Data Con LA 2020
Description
HopSkipDrive, a startup focused on youth mobility, wanted to invest Data Culture to ultimately improve the abilities of its entire staff to quickly make data driven business decisions that created positive change. The key to achieving an effective Data Culture was treating internal data like a product, hiring a Data Product Manager to lead this initiative and create targeted solutions for specific data utilization problems. This talk by HopSkipDrive's Data Product Manager, Cindy Lin, will cover the initial job description, the steps she took once onboard including the process of creating and executing the HopSkipDrive Data Culture model and roadmap, and the outcomes she has achieved so far.
*What led to hiring this role and why was Cindy hired?
*Once onboard, retaining executive buy in
*Establishing the data culture model and roadmap
*Executive of the data culture strategy
*Impact of the above
Speaker
Cindy Lin, HopSkipDrive, Data Product Manager
5. Someone who solves DATA problems through:
Customer Empathy 💖
Creativity 🎨
An Analytical Mindset 📈
Relationship Building 🤝
The solution is still the Product 💡
6. Typical Data Problems
● Poor Data Quality
● Insufficient Data
● Lack of Data Analysis Skills
● Communicating with Data
● Data is not Utilized
● Data is Selectively Utilized
Business Intelligence and Management are familiar
with these data problems, and typically implement
a Data Culture initiative to address these
problems. But who owns a company’s Data Culture
and what does a successful one actually look like?
7. Data as a Product and the Data Product
Manager Role at HopSkipDrive
10. Strategy 1: Treat Data as a Product
• HopSkipDrive Data Problem 1: There were over 200 views in Tableau, many
being underutilized while the BI team still fielded questions already answered in
these views.
• Solution:
▪ Treat Data as a Product
10
What are the problems or barriers preventing
better adoption of existing data dashboards?
11. Strategy 2: Bring the Product to the People
• HopSkipDrive Data Problem 2:
▪ Employees have access to data but struggle to develop insights that can then
be used to make decisions.
• Solution:
▪ Build up analytical skills in each department.
Who should be trained to what level in each department?
12. Data Product Manager Evaluation Criteria
1. Train members of each team to use
Tableau
2. Curate published data sources
3. Build data products and models
4. Enhance Data Infrastructure
● Product/Project Management Skills
○ Clear oral and written communication
○ Presentation Skills
● Data Skills
○ Analytics
○ Tableau dashboard development
○ SQL
○ Data Warehousing
○ Excel/Gsheets
○ GCP
○ PostgreSql
○ BigQuery
○ Data Prep or Alteryx or other ETL tool
● Programming Skills
○ Linux
○ Python or other scripting language
Key Outcomes
Competencies
Accelerate adoption of BI self-service enabling
the entire staff to take action, at greater scale,
while maintaining excellent service.
Mission
13. Who did they hire?!
Education:
● Civil Engineer B.S. from WPI
Past Experiences:
● Process Engineer and Business Analyst experience at
a Fortune 500 manufacturing company.
● Business Intelligence Analyst at a social media
startup.
● Product Manager at a research and advisory firm.
● Product and Analytics Consultant for a gaming studio.
Looking for: A product manager role at a mission driven
technology company.Cindy Lin (that’s me!)
15. Interview Goals:
● Meet the team
● Establish rapport and build
trust
● Understand key job
responsibilities and how data
fits into their work lives
● Identify problems.
Start Customer Interviews
Interviewed 24(!) People
16. Unnecessary time
spent looking for
answers
Multiple steps required
to answer one question.
Teams rebuilds a report
that already exists and
used by another team.
Discussion spent
clarifying data
Actionable data is
underutilized
Pain Points
Inconsistent terminology
across teams.
Multiple data sources
not integrated so data
does not match between
sources.
Using data is tedious so
deeper exploration
becomes difficult.
Accurate deep insight
can only be generated by
a few people.
18. Trust - Data is trusted
Process - Framework
to process data
Transparency -
Metrics and insights
shared regularly from top
Tools - Data available and easily
understood with tools designed for
teams.
