it? Who has the most experience

companies
with it? Who feels its weight every day? They probably aren’t all
analysts and data scientists.

Now recall a data analysis that was focused on the challenge. Were
the people you just pictured in the room, or on the thread, from the
beginning? Were they consulted and considered throughout? If not,
you might have a homogeneity problem (pattern #5).

These are 16 patterns we see in companies that are struggling to

evolve into data-driven cultures. They share a root cause:

Most companies have invested more in Selective Data Empowerment

than Collective Data Empowerment.

What do we mean? Selective Data Empowerment makes

advanced users more productive with data while Collective Data
Empowerment makes everyone more productive with data.

On to the patterns…

8
1. Unused data
People don’t know what data your company has, and can’t find data when they
need it. How can you use something if you don’t know it exists?

2. Debilitating delays
Data and analysis arrives too late to capitalize on opportunities. People struggle
to understand it, and their questions aren’t answered fast enough.

3. Wasted resources 5. Homogeneity

High-paid data people spend most of the day making data more usable, answering Project outcomes reflect the knowledge, abilities, and biases of a small,
the same questions, and refreshing reports. They can’t focus on the advanced homogenous group. These advanced practitioners convey insights that
tasks that only they can do. seem cryptic to most of their colleagues.

Data scientists, according to interviews and expert Very few companies expect only professional
estimates, spend from 50 percent to 80 percent of writers to know how to write. So why ask only
their time mired in this more mundane labor of professional data scientists to understand and
collecting and preparing unruly digital data, before analyze data, at least at a basic level?
—Jonathan Cornelissen, Harvard Business Review10
it can be explored for useful nuggets.
—Steve Lohr, The New York Times9

6. Needless repetition
4. Lost knowledge When people can’t see what others are working on, they waste
time and money on duplicative work.
Teams lose the knowledge they create if it falls outside of narrowly-defined
project goals. If it doesn’t have immediate value, and there’s no obvious home
for it, it falls through the cracks.

9. https://www.nytimes.com/2014/08/18/technology/for-big-data-scientists-hurdle-to-insights-is-janitor-work.html 9
10. https://hbr.org/2018/07/the-democratization-of-data-science
7. Puzzling barriers
Access controls and other restrictions persist without regular reconsideration These are the know-it-alls who, almost on
and changes. It’s the classic battle between legacy and logic.
principle, ignore the data, ignore the evidence and
8. Lethargic data literacy recommendations, and do what they want because
Data and insights do little to elevate data literacy if people can’t trace the path
they know best (after all, their paycheck proves it!).
from question to answer. Fancy tools can boost the productivity of data elites, but —Carl Anderson, oreilly.com12
they don’t lift the general baseline by themselves.
11. Reproducing nothing
9. Competing definitions The reproducibility crisis13 in the scientific community should make
You are entitled to your own opinions, but not your own facts. What about your businesses interrogate their own data practices. When people can’t
own definitions? When Team A defines something one way, and Team B defines it reproduce an analysis, they can’t check it for accuracy or reuse it.
another way, an innocent disconnect can cause serious problems.
12. Leaps of faith
10. Dominating HiPPOs People make choices based on insufficient, inaccurate, or outdated data and
Data-driven cultures do not put the Highest Paid Person’s Opinions above the analysis — when they use data at all.
data. Carl Anderson, who wrote the book11 on data-driven cultures, describes
the HiPPO’s destructive tendencies: 13. Dark data
Too much data work and analysis takes place without transparency,
audibility, or accountability. Signs include rogue databases, unsanctioned
software, and too many emailed spreadsheets.

11. http://shop.oreilly.com/product/0636920035848.do?intcmp=il-data-books-videos-product-na_20151001_radar_carl_anderson_post_being_data_driven_top_cta
12. https://www.oreilly.com/ideas/being-data-driven-its-all-about-the-culture 10
13. https://fivethirtyeight.com/features/failure-is-moving-science-forward/
 14. Separate lanes for data, analysis, and context
Getting the full picture of an analysis or dataset requires looking for puzzle pieces
in different places. The context people need so they can use data gets stuck within
email threads, one-off DMs, and hallway conversations.

15. Analysis paralysis

People wait for the “perfect” data before deciding, even though the data they can
access should be enough. A mix of factors drives this behavior, ranging from low
trust to the steep learning curves typical of data tools.

