Boyd Evert

This is the fastest way data ops debt gets ignored. Someone flags an issue. “Some fields aren’t lining up.” “A few records need cleanup.” It sounds small. So teams move on. They add people. They add rules. They add pressure. Nothing really improves. Why? Because the problem wasn’t speed. It was debt. Data ops debt doesn’t announce itself. It shows up quietly. In rework. In reports you don’t trust. In SLAs that look fine but feel wrong. The teams that fix it don’t talk about quality. They prove it. They standardize one field. They add one validation. They clean one flow end to end. No big program. No new tool. They remove the friction. Data ops debt isn’t solved by explaining better. It’s solved by doing the work. Showing beats telling.

9

The #1 mistake analysts make? Jumping straight into the data. I used to do it all the time. You get a problem statement, query data, pull the numbers - and three days later, you realize you solved the wrong problem. Stakeholder’s unhappy. Your analysis goes unused. And you're left wondering what went wrong. After 3 years of hitting that wall, I’ve changed one small habit that’s made a big difference: I don’t start with data. I start with clarity. Before I look at a single metric, I now: → Break down the problem → Map out a hypothesis tree → Imagine what data could prove or disprove each path Nine times out of ten, this shows me if the problem statement itself is unclear. That’s when I pause and go back to the stakeholder. Ask questions. Align on what we’re really solving for. Only then do I go deep into the data. This shift has saved me 25% of my time and spared me from presenting multiple wrong analyses. It’s made my work sharper, and more importantly, more useful. If you’re an analyst, ask yourself: *Am I solving the right problem, or just the most convenient one?*

8

Why percentages are manipulation disguised as math: "Complaints decreased 50%!" Could mean: - 1000 → 500 (significant) - 10 → 5 (not significant) - 2 → 1 (why are we even tracking this?) Or: - We changed what counts as a complaint - We made it harder to submit complaints - We stopped measuring for two weeks Always show the absolute numbers. Always show the context.

109

11 Comments

Building a modern data platform isn't just about technology, but having the right partners. Rittman Analytics just published a case study on our work together at BARK. They didn't just migrate us to BigQuery and Looker—they embedded with our team and fundamentally elevated how we work. Through hands-on coaching, skills transfer, and establishing best practices, they ensured we could independently manage and scale the platform long-term. The technology transformation was table stakes. The real value was building a more capable, confident data team with modern processes we now own completely. The result: a single source of truth, self-service analytics across the business, and a strong foundation layer for AI. If you're evaluating data consultancies, this is the model to look for—partners who work themselves out of a job by making your team better. Thanks to Mark Rittman Lydia Blackley Bailey Sharp-Ledger Lewis Baker Olivier Dupuis https://lnkd.in/ehu6PhxJ

38

5 Comments

Your data team is shipping 'wins' that die in 90 days. Meanwhile, one manufacturer's data team is generating +8-figures in revenue. Difference? Most companies think they’re scaling data. They’re actually scaling chaos. Every “use case” sales dashboard, churn model, ops KPI report gets celebrated. But under the surface? Each new use case silently rebuilds the same data logic, pipelines, and metrics… again. By year two, you’ve built ten isolated “wins” that don’t compound. But when the CEO asks, “What’s revenue this month?” ...4-different numbers. Here is what giant tech industrial manufcaturer figured out. They used to chase use cases too, project after project, team after team, until they hit the wall. Scaling 100+ projects with different systems became unmanageable. The data team ships dashboards, the business claps, and three months later everyone’s back in spreadsheets. They changed. Instead of trying to scale use cases, they started scaling the data products that power them, reusable, governed assets that any business unit could consume. They stopped treating data as a “project one-time-deliverable” And now of now they are main growth engine earning the organizaton high 8-figures Here is the playbook: 1. Audit your use cases for duplication. Pull the top 10 dashboards, models, or use cases you’ve built. Highlight every dataset, metric, or definition that shows up twice. Ex. If you’re solving “forecast accuracy” in Sales and “revenue leakage” in Finance, both depend on the same “Revenue Data Product.” 2. Define what “done” actually means. A data product isn’t “done” when the table is populated. It’s “done” when anyone can use it without Slack messages or rework. That means it has clear ownership, definitions, documentation, and guaranteed quality. If your team disappears tomorrow, the business should still run on it. 3. Separate your foundation from your use cases. Most data teams build monoliths: Every use case has its own pipelines, joins, and metrics. Flip it. Build modular. Your foundation (data products) should live in a shared layer. Use cases sit on top, consuming those products. Most companies rebuild pipelines every project because they don’t separate raw → curated → product layers. 4. Treat Data as a Product, Not as Project This mindset shift changes everything: Projects have deadlines, budgets, and handoffs. Products have owners, roadmaps, and feedback loops. That means: Stop “handing over” data models. Start iterating on them with users. Measure adoption, not completion. Schedule feedback sessions Adoption happens when trust and results are shown. 5. Measure impact, not volume. Stop counting “number of dashboards shipped.” Start tracking “impact.” and number of data products reused across teams That’s how you measure compounding ROI instead of constant rebuilds. I often say data leader should act as CEO. Each initaitve should be foundation for the next one

