Analytics Best Practices
- Huntington Beach, CA, December 2025
Photographic imaging technology slowly advanced for decades before high-quality, affordable digital cameras hit the market in the early 2000s, but when they suddenly appeared, companies that tied their business strategy to film suddenly disappeared.
When new technologies take people by surprise, they’re constantly playing a game of catch-up.
You can tell how unprepared people are by the questions they ask. A CIO asked me whether AI would increase demand for business analysts as they find more insights in the data. Then another asked if AI would eliminate the need for business analysts by automating their insights.
Jonathan Grudin, in his introduction to The Discipline of Organizing, says, “Human beings do not reason well about exponential growth; our experience is linear, not exponential. What we overlook is that exponential growth can proceed for a long time under the radar.”
Did you recently change your whole analytics strategy to focus on artificial intelligence? What was your plan before this change? Will your strategy survive the test of time?
Constancy of Purpose
I had the rare experience of working at a Japanese-owned semiconductor company in the 1990s, which explains why I’m always quoting W.E. Deming. After WWII, Japanese car manufacturers retooled and overtook their American competitors in just two decades by applying Deming’s principles to every part of their operation – including data management.
Deming urged leaders to create “constancy of purpose” with an unwavering, long-term commitment to a clear goal. That’s a personal character quality of managers at every level, not a top-down strategy. He offers three ways to make constancy the foundation of your leadership:
Care about quality. “Innovation, the foundation of the future, cannot thrive unless the top management have declared unshakable commitment to quality and productivity,” Deming says.
- Understand the work. “[but] it is not enough that top management commit themselves for life to quality and productivity. They must know what it is that they are committed to— that is, what they must do.”
- Remember the theory behind your work. “Experience alone, without theory, teaches management nothing about what to do to improve quality and competitive position, nor how to do it.”
Compared to their American counterparts, these Japanese managers were thinking in reverse, playing the long game. Executives thinking about their analytics strategies today should learn from this.
Applying Deming
Whenever I get the chance to talk with a CEO (or anyone with a C-title), I ask them, “What’s your data strategy?” They almost always name a new technology, sounding like they just read about it in a magazine. They could use some constancy.
Here are some ways you can apply Deming’s wisdom to your data strategy:
1. Deepen Knowledge.
Why would it ever make sense to hand over knowledge work to someone who knows nothing about your company? Shallow, outsourced solutions won’t improve data quality, and worse, you’re missing the opportunity to develop knowledge instead of software.
Jessica Talisman says that a lack of purpose “perfectly abstracts the current state of knowledge representation in organizations…knowledge is undervalued as an underpinning or bedrock for organizations and society for a few years. We are now seeing the impacts of this undervaluation.”
She’s exactly right:
If you care about quality, you’ll deepen institutional knowledge across your teams and systems.
2. Organize Data.
A grocer who “understands the work” wouldn’t keep their inventory of food in one giant pile in the middle of a grocery store. Neither should data leaders ignore the discipline of organizing data.
Developing knowledge maps, master data hierarchies, data dictionaries, and metadata reporting are all examples of organizing data. It might surprise you how many data teams don’t do any of these, even though this discipline pre-dates the Dewey Decimal System. If you embrace this mindset, you’ll see connections between structured business data and the large language models that power AI.
3. Adopt System Thinking.
“Starting with the theory” changes the way you think about every data solution. Instead of viewing data solely as inputs to people’s decisions, systems thinkers recognize the inputs, outputs, flows, and feedback loops that shape the data asset. Systems thinkers survive the test of time by constantly adjusting their understanding of how the business works.
People who think like this always make certain they understand the purpose of the work before deciding what tool to use. This kind of thinking once helped my team eliminate batch processing in our data platform at zero cost. SAP charged millions for this feature.
Strategic Pillars
Even last week, a colleague I worked with at Broadcom in the early 2000s asked what it would take to reproduce those solutions in their company today. I told him it wasn’t a technology; it was these three strategic pillars. His question proved it: yes, a data strategy can survive the test of time.
To remind you of this week’s data concept, enjoy Alive And Kicking, by Simple Minds, from the Frictionless Data Spotify playlist.