16. Low trust

It’s hard to trust data if you don’t know things like who has touched it, where it
comes from, and how it has changed over time. No wonder 56 percent of CEOs
worry about the integrity of the data they use for decision-making14.

Now that you know the patterns that emerge in companies that fail to become data-
driven, how do you unlock the benefits of a truly data-driven culture? You activate
your hidden data workforce. Help them find their way in the next chapter.

14. https://home.kpmg.com/xx/en/home/insights/2017/08/trusted-analytics-matter-to-ceos.html 11
Finding your way
with Collective Data
Empowerment
You are in a dark forest. Above, the dense canopy of limbs and leaves hides the
sun’s position. Your phone died miles ago. Animals cry in the distance as you strain
to listen for highway sounds that will lead you back to safety.

Before your mind thinks it, your gut feels it: you’re lost.

All you have is intuition, so you start walking in the way that feels right. You wish
you had a compass. You wish you remembered that thing about moss growing on
trees. Most of all, you wish you weren’t alone.

Working with data in an enterprise can feel like a less-terrifying version of this.
Your hidden data workforce is lost in the wilderness. The tools they need are
missing (like the compass) or unusable (like the cell phone). The knowledge and
practices they need are hidden (like the sun’s position) or forgotten (like the
moss trick). And the people who can help just aren’t there.

But, what if…

12
 A short time later, a park ranger finds you wandering. She
doesn’t just point you to the highway, she gives you a
compass and a map. And she walks with you, teaching you
navigation techniques along the way. With your new tools
Tools & education:
and knowledge, you’ll be prepared for the next hike. In
fact, you’re excited for it! You share what you learned with
Airbnb’s data portal
your friends and start hiking with them every weekend. Airbnb is building a truly data-driven culture across the
The group grows and attracts people of all skill levels and
entire company. The vision of Data University16 is one of
backgrounds. Each new member contributes something
different to the collective. You all become better hikers,
Collective Data Empowerment.
equipped, prepared, and confident. Each excursion is slightly
more challenging than the last. Before you know it, you’re
an experienced and skilled hiker surrounded by thoroughly Data University is data education for anyone at
capable peers.
Airbnb that scales by role and team.
If you’re an expert data practitioner, how can you lead your Our vision is to empower every employee to make
hidden data workforce through the wilderness? If you’re a data informed decisions.
subject matter expert with less data literacy, what will it take
—Jeff Feng, Product Lead, Data at Airbnb15
to use data more frequently, more confidently? And if you’re
a data executive with a mandate to create a data-driven
culture, what’s the quickest path out of the dark forest? Central to the program is Airbnb's homegrown data portal. It’s
purpose-built to foster collaboration between employees of all
Go for Collective Data Empowerment, the thoughtful data literacy levels and to empower them with data.
combination of tools, practices, and strategies that makes
everyone more productive with data. Here are some of
our favorite examples of Collective Data Empowerment
in the enterprise.

15. https://medium.com/airbnb-engineering/how-airbnb-democratizes-data- 13
science-with-data-university-3eccc71e073a
 Practices:
Salesforce’s socratic
diagnostic process
Earlier in the white paper, we wrote, “When companies win, when companies When we decided to build our own forecasting
fail, people feel it.” So, what does a company with a $108B market cap do
when it misses a sales target? It digs into the data and the culture. We can
technology, it started with domain understanding.
learn a lot from how Salesforce took a rigorous and inclusive approach to By that, I mean understanding the sales realm, sales
understanding the context, asked laser-targeted questions about the problem,
pipeline, teams, structures, and processes. We sat down
and designed a process and solution around the people with the most
knowledge and the most skin in the game. with sales leaders, reps, and all of the constituents to dig
into the sales pipeline as a framework.
The hidden data workforce — in this case, the sales team — was involved at
—Robin Glinton, VP Data Science Applications at Salesforce16
every step, from finding the heart of the problem to creating and improving
the solution. The resulting tool worked so well, it’s now a product you can buy
from Salesforce. You might not need an AI-powered sales forecasting tool, but
if you want to create a data-driven culture fast, you do need the Collective
Data Empowerment thinking that helped this Salesforce initiative succeed.

