Phil Marriott-Smith

Every data governance program has one. The lone metadata evangelist. They carry the lineage dream like it’s a noble quest. They say things like “let’s just document the fields” as if that’s ever worked on its own. Meanwhile: → The business doesn’t know the tool exists → IT’s too busy rewriting pipelines → Product is shipping AI features without knowing the source → The only thing being updated is the org chart And still… they believe. In visibility. In documentation. In a world where someone might care what “Customer_ID” actually means. But here’s the truth: Metadata isn’t a tooling problem. It’s not even a documentation problem. It’s a relevance problem. If no one needs it to do their job better it won’t survive. So if you’re launching a governance effort, ask this first: → Who will use the metadata? → Why does it matter to them today? → What’s the lowest-effort way to deliver that value? If you don’t have good answers, you don’t have a metadata strategy. You have a desert. ✔ Follow for daily, no-fluff insights on data, AI, and what actually works. 🔔 Tap the bell on my profile to get notified when I post. ♻️ Repost if you found this valuable.

787

117 Comments

It’s easy to follow certain architectural patterns or throw buzzwords around. It’s much harder to truly internalise and reflect on their underlying message. If you are a Data Engineer, you know Medallion architecture. You can also scream with confidence that we should obviously shift left. But even though we can make both of them work, there is pretty intense tension between them. Fundamentally, they are telling completely different stories. Take a look at Medallion. It’s basically a multi-hop architecture with three layers: bronze, silver and gold. Hence the name. Bronze is raw data from the system. The textbook says that it should land “as is”. It reflects the true state of the source. Whatever is there stays like that. Then you have a Silver layer. Depending on the author, you can find more and less extreme versions of this. In some, you do as few transformations as possible: just clean a bit, change types and do basic validation. In other camps, you do a bit more. Maybe some renaming, joining, or actual modelling (usually the Data Vault approach). Then you have a Gold layer with the highest level of transformation. Data should be reliable, top-quality, prepared for the end user. You implement business logic there. That’s it at a high level. Now take a look at Shift-left. Schema enforcement and validation at ingestion. Data Contracts that often block code changes if breakage is detected. Basic transformations already included before anything lands in your lake. There is a big push to do it already in the Stream so there is no latency and anything can fetch trusted data. Do you see it? One approach tells you to “dump whatever on the data lake and worry later”. Medallion is fundamentally Shift-right. Shift-left is the other extreme. In a certain regard, it resembles the old ETL. Do much more before even landing data into the lake. It looks like we are trying to find the balance nowadays. We went from classic ETL and hit bottlenecks. So we invented the lake and ELT to move fast. But quickly realised we have another shitshow. So we want to be more careful again about what we ingest. But we don’t want to repeat the old ETL mistakes. So what do we do with that? For now, it looks like we stick to the Medallion but we see the problems. So for certain assets, we shift left, i.e. skip the Bronze and ingest into the Silver straight away. At least that’s one approach. Let’s see where the balancing act lands. I think the final sentence has not yet been said here. Super interested to see how things will evolve over time. —-— 👋🏼 I'm Slawomir. I guide companies at any stage to turn data into clear action and real business impact. ♻️ Repost to help your network.

48

12 Comments

I’ve been in business for nine years this month. Here are nine (of the many) things I’ve learnt about data storytelling: 1. As sexy as data visualisation sounds, we’re now saturated in data visuals that are too easy to ignore. 2. The audience decides what works, not me. 3. The hard part isn’t finding numbers, it’s understanding what matters. 4. Without context, data is too easy to misread. 5. The same dataset can tell multiple truths, depending on the lens. 6. People repeat the story back, not the numbers. 7. The majority of people who want to tell data stories don’t, because it’s hard. 8. Data analysts have more power and influence than they realise. 9. People connect with the message, not the method. What would you add? ---- Kia ora, I'm Kat 👋 I wrote The Data Storyteller's Handbook to help improve how we communicate data insights. Check it out here: https://lnkd.in/gWs2kTP Use the code 9YEARS for 50% off the PDF (September 2025 only).

216

35 Comments

If Your Data Governance Council Isn't Asking These 22 Questions, It's Already Falling Behind on Innovation.   The era of Data Governance as a passive, policy-enforcing committee is over. In the era of data driven business, if your council isn't actively fueling the business strategy, it's just overhead. It's not about control. It's about acceleration.   Many councils are stuck reviewing data catalogues and debating definitions. Meanwhile, the business is screaming for data-driven innovation. This gap is what makes them obsolete. The shift starts by asking fundamentally different questions.   We've moved beyond basic governance. The modern council needs an Innovation Agenda. It’s built on 5 pillars, driven by questions that challenge the status quo:   Category 1: Strategy and Ideation (Focuses on where to innovate and how to generate ideas) 1.    What methods are we using to identify new business opportunities where data can be a key driver? 2.    How can we break down data silos to encourage cross-departmental collaboration and spark new ideas? 3.    What competitive intelligence and market data are we analyzing to identify opportunities for differentiation? 4.    How are we leveraging customer data and feedback to understand unmet needs and drive personalized solutions? 5.    What criteria do we use to prioritize and allocate resources to the most promising data-driven innovation initiatives? 6.    How can we monetize data (directly or indirectly) to create new revenue streams or business models?   Category 2: Technology, Data and Analytics Capabilities (Focuses on the tools and data infrastructure needed)
   7.    What investments in technology (e.g., cloud platforms, data lakes) and data integration strategies are necessary to provide a comprehensive view for innovation?
 8.    How are we leveraging advanced analytics, AI, and ML to uncover non-obvious insights that drive innovation?
 9.    What infrastructure and processes do we have to support the rapid analysis and application of real-time data?
 10. What tools and processes ensure decision-makers have access to high-quality, trusted data? 11. What interoperability and industry-standard practices ensure our data-driven innovations scale across ecosystems and value chains?       Category 3, 4 and 5 continue in 1st and 2nd comments.   What's the one question you believe every Data Governance Council MUST be asking today? Share it in the comments below.   Transform Partner – Your Strategic Champion for Digital Transformation   Image Source: UNHCR

104

16 Comments

Creating a custom data analytics solution for a business is apparently only half the job. I thought I understood that by now… But no one warned me how tricky the implementation phase would get! Once the system was built, I had to train the staff to actually use it …and that challenged me on a whole new level. Here’s what I learned as a result. 📌 I can’t be too “high-level” with my explanations. I had to simplify every concept. 📌 A custom-built solution is never truly “done” until it replaces the legacy system in use (in this case, pen-and-paper books) and is easy enough for the team to use while getting more done in less time. 📌 Teaching requires deep patience, especially when you’re someone who naturally learns fast. I’ve had to slow down, adapt to each person’s pace, and create an environment where learning feels safe and exciting. Even talking to business owners has been a lesson in communication: 📌 No tech jargon. 📌 No buzzwords. 📌 Just clear value in as few words as possible I am very much still a work in progress But this journey is stretching me in all the right ways. I now see how vast data analytics application really is and can be. Approaching it from this angle has allowed me to apply every skill I have and test the skills I’m still developing. So if you’re a data analyst reading this, you need to start thinking in terms of real-world application. What kind of problems do you want to solve with your skillset? That’s where your career clarity lies. And if you are still struggling to find yourself or figuring things out, it's okay too! Rome wasn't built in a day; you will get there! I want to help! Drop where you’re currently at in your journey in the comments I’ll do my best to give you a head start. 😊 #DataAnalytics #BuildInPublic #DataDriven #Growth #Career #KeepLearning

51

10 Comments