Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
with Lars von Sneidern
Hi! Thanks for letting me
crash the party.
What’s up
tonight?
• Who I am and what I am doing here
• What does business reality look
like?
• Foible #1: Connecting sil...
Who is this person?
• Yes, Lars is my real name
• I am the Customer Data and
Insights Lead for Connective DX
• What the he...
About Us:
• 19 years old
• 80 people strong
• 2 offices (Portland & Boston)
• 5 “Best Places to Work” awards
• Recognized ...
Experience
Strategy
Customer Data
& Insights
DX Strategy
& Roadmap
Technology Consulting
Data Consulting
Innovation
Analyt...
Select technology
partners
Select Clients
Need
Compare
Discover
Research
Purchase
Decide
Use
Get Help
Share
Personalize
PartnerEvangelize
Tribe
Search
Customer
Expe...
It starts with a framework;
Builds a process;
Establishes governance;
Links with strategy;
Is built with technology;
And m...
Known user
attributes
Voice of
customer
Observed
behavior
Known user
opinions
Known user
behaviors
Touchpoint
opinions
Cus...
Customer
Data
Platforms
The reality of
“business” data
What do we mean by business data?
Foible #1:
Connecting Silos
What are your silos?
Customer
Support
Connecting Data Silos
CRM
Web
Analytics
CMS
Email
Personalization
Cobweb
Ad
serve
Hub-and-Spoke
Customer
Data
Platform
Per...
Use case: How do we create a
single view of customer data?
Option 1:
Figure out and
use a universal
ID
Option 2:
Behavior ...
You and UUIDs
Customer
Support
This can get big in
non-relational
systems
HOLY SHIT
WHAT IS
GOOGLE DOING
Who’s in charge of
your data?
…depends on who
you ask!
Remember
Master Data
Management?
These days many
people are talking
about Governance
instead
Regardless, the idea
is having one
Source of Truth
Foible #2: Scheming
the schema
1. Do you have the right data?
2. Are you normal?
3. Can you scale?
Do I have right data?
What are you trying to accomplish?
Remember
Data Dictionaries?
Yeah, let’s do that
more.
To Normalize, or not?
IT IS A QUESTION!
Pretty Normal
Too Normal?
To Normal or to Not
Pros:
• Keeps data clean
• Is a best-practice
• It scales
Cons:
• Will cause table
bloat
• It’s exactl...
What about
Scalability?
I have seen several examples of
PostgreSQL systems that were
built to be quick and easy, but had
some major performance is...
Foible #3: Ongoing
integrity
Noting is as it has
always been
Data Drift
Three major types:
Structural:
Changes at the
source.
Ex: Fields added,
deleted, or
changed
Semantic:
The meani...
API “Jitter”
Makes your
ETL, SOL
So, what can you do
about it?
checking
Foible #4:
Making people care
Developers need
____?
Developers need
requirements.
Business people
need ____?
Business people
need insights.
The key is
Empathy
Thanks!
Four Short Foibles of Organizational Data
Four Short Foibles of Organizational Data
Four Short Foibles of Organizational Data
Four Short Foibles of Organizational Data
Upcoming SlideShare
Loading in …5
×

Four Short Foibles of Organizational Data

335 views

Published on

Anyone that works with data downstream in an organization has seen things go...wrong, while upstream managers and business leaders are being held accountable. Whether it's a failure in process, or something technically goes wrong, working with data is not always easy. What happened? How can we prevent it from happening again? What's next?

This talk, given at the Portland Data Science Group on October 27, 2016, uncovers 4 common foibles of working with organizational data.