16. https://www.salesforce.com/quotable/articles/what-salesforce-learned-about-
forecasting-after-missing-internal-sales-goals/?d=7010M000001yEPO
14
Strategies:
AP’s data journalism program
We thought if we could put this data into the hands
Effective data sharing goes far beyond simply providing access and sending
out spreadsheets. The Collective Data Empowerment approach to data of our members and customers, they could find stories
sharing helps people across the data literacy spectrum understand where in the data we’d never see.
the data comes from, what it means, what they can do with it, and other
—Troy Thibodeaux, Editor, Interactive Newsroom Technology at AP
context to improve understanding and usability.

Knowing this, The Associated Press collects, prepares, and packages the
data so its members can easily discover and report local stories. Every
dataset ships with ready-to-fire SQL queries and simple instructions so AP uses data.world to make data distribution
members can easily segment the data for their local news markets. faster, more efficient, and transparent. Do you struggle to
share valuable data with other teams? Download this case
Through AP’s data journalism program17, journalists from 300+ study and get inspired!
organizations have gained access to a user-friendly source of vetted,
story-rich data to power local news.

Now that you know what you can do through Collective Data
Empowerment, how do you get started? Find out in the next chapter.

17. https://page.data.world/data.world-case-study-associated-press 15
Three ways to find out
what your data people
aren’t telling you
So, you’re a data executive. You’re on a mission to build a data-driven culture. You
spend your days leading organizational change, building relationships, choosing
vendors, and other “senior management things.”

But how often are you actually doing data work? Or closely watching individual
data projects as they unfold?

The best data leaders want to see how the sausage gets made. You’ll know what’s
broken, what’s going well, and what to do about it.

Here are three ways to get the truth.

16
 1. Do data project ride-alongs
Our research has revealed that typical data projects contain seven stages (see image Do enough of these and you’ll have a good sense where things are flowing,
below). In a data project ride-along, check in with stakeholders at each stage and observe slowing, working, and breaking. You’ll feel the friction and hear some ideas
how their tools, objectives, needs, collaborators, and challenges change along the way. worth investigating. Consolidate your findings, share them widely, and discuss
them openly. Make sure to involve less-technical people, too, or it’s not
Here are some questions we find helpful: Collective Data Empowerment!

• Who is involved at this stage? What are their roles? Are they data practitioners,
subject matter experts, or…?
• Is anyone missing from this stage? Who? Would adding their perspective or skill
improve the project? Which aspect (e.g., speed, reproducibility, etc.) How?
• What would make this stage more efficient? What are three ideas to consider?

Taken from Practices for Better Data Teamwork courseware.

17
 C) Make them blameless, like Google does
2. Conduct data For our own post-mortems, we use a slightly adapted variant of Google's post-
mortem template19. Use our version as is, adjust it to fit your team, or adapt to each

project post-mortems individual project. Either way, you’ll want to make sure you capture lessons around
these main elements:

Many companies do post-mortems to squeeze more knowledge out of every

project and apply lessons learned to the next project. You don’t have to lead
Project overview
every data project post-mortem. Nominate someone to be your data practices
This should have your goals, objectives, and success criteria listed for the
champion, teach them the steps, and pass the torch.
project. Don’t forget to log whether the project was successful or not!

Here are some things to keep in mind for your first data project post-mortem.
Accomplishments
What went well? Which outputs are most useful beyond the project? Name the
A) Make your people part of the process top three project highlights. Be specific here. You want to document and share
Project participants should share what they learned with others and listen what great data work looks like with everyone else, right?
respectfully to their collaborators.
Improvement areas
B) Do it within a week of completing the project Now, what went wrong? How can we avoid or address those issues moving
Speed is essential to improving data practices. The memories are fresh, and forward? Do we know the root causes of any project delays? What are the
participants can immediately apply what they learn. immediate and long-term effects of these issues. Specificity is important here, too.

Lessons learned
What can we carry forward in our own work that will help us? How can other parts
These post-mortems are blameless because we of the company learn from our successes and failures? Document and share.
assume everyone makes mistakes from time to time.
Post-mortems aren’t criminal investigations, they’re Action items
What are some outstanding action items remaining from the project? Do they
an affirmative process designed to make us all a have clear owners assigned to them? What does the follow up plan look like?
little smarter. Cross all t’s and dot all i’s.

—Ken Norton, Partner at GV18

However you archive and distribute these lessons learned is up to you, but
make sure they’re accessible, easy to understand, and actionable.

