Are We Dangerously Obsessed with Data Collection?
2 likes189 views
The document discusses whether organizations are too obsessed with data collection at the expense of effectively using data. It notes that while data collection has potential value, there is no inherent business value in simply collecting data. Vendors and consultants have an incentive to overpromise the value of data collection. However, data usage is difficult work that requires mental models and creative thought rather than just technology. Effective data usage involves performance measurement, hypothesis validation, and operational enablement. Overall, the document argues organizations should focus more on realizing actual value from data usage rather than just data collection.
![Hypothesis
Validation
What idea do we have that,
if true, would lead to action
that would have a positive
business impact?
We believe [some idea]
because [some
observation or evidence].
If we are right, we will
[take some action].
Hypothesis!
Check for actionability!](https://image.slidesharecdn.com/2024-02-columbusdaw-datacollectionobsession-timwilson-240208155823-4a06753f/85/Are-We-Dangerously-Obsessed-with-Data-Collection-39-320.jpg)
Are We Dangerously Obsessed with Data Collection?
- 1. Are We Dangerously Obsessed with Data Collection? Tim Wilson Columbus Data & Analytics Wednesday February 2024 Yes!
- 2. @tgwilson Consider any task that touches or impacts data to be data work
- 3. @tgwilson We can break data work down into two distinct types of activity
- 6. @tgwilson There is no inherent business value in collecting data. There is only potential value.
- 7. Analytics teams should only build, maintain, and extend the analytical platform And yet
- 8. Analytics teams should only build, maintain, and extend the analytical platform for stakeholders to do their own analysis on. Everything else just holds the business back. Digital Analytics Industry Veteran And yet
- 9. @tgwilson The Expectations Chasm I have gathered all of the data for you! I need insights and recommendations!
- 10. @tgwilson We assume that marketers just know how to put data to productive use.
- 11. @tgwilson They have business questions, so isnt it just obvious how to use the data?
- 13. @tgwilson My parents bought me a nice, new car & assumed I would just know how to drive it!
- 14. @tgwilson Who is responsible for creating and delivering the driving lessons?
- 15. @tgwilson Why is this disconnect so common?
- 16. @tgwilson There are powerful economic forces driving the idea that data collection has inherent value Data Collection Data Usage
- 17. @tgwilson Consider: The incentive structure for technology vendors High Fixed Cost + Nominal Variable Costs The more customers we sign, the more profit we make!
- 18. @tgwilson Consider: The incentive structure for technology vendors High Fixed Cost + Nominal Variable Costs The more prospects who believe (our) technology is the answer, the more profit we make!
- 19. @tgwilson Consider: The incentive structure for technology vendors Implement our platform and you will immediately see business value! Implement our platform and, with a lot of additional hard and messy work on your organization, your processes, and your people you will hopefully eventually see business value! vs.
- 20. @tgwilson Consider: The incentive structure for consultancies Nominal Fixed Cost + High Variable Costs High profitability comes from repeatable processes with a scalable workforce!
- 21. @tgwilson Consider: The incentive structure for consultancies Implementing technology does this more than deeply engaging with the business Echoing technology partners messaging is low-cost content
- 22. @tgwilson Vendor Gets $$$! Good luck with that! Consultant Gets $$$! Buy the Tech Implement the Tech Business Value Story Time!
- 23. @tgwilson This message from vendors and consultants is flawed Buy this data collection machine. It is inherently valuable!
- 24. @tgwilson Consider: The incentive structure for data/analytics teams Technology updates are tangible Technology updates fit into a standard dev process Technology updates buy time (Were working on it!)
- 27. @tgwilson Data usage is hard and messy work! Reality
- 28. @tgwilson It is more about mental models and creative thought than any specific technology Reality
- 29. @tgwilson Data usage can be broken down into three buckets
- 30. Performance Measurement Where are we today relative to where we expected to be today at some point in the past? Hypothesis Validation What idea do we have that, if true, would lead to action that would have a positive business impact? Operational Enablement We are directly putting the data itself to tactical use as part of a clearly defined process.
- 31. Performance Measurement Where are we today relative to where we expected to be today at some point in the past? What are your KPIs?
- 32. Performance Measurement Where are we today relative to where we expected to be today at some point in the past?
- 33. Performance Measurement Where are we today relative to where we expected to be today at some point in the past? Two Magic Questions: What are we trying to achieve? How will we know if weve done that?
- 34. Performance Measurement Where are we today relative to where we expected to be today at some point in the past? Two Magic Questions: What are we trying to achieve? How will we know if weve done that?
- 35. Performance Measurement Where are we today relative to where we expected to be today at some point in the past? Two Magic Questions: What are we trying to achieve? How will we know if weve done that? Meaningful KPIs (Metrics + Targets)
- 36. Performance Measurement Where are we today relative to where we expected to be today at some point in the past?
- 37. Performance Measurement Where are we today relative to where we expected to be today at some point in the past? What do we do if were missing expectations?
- 38. Performance Measurement Where are we today relative to where we expected to be today at some point in the past? Hypothesis Validation What idea do we have that, if true, would lead to action that would have a positive business impact? hypothesis h朝-p辰-th-ss a tentative assumption made in order to draw out and test its logical or empirical consequences Source: Merriam-Webster definition #2
- 39. Hypothesis Validation What idea do we have that, if true, would lead to action that would have a positive business impact? We believe [some idea] because [some observation or evidence]. If we are right, we will [take some action]. Hypothesis! Check for actionability!
- 40. Hypothesis Validation What idea do we have that, if true, would lead to action that would have a positive business impact? Operational Enablement We are directly putting the data itself to tactical use as part of a clearly defined process. The ordering process on an ecommerce website A next best action B2B lead management process A dynamic landing page personalization process Etc.
- 41. Operational Enablement We are directly putting the data itself to tactical use as part of a clearly defined process. Why do we think the process will deliver value? How will we determine it delivers that value as expected over time? Hypothesis Validation Performance Measurement!
- 42. Operational Enablement We are directly putting the data itself to tactical use as part of a clearly defined process. Generative AI is an operational enablement use of data.
- 43. @tgwilson Data Collection & Management Data Usage Gather the Data Integrate the Data Govern the Data Performance Measurement Hypothesis Validation Operational Enablement Potential Value Realized Value
- 44. Thank you! Use the data! Images: Midjourney AI These 際際滷s bit.ly/daw-2024 Podcast analyticshour.io Social @tgwilson (X/LinkedIn) @tgwilson.bsky.social (Bluesky) Email tim.wilson@gilliganondata.com /imagine prompt Tims inner being in a mid-century modern style
- 45. @tgwilson Is Our Obsession with Data Collection Keeping Us from E鍖ectively Delivering Value? Every digital analyst wants to deliver meaningful value to their organization. Why, then, does it often seem like we're just constantly fighting fires to get the data into good enough shape to do proper analyses, with no time left at the end of the day to actually put the data to work? In this talk, Tim will make the case that, as an industry, we've let ourselves get enamored with data collection and management to the detriment of putting enough focus on actually putting that data to meaningful use. And, he will describe a useful framework for the di鍖erent ways that data can be put to use.