89

25 Comments

The Business Analytics & Information Systems (BAIS) landscape has shifted more in the last 24 months than in the previous decade. If you are still thinking of this field as just "spreadsheets and SQL," it’s time to zoom out. As we move through 2026, here are the 3 major shifts defining the future of BAIS: 1️⃣ From Insights to Action (The Rise of Agentic AI) 🤖 We are moving past descriptive dashboards. The new gold standard is building AI Agents—systems that don’t just show you a sales dip, but autonomously trigger supply chain adjustments to fix it. 2️⃣ The Democratization of Data 🗣️ Natural Language Querying (NLQ) is changing the game. Technical barriers are falling, meaning the value of a BAIS professional has shifted from "knowing how to code" to "knowing which business questions to ask." 3️⃣ Governance is the New Competitive Advantage ⚖️ With AI regulation tightening, ethical data oversight is no longer a "nice-to-have" in the curriculum—it’s a core business requirement. Managing bias and ensuring data privacy is now a primary driver of brand trust. The Bottom Line: The market isn't just looking for data crunchers; it's looking for Data Translators—professionals who can bridge the gap between complex machine learning models and C-suite strategy. The demand for these roles is projected to grow by 35% through 2032. There has never been a more exciting time to be at the intersection of tech and business. #BusinessAnalytics #InformationSystems #AI2026 #DataStrategy #FutureOfWork #DigitalTransformation

31

📉 𝗧𝗵𝗲 𝗣𝗼𝘄𝗲𝗿 𝗼𝗳 𝗧𝗵𝗼𝘂𝗴𝗵𝘁𝗳𝘂𝗹 𝗗𝗮𝘁𝗮 𝗩𝗶𝘀𝘂𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻 America is experiencing a profound decline in face-to-face social connection. Between 2003 and 2023, in-person socializing dropped across every demographic, with some groups seeing decreases of 40%. This matters for policy, civic engagement and how we strengthen social cohesion. But here's the thing: 𝗵𝗼𝘄 𝘄𝗲 𝗽𝗿𝗲𝘀𝗲𝗻𝘁 𝗱𝗮𝘁𝗮 𝘀𝗵𝗮𝗽𝗲𝘀 𝘄𝗵𝗲𝘁𝗵𝗲𝗿 𝗽𝗲𝗼𝗽𝗹𝗲 𝘂𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱 𝗶𝘁. The original visualization used to make this important point displays vertical bars with rotated text, forcing readers to tilt their heads and making comparisons difficult. Also, no data grouping means important patterns get buried. So, I reimagined it with three simple design changes. The difference?   • Instant readability. No neck acrobatics required   • Clear patterns emerge within each demographic category   • The most affected groups jump out immediately   • Easier to spot surprising insights (like the 40% decline among lower-income Americans) 𝗖𝗵𝗲𝗰𝗸 𝘁𝗵𝗲 𝗿𝗲𝗱𝗲𝘀𝗶𝗴𝗻𝗲𝗱 𝘃𝗶𝘀𝘂𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝗶𝗻 𝘁𝗵𝗲 𝗰𝗼𝗺𝗺𝗲𝗻𝘁𝘀 👇 The difference will change how you think about presenting data. You'll instantly see: ✓ Which groups are most isolated ✓ The surprising age pattern everyone misses ✓ Why the original design obscured the real story When communicating to policymakers, stakeholders or the public, a sharp visualization can be the difference between action or indifference. 𝗞𝗻𝗼𝘄 𝘀𝗼𝗺𝗲𝗼𝗻𝗲 𝘄𝗵𝗼 𝘄𝗼𝗿𝗸𝘀 𝘄𝗶𝘁𝗵 𝗱𝗮𝘁𝗮 𝗼𝗿 𝗽𝗼𝗹𝗶𝗰𝘆? 𝗧𝗮𝗴 𝘁𝗵𝗲𝗺 𝗮𝗻𝗱 𝗮𝘀𝗸 𝘄𝗵𝗮𝘁 𝘀𝘁𝗮𝗻𝗱𝘀 𝗼𝘂𝘁 𝗺𝗼𝘀𝘁 𝗶𝗻 𝘁𝗵𝗲 𝗿𝗲𝗱𝗲𝘀𝗶𝗴𝗻. #DataVisualization #PublicPolicy #CivicEngagement #DataStorytelling #SocialCapital #DataDriven

1

1 Comment

Most dashboards & reports fail for one specific reason. It rarely has to do with the tool or the data. The problem is much simpler: metric definitions. A dashboard full of metrics means nothing if no one agrees on what they mean. I see this pattern constantly: ❌ sales trusts a spreadsheet over the dashboard. ❌ finance / Product report conflicting churn %. ❌ "active user" has 3 different definitions. Before you build the dashboard, build the Metrics Dictionary. It can be a simple Google Sheet that lists out your metrics + descriptions. Here are the columns we include in it: 1. unit → is it €, $, % 2. filters → how to slice the data 3. calculation → how it's calculated 4. department → who cares / uses it 5. metric owner →who is accountable 6. refresh rate → how often it updates 7. description → what the metric means 8. data source → where the data comes from I attached the template we use in the comments. Copy it, tweak it, use it. ✌️

18

4 Comments