18. https://www.kennorton.com/newsletter/2016-04-20-bringing-the-donuts.html
19. https://landing.google.com/sre/sre-book/chapters/postmortem/
18
3. Read and sign the
Manifesto for Data
Practices with your teams
After you’ve done some data project ride-alongs and post-mortems,
you will know much more about your company’s actual data practices.
Specifically, what to improve and what makes them great. The last piece
to bring everything together is a commitment to doing things better.
Enter the Manifesto for Data Practices.

The Manifesto for Data Practices is a set of four values and twelve
principles that we believe describe the most effective, ethical, and
modern approach to data teamwork. Leaders from diverse backgrounds
including academia, business, journalism, open source, and the public
sector co-authored the Manifesto to build a shared understanding of
modern data teamwork.

Consider taking the Manifesto as is or adapting it for your needs. When

asking others in your company to join you, remember this. The most
important thing is a public declaration of your commitment to improving
data practices. When you do so, others will join in solidarity. If your
organization signs the manifesto, you’ll be joining 2,000+ others who
already made the same commitment. Read it, sign it, and share it here.

Our team has also designed a repeatable workshop based on the

There and back again
Manifesto, called Practices for Better Data Teamwork. Feel free to
When you’ve done these activities, share them to empower your colleagues to empower
take the slides and run it yourself, or let us know if you’re interested in
others. Adapt them. Reinvent them. Whatever you do, don’t let the momentum die.
collaborating on a custom workshop for your company or organization:
[email protected].
Ready for more? Check out the Data Teamwork Toolkit next!

19
The Data Teamwork Toolkit
This collection of tools helps businesses cultivate collaborative, connected data practices
across the entire organization. Give them a try!

20
Survey Template:
Data Tech Stack Audit
Everyone works with data differently: each team and role has a set of tools to best Encourage inclusiveness.
address their specific goals and data needs. If you’re tasked with bringing your This audit is most effective when individual contributors and their
organization together to find, share, and use data more collaboratively, one of your first managers fill it out. Their teams may be using free tools they (as
steps is to first understand how everyone works with data right now. managers) don’t know about. We made it as short as possible to
maximize response rate.
This survey form is a template we’ve created to help. To use this in your own
organization, copy the questions into your survey tools of choice, make any edits you
need to fit your business structure, and share with stakeholders on each team.
Encourage openness.
It’s important that you don’t use this to threaten potential rule-breakers
with reprimands. Save that for another time. All you’ll achieve with that
is to push unknown tools and processes even further into the shadows.
Your goal is to learn what people are doing, why they’ve adopted the
tools they have, and how you can help bring everyone together in a
secure, scalable way.

Encourage collaboration.
Your IT manager, CTO, or CIO may find significant value in the survey
responses. They’re great partners in getting your organization to work
together on a common, secure, and trusted set of data tools.

21
21
SURVEY TEMPLATE: DATA TECH STACK AUDIT

Step 1: Tell us about yourself.

1. What’s your name? (First and last) 4. Who else works on data with you?
a. Just others on my team
2. In which part of the business do you work? b. Other teams in the business
a. Sales c. People outside our business, such as clients
b. Marketing d. Contracted agencies or strategic partners
c. Finance e. Other:
d. Customer Success
e. Managed Services
f. Product/Engineering
g. IT If you said you work on data with other teams or with people outside the
h. Operations business, please specify below. [short answer format]
i. Other:

3. Which data tasks do you participate in? (Select all that apply)
a. Manage security and/or governance of data sources
b. Manage data infrastructure
c. Generating or extracting data from another source
d. Cleaning and preparing data for analysis
e. Performing detailed, custom analysis
f. Generating reports based on pre-existing queries or configuration
g. Use reports generated by other
h. Other:

22
SURVEY TEMPLATE: DATA TECH STACK AUDIT

Step 2: Which data tools

help you get your job done?
This will help us understand what’s essential to keeping our business going. 3. What tools do you use to analyze and understand data?
Please share the names of any tools you use, whether it’s an app, paid for by the
company or a freebie used on your own, or even if it’s as simple as email. 4. What tools do you use to share the results of your data work with others?
Remember, when we’re asking about data, we mean all sorts of things: structured
and unstructured content, customer feedback and quotes, charts and dashboards 5. Are there additional tools you wish our business made available to you?
that let you keep a pulse on the business, and more. Data doesn’t just include a. Yes
numbers and spreadsheets! b. No

