Good Stuff Happens in 1:1 Meetings: Why you need them and how to do them well
Data quality and Agile BI Oct 2018
1.
2. Sourav Dutt, Biogen
Mark Frisch, Alteryx ACE
Oct 9, 2018
From Cleaning Dirty Data to Building Organizational Trust
at Biogen
Disclaimer: The views and opinions expressed in this presentation and any related discussion(s) are solely those of the
individual presenter(s) and do not express the views and opinions of Biogen”
3. Building trust in the data is
fundamental to driving usage
and adoption of any analytics
product
Communicating impact of data
quality is essential to getting
support
Data Quality (DQ)
Requirements needs to be well
Defined,
Understood, Monitored,
Remediated and Resolved
Alteryx can enable business
owners to take responsibility
of their data quality & quickly
adapt to changing business
need without reliance of
advanced SQL knowledge
Key Messages for this presentation
4. Number doesn’t match my
Salesforce App, , Begins
to mistrust the data in the
report
1
Country Data Steward
Data Steward reaches
out to central database
team and waits for a
response
Time spent in waiting for
a response is perceived
as credibility with data
and process
Several reasons exist for the
discrepancy – CRM entry, special
filters, master data gaps, etc.
2 3 4
5
6
Sales Rep Reviews report for his
performance.
Asked to explain the
discrepancy
Data base team gets back to Data
Stewards with findings
Data steward team explains Sales
Rep why the difference.
If Sales Rep Satisfied, case closed.
Else again start with Step 1
What are these filters?? Again Step 1 Again Step 2 Again Step 3 Again Step 4
There was a perception of large scale data quality problems,
but we never really knew how bad !
Problem
7. Monthly Management Summary ReportData Auditing Summary Report
Actionable Remediation Reports
Record Details
No. of Fails
% Pass
% Pass
Examples of Data Auditing Apps & Monthly Management Reports
8. Making an impact
✓ Raise awareness
✓ Clarify ownership
✓ Increase reporting accuracy
✓ Highlight process gaps
✓ Trigger improvements
Monitor, Remediate and Resolve
Communicate and Train to ensure Self Service
9. RESULTS
117 Alteryx modules
validating 20M+ data
points as Pass or Fail
Volume
As our source systems
mature our DQ rules are
ready for future changes
Agile
Available
Results presented in seven
data auditing reports
Automatically updated daily
Adoption
Transparency of the
monitoring process drives
corrective action and builds
trust in data based solutions
10. CALL TO ACTION – GO AFTER DATA QUALITY IT CANT BE AN
AFTERTHOUGHT?
• Identify couple of critical data asset
Define
• Call out the impact of bad data for these data assets
• Ideally financial impact preferred work prioritizationPrioritize
• Document entity relationship diagrams
Profile Data
• Build working prototypes
• Thoughtful Sprints with an emphasis towards self-relianceStart Small
• Take the time to confirm success
• Share solution outcome broadly and confirm if its usefulAlign & Validate
• Ensure End View/Table is truly “useful” and column names are not geeky
• Alteryx Server for Scheduling and DeploymentBuild, Test & Deploy
• Handover Alteryx DQ Workflow to Owner
Communicate & Train Data Owner