Published in: Data & Analytics
  • Be the first to comment

  • Be the first to like this

Four Short Foibles of Organizational Data

  1. 1. with Lars von Sneidern
  2. 2. Hi! Thanks for letting me crash the party.
  3. 3. What’s up tonight? • Who I am and what I am doing here • What does business reality look like? • Foible #1: Connecting silos • Foible #2: Scheming the schema • Foible #3: Ongoing integrity • Foible #4: Making people care
  4. 4. Who is this person? • Yes, Lars is my real name • I am the Customer Data and Insights Lead for Connective DX • What the heck does that mean? • I make sailors blush • I love bulleted lists
  5. 5. About Us: • 19 years old • 80 people strong • 2 offices (Portland & Boston) • 5 “Best Places to Work” awards • Recognized by Forrester as a leading Digital Experience Agency
  6. 6. Experience Strategy Customer Data & Insights DX Strategy & Roadmap Technology Consulting Data Consulting Innovation Analytics Experience Design Customer Journey Mapping Content Strategy Digital Experience Design Experience Optimization Technology Platform & Systems Consulting Content Management Application Development Global Content Delivery Mobile Commerce Digital Enablement DX7 Assessment Tools & Frameworks Product Acceleration Training & Education Staffing Connected Expertise
  7. 7. Select technology partners
  8. 8. Select Clients
  9. 9. Need Compare Discover Research Purchase Decide Use Get Help Share Personalize PartnerEvangelize Tribe Search Customer Experience Journey The Customer Experience Journey Earning a unique role in your customers’ lives. From function, to value to meaning.
  10. 10. It starts with a framework; Builds a process; Establishes governance; Links with strategy; Is built with technology; And measures success.
  11. 11. Known user attributes Voice of customer Observed behavior Known user opinions Known user behaviors Touchpoint opinions Customer insights
  12. 12. Customer Data Platforms
  13. 13. The reality of “business” data
  14. 14. What do we mean by business data?
  15. 15. Foible #1: Connecting Silos
  16. 16. What are your silos?
  17. 17. Customer Support
  18. 18. Connecting Data Silos CRM Web Analytics CMS Email Personalization Cobweb Ad serve Hub-and-Spoke Customer Data Platform Personalization Testing CRM Web Analytics CMS Email Ad serve Testing
  19. 19. Use case: How do we create a single view of customer data? Option 1: Figure out and use a universal ID Option 2: Behavior and cookie matching
  20. 20. You and UUIDs
  21. 21. Customer Support
  22. 22. This can get big in non-relational systems
  23. 23. HOLY SHIT WHAT IS GOOGLE DOING
  24. 24. Who’s in charge of your data? …depends on who you ask!
  25. 25. Remember Master Data Management?
  26. 26. These days many people are talking about Governance instead
  27. 27. Regardless, the idea is having one Source of Truth
  28. 28. Foible #2: Scheming the schema
  29. 29. 1. Do you have the right data? 2. Are you normal? 3. Can you scale?
  30. 30. Do I have right data?
  31. 31. What are you trying to accomplish?
  32. 32. Remember Data Dictionaries? Yeah, let’s do that more.
  33. 33. To Normalize, or not? IT IS A QUESTION!
  34. 34. Pretty Normal Too Normal?
  35. 35. To Normal or to Not Pros: • Keeps data clean • Is a best-practice • It scales Cons: • Will cause table bloat • It’s exactly not easy • It takes time
  36. 36. What about Scalability?
  37. 37. I have seen several examples of PostgreSQL systems that were built to be quick and easy, but had some major performance issues as they didn’t grow out of the Proof of Concept phase. If you are planning on success, plan for scalability too.
  38. 38. Foible #3: Ongoing integrity
  39. 39. Noting is as it has always been
  40. 40. Data Drift Three major types: Structural: Changes at the source. Ex: Fields added, deleted, or changed Semantic: The meaning changes. Ex: Field itself doesn’t change, but what is in it does Infrastructure: Changes in platform and tech. Ex: A new platform is added or is changed Source: Girish Pancha http://www.cmswire.com/big-data/big-datas-hidden-scourge-data-drift/
  41. 41. API “Jitter”
  42. 42. Makes your ETL, SOL
  43. 43. So, what can you do about it?
  44. 44. checking
  45. 45. Foible #4: Making people care
  46. 46. Developers need ____?
  47. 47. Developers need requirements.
  48. 48. Business people need ____?
  49. 49. Business people need insights.
  50. 50. The key is Empathy
  51. 51. Thanks!

×