6. If you said “yes,” which tools? If you’re not sure exactly which tools would be
1. Where do you store your data?
best for the job, share more about where you need help.
a. On my hard drive
b. In a cloud-based folder (like Box, DropBox, Google Drive) that I haven’t
7. Which data-related skills help you most in your day-to-day work?
shared with my team
a. Project or task management
c. In a cloud-based folder (like Box, DropBox, Google Drive) that’s shared
b. Data management and ETL
with my team
c. Infrastructure management
d. Our company’s internal data warehouse or lake
d. Fluency in SQL or other language
e. Email and/or chat archives
e. Dashboarding, visualization, and data interpretation
f. Company CRM platform
f. Deep understanding of our customer, market, product, or other specific
g. Other:
area of expertise
g. Being able to talk directly with the other people involved in a data
Note: the following questions should be freeform rather than multiple choice or
project, whether that’s the person asking the question or the people
other defined lists. Give respondents the opportunity to share things you might
doing the deep analytics or anything in between
not know about.
h. Other:

2. What tools do you use to manage data storage and access?

8. Anything else you’d like to share about the way you work with data? (For
example, specific challenges you face in getting your work done well or easily,
opportunities for our business to improve, etc.)

23
SURVEY TEMPLATE: DATA TECH STACK AUDIT

Ready to follow up?

Once you’ve built your survey and are ready to deploy, here are few best practices to help it be successful.

Be appreciative.
Thank your respondents: their time and responses are appreciated. Most survey tools will offer to
automatically follow up with a thank you email, but you might also consider rewarding the teams with
greatest participation.

Be transparent.
If you can, communicate your findings from the survey and plan to share how you’re using those insights.
This is incredibly important: people are more likely to share honest information if they know how it’ll
make a positive difference in their day-to-day work.

Be vocal.
Setting out to cultivate a connected, collaborative data-driven culture requires buy-in from every single
part of the business. Think about it as a marketing campaign: share your vision, the value, and the
successes you see along the way. Educating everyone on the importance of this initiative helps ensure
success. If you’d like help getting started on this front, data.world can help.

24
Modern Data
Project Checklist Data projects can get messy, fast. That’s why we created this checklist
to break them down into a clear, repeatable process. Try it on your
next project. See how it clarifies questions, gets everyone on the same
page, helps them find and use more valuable knowledge, and delivers
a better analysis.

This sequence is modeled on the practices of the world’s most

accomplished data teams. But every company is different, and no two
data projects are exactly alike. As you work through the checklist,
adapt it to your needs.

Hint: copy this checklist into your favorite task management tool to
manage the workflow.

If you’d like extra detail on any step or suggestion, we’re happy to

help. We’d also love your feedback! Did this help you as much as it has
helped so many other businesses on data.world? How did you tweak it
to match your unique team? Let us know: [email protected]

25
MODERN DATA PROJECT CHECKLIST

Plan
Record:
Write down the question or hypothesis.

Review the context: Identify the stakeholders:

‰‰ I know HOW this relates back to a business goal.
‰‰ I know WHY the question was asked.
Role Contribution Name
‰‰ I know WHO will use the analysis.
‰‰ I know WHAT they will do with the analysis.
Requester(s) Asks the question
‰‰ I know WHEN they need the analysis.

Subject Matter Expert(s) Understands the market,

Clarify the question: product, or topic involved;

not necessarily technical
Shared definitions and clear questions reduce friction and
back-and-forth between stakeholders, increase the number
Data consumer(s) Makes decisions from the
of people who can benefit and contribute to the analysis,
analysis produced; often the
and align results to expectations.
same as the requester

‰‰ It is only ONE question (if not, separate them).

Data steward(s) Manages the data needed,
‰‰ There is little chance that another stakeholder will
or knows where to find it
misinterpret the question.
‰‰ All ambiguous terms are defined.
Data practitioner(s) Answers the question using
‰‰ There are no acronyms.
data and related context
‰‰ It is concise.
‰‰ It can be answered with available or obtainable data.

26
MODERN DATA PROJECT CHECKLIST

Getting Started
Build your bank of resources:
‰‰ Determine what data you need, who owns it (data stewards), and where it currently lives.
‰‰ Organize your data by grouping related files into datasets.
‰‰ Place datasets into a shared environment for access, analysis, and discussion.
‰‰ Schedule data updates if it will require refreshing over course of project.
‰‰ Review usage, security, and licensing restrictions.
‰‰ Clean and normalize the data. Use internal fact tables, ontologies, or directories to streamline this work where you can.

