Bruno Aziza has run data organizations at some of the largest companies on earth, so I asked him the question I ask every data leader, in its purest form: “What is the ideal data culture, and how do you create it?”
Most answers to this question are shopping lists. Buy the platform. Hire the data scientists. Put a dashboard in every meeting room. As if culture were a procurement category.
Here’s the reframe: a data culture is not what your organization believes about data. It’s what your organization does at 4 p.m. on a Friday when the data says something nobody in the room wants to hear. Everything before that moment is branding.
Look at it through the SPICE lenses — Strategy, Politics, Innovation, Culture, Execution:
Strategy. The ideal data culture starts one level above data: with decisions. Companies that catalogue their data before they catalogue their decisions end up with pristine lakes and thirsty executives. Strategy means naming the twenty decisions that actually move the business, then building the data, the literacy, and the access around those — and being willing to let the rest of the data estate stay imperfect. Culture follows priority. It never follows inventory.
Politics. Every data culture has an enemy, and it isn’t ignorance. It’s the executive whose authority rests on being the person who knows — the one whose value is the information only they carry. Self-serve data is an existential threat to that person, and they will fight it with the most effective weapon in corporate life: reasonable-sounding delay. “We need to get the definitions right first.” Sometimes true. Often a filibuster. The politics of data culture is the politics of making knowledge public inside a building that rewards keeping it private.
Innovation. The genuinely new idea here isn’t generative AI on top of your warehouse, exciting as vendors find it. It’s treating data education as a leadership obligation rather than an employee perk. Aziza’s own line of questioning in that conversation — how much of asking good questions of data depends on users knowing the limits of the data — is the tell. The innovation is a workforce that knows what data can’t say. That’s rarer, and more valuable, than a workforce that can prompt.
Culture. You can read a data culture in one meeting. Does anyone ask “compared to what?” Does the most junior person in the room ever correct a number the senior person cited — and survive it? Is “I don’t know, let me check” a respectable sentence? Culture is the sum of what gets rewarded when data and hierarchy collide. Everything else is a values poster in the analytics team’s Slack channel.
Execution. So here’s the machinery: start with one decision your leadership team makes monthly. Instrument it end to end — the data behind it visible to everyone in the room, the outcome tracked against the prediction, the misses reviewed without a defendant. Publish the scorecard internally. Then add the next decision. An ideal data culture isn’t declared; it’s accumulated, one honestly-reviewed decision at a time, until asking for the evidence becomes as reflexive as asking for the budget.
Today, most companies have data assets and call the pile a culture. Tomorrow, the winners will have fewer dashboards and more decisions they can defend line by line.
Ask yourself: if your chief data officer quit tomorrow, would your data culture survive the quarter — or was the culture just one person’s job description?
Sources
CXOTalk Ep. 784, “Data and Analytics Strategy with Bruno Aziza, Head of Data and Analytics at Google Cloud” (Apr 14, 2023) — my questions on data education, data haves and have-nots, and the ideal data culture: https://www.cxotalk.com/episode/mastering-data-and-analytics-strategy-with-bruno-aziza-head-of-data-and-analytics-at-google-cloud
Leave a Reply