Community - Fellow team
members available to support
Empowerment -
Individuals can ask and
answer their own questions
Talent - Data talent is
intentionally recruited,
developed, and retained
Mindset - Data used to
consistently challenge
assumptions
Metrics - Co. wide
agreement on metrics
Data Culture Model
EmployeeCompanyTeam
19. Initial Roadmap (Prepared Nov 2019)
Dec 2019 Jan 2019 Remainder of Q1 Q2
Company Align with Exec Team
Define data request
process
‘19 Q4 Data Survey
Start Data Training
Identify and scope major
Q1 data projects
‘20 Q1 Data Survey
Complete Company
Dashboard
Conduct all 4 Data
Trainings
‘20 Q2 Data
Survey
Team Pilot Team Dashboard
Determine roll out order
to teams
Start scoping team
dashboards 2 and 3
Identify Data Champions
Complete all team
dashboards
Revise team
dashboards based
on findings
Employee Develop Level 1 training Develop Level 2 and 3
trainings
Develop Level 4 training
20. The Metrics that Measure:
● Individual:
○ % of individuals trained by training level
● Team:
○ Weekly views of Dashboards / User
● Company: Survey
○ Please rate the quality of data shared on company and team
performance.
■ 1 (Very low quality and cannot be trusted) to 5 (Very high quality and
is completely trustworthy)
○ Please rate the value of data shared on company and team
performance to your job.
■ 1 (Not valuable at all for my job) to 5 (Integral for doing my job)
22. Four Level of Data Trainings
Data Evangelist
Train team analysts to build their
own data sources
04
Target Audience: Advanced Team Analysts
● Reviews SQL and database basics.
Data Evaluator
Train team analysts to build their
own reports
03
Target Audience: Team Analysts
● Reviews Tableau software & published data
sources.
Data Explorer
Share details on HopSkipDrive
specific data
02
Target Audience: Tableau Users
● Reviews what data is captured, where it is
found in Tableau, and the what and why for
company metrics.
Data Educated
Establish baseline for data literacy
01
Target Audience: All
● Reviews data types (qual vs. quant), examples
of data usage, data visualizations, and data
principles.
23. Team Dashboards
Treat each of these team dashboards as
a data product for that team.
Goal of team dashboards
● Develop a tool specialized for
teams’ focus areas.
● Minimize the time spent looking
for answers and make it clear
what to do next.
● Continuously iterate (a.k.a. be
clear about scope)
25. Data Gym
First 30 Minutes:
A review of a data topic
publicized a few days before.
Ex. Data requests, completed BI
analyses, models completed by
other teams, videos from experts
Second 30 Minutes:
Available time to answer any
individual data questions.
Ex. specific Tableau issues, excel or
sheets formulas, where to find data
One hour weekly session available to every employee
Dedicated Slack Channel - #b-i-secrets
27. Data Trainings Success
Data Evangelist
Target Audience: Advanced Team
Analysts
04 N/A - not yet launched
Data Evaluator
Target Audience: Team Analysts
03 77% of Target Audience Trained
Data Explorer
Target Audience: Tableau Users
02 40% of Target Audience Trained
Data Educated
Target Audience: All
01 53% of Target Audience Trained
28. Seven Team Dashboards Launched!
Those deployed prior to the pandemic were of little use and new ones were not
prioritized over other data projects focused on continuing operations with the
smaller team.
Recently revamping dashboards as schools begin to reopen.
29. Q2 2020 - Data Culture Survey Results
• Improvement in employees’ perception of
the quality and value of data across almost
all teams
• Biggest Change:
▪ Marketplace team reports the highest jump in
ratings.
• Status Quo / Regression:
▪ Sales rated QUALITY essentially the same and
VALUE lower compared to previous survey
30. So are all the elements of a
Product Manager there?
Problem
Customer Empathy
Creativity
Analytical Mindset
Relationship Building
Product
✅
✅
✅
✅
✅
✅
What’s next?
Customer Data Problems
31. Questions?
I want to hear how your stories about Data Culture
initiatives. Connect with me:
www.linkedin.com/in/cindy-lin-product/
https://twitter.com/cindylinpm