27
MODERN DATA PROJECT CHECKLIST

Prep Data
Make your data easy to understand:
‰‰ Add context to improve usability and discovery.
‰‰ Briefly describe each file to help others understand the inputs.
‰‰ Create a data dictionary that defines columns in your datasets.

What question are your trying to answer?

Who are the stakeholders? (Use the project stakeholders table from
earlier in this resource)

When will the project be completed?

How can people use the analysis?

What restrictions apply to this analysis and/or data?

What related materials or projects should people reference to

understand this project?

Will this project and/or analysis be updated, and on what schedule?

Who should they contact if they’d like to contribute?

28
MODERN DATA PROJECT CHECKLIST

Analyze Check in:

Dig in: Build in time to gut-check your project with your original requester and
Subject Matter Experts. They can ensure your work is still aligned with
their initial objective, resolve any confusion, and spot early insights your
Empower your team to work with the data and share what they learn as they go.
practitioners may be unaware of.

‰‰ Connect the right tools to your common data workspace. Use the Data
Try these 10 prompts to guide your check-in session:
Tech Stack Audit template if you’re not sure what tools your team needs
to integrate.)
1. Here’s what we’ve learned so far, and why we think it’s relevant.
‰‰ Create, save, and share queries so people can build on the work of their
(Useful if you’ve been saving insights as you go!)
peers.
2. Does this help answer your question?
‰‰ Share scripts created through R, Python, Jupyter, or similar tools to
3. What’s surprising (or not) about this project so far?
document your work and ensure reproducibility.
4. Are any of these insights useful to you today? If so, how?
‰‰ Encourage people to share insights and visualizations, ask questions,
5. What would make these insights easier for you to act on?
and post ideas throughout the life of the project. (Showing your work
6. How would you summarize what we’ve learned to your team (or
increases trust in your analysis. It also lets others learn from your
manager)?
process, springboarding future projects.)
7. Do you agree or disagree with this interpretation of the data? Why?
8. Who else could use what we’ve shown you today?
9. What would they need to easily understand and act on the
information?
10. What other questions do you have now that you’ve seen our
progress?

If you’re on track, keep moving through the process. If your check-

in uncovers gaps or opportunities, go through the “Prep Data” and
“Analyze” exercises so fresh resources meet the same standards for clean,
contextualized data.

29
MODERN DATA PROJECT CHECKLIST

Deploy
Package and ship:
Once analysis is complete, make sure the right people know about it. Keep in mind that different audiences will benefit from the data project’s results
being packaged in different ways. Try prepping your final results for these three data consumer types.

Type Needs Emphasis

High-level Values anything that helps them keep a pulse ‰‰ Specific, concise conclusions and recommendations
consumer on the business; values transparency and ‰‰ Visualizations and text explanations
content that can quickly improve business ‰‰ Supporting third-party research
decisions ‰‰ How-to guidance on using the analysis
‰‰ Link to the data project

Data Helps discover and translate the story in the ‰‰ Project summary and process updates
interpreter data; needs detail and summary information ‰‰ Activity records
and/or to translate and convey to other stakeholders, ‰‰ Access to, and understanding of, data used
strategist especially clients and executives ‰‰ Visualization-ready data
‰‰ Saved queries to quickly explore raw data
‰‰ Ability to securely share project, insights, and data with teams or individuals
‰‰ Open channel to ask questions about the project
‰‰ Ability to subscribe to project activity alerts

Data analyst Checks conclusions; reuses and adapts relevant ‰‰ Access to detailed context, data, and source information
part of the project for other projects ‰‰ Option to view, clone, or create new queries to add new insights to the project
‰‰ Open channel to make suggestions on additional data, hypotheses, or other improvements

30
About data.world
data.world makes it easy for everyone—not just the “data people”—to get clear, accurate, fast answers to
any business question. Our cloud-native data catalog maps your siloed, distributed data to familiar and
consistent business concepts, creating a unified body of knowledge anyone can find, understand, and use.
data.world is an Austin-based Certified B Corporation and public benefit corporation and home to the
world’s largest collaborative open data community. Visit data.world for more information and expert
guidance.