#research-piece I am a data person who is keen on understanding how my career can help shape the global transition to a green economy[^1]. This [2024 DNV Report](https://brandcentral.dnv.com/original/gallery/10651/files/original/b7ba9211-780b-4c6a-bb6c-396f1425af4e.pdf)suggests a clear link between being "digitally advanced" and being better at almost everything, including decarbonisation. This note captures what the report actually shows, where the gaps are, and what opportunities exist for someone with data skills. >[!info] > I'm reading this critically. I want to understand what's really happening versus what's marketing language. ## The methodology **Research design:** DNV commissioned FT Longitude (Financial Times research) to survey senior energy professionals, plus conducted in-depth interviews with industry leaders and experts. ![[digit-role-green-econ-meth.png]] Figure: Survey sample breakdown by sector, geography, company size, and respondent seniority from the DNV report. Critical gap: The report doesn't clarify how many companies these 1,289 respondents represent. Are we looking at one opinion per company, or is one company over-represented? If one person answered for a 50,000-person utility versus someone at a 200-person firm, they carry equal weight in the aggregate. This matters for interpreting the results. I'd want to know the company distribution before trusting the numbers too heavily. ## Digital Leaders are better at almost everything The report segments respondents into three groups: - Digital Leaders (28%): say they're industry leaders in digitalisation - Digital Laggards (37%): say they're behind - Middle 35%: neither clearly leading nor lagging The interesting part: Digital Leaders don't just perform better at digital tasks. They're more optimistic about hitting revenue targets (77% vs 55%), profit targets (68% vs 51%), and crucially, decarbonisation targets (65% vs 31%). This could mean: 1. Being digitally mature helps you hit climate goals (plausible) 2. Optimistic companies invest in both digital and climate work (also plausible) 3. Digital leaders self-select for having more realistic (or ambitious) targets in all areas The report doesn't distinguish between these. It just shows they correlate. ## What "digitalisation" actually means here The report defines it as: using digital technologies to optimise energy generation, transmission, distribution, and consumption. In practice, this includes: - Smart grids for efficient power supply - Predictive maintenance in power plants - Real-time data for decision-making - Automation of operations But this is still too broad in my opinion. When the report lists "optimising processes" as the most impactful application (68% of leaders report major impact), what does that actually mean? You could optimise a single spreadsheet limited to one team or you could just redesign your entire supply chain. Where the report gets concrete however is Integration of systems and databases. Two-thirds of Digital Leaders say this had massive impact. Why? Because energy companies have legacy systems that lock data away. Getting data out and overlaying it with the actual network topology—what Sabine Erlinghagen (Siemens) calls "creating a digital twin"—suddenly makes the raw data usable. You can see the whole system instead of isolated data points. ## The opportunities for data skills From the report, here are areas where data work directly drives value: ### System integration and data liberation - The problem: Data is trapped in legacy systems. One company has financial data here, operational data there, sensor data nowhere - The opportunity: Someone has to design how to get all that data accessible. That's data architecture and engineering work ### Monitoring at scale without sensors everywhere - Example from the report: Power transformers don't tell you they're failing until they fail. With EVs adding load to the grid, transformers are now overloaded more often. But putting sensors on every transformer would take years - The solution: Use the meter data you already have, overlay it with the grid topology, and use AI to estimate which transformers are at risk - This requires: data engineering (getting the right data), analytics (understanding the pattern), and enough domain knowledge to know what you're looking for ### Data quality as a prerequisite, not a side project - 94% of Digital Leaders say they'll prioritise improving data quality in the coming year - 69% expect to use AI in operations in the coming year - The connection: AI needs good data. Period. If you want to deploy AI in your organisation, you first need someone to audit, clean, and structure your data. That's a data role that's critical but often unglamorous ### Building decision support systems (not autonomous systems) - Important caveat from the report: "AI will not be used to automate decision-making for critical energy infrastructure for the foreseeable future" - What it will do: help operators understand complex systems faster and make better decisions - This is a significant constraint but also a reality. You're building tools for people, not replacing them ## The barriers are different for both leaders and laggards The report breaks this down (Figure 2.4), and the differences are revealing. ![[barriers-to-digit.png]] Figure: Barriers to digitalisation differ significantly between Digital Leaders and Digital Laggards. For Laggards, the biggest barrier is resistance to change (39%). For Leaders, the biggest barrier is the cybersecurity risk (36%), but this is closely followed by resistance to change again (33%). In energy, failure is genuinely unthinkable. You can't "move fast and break things" when you're managing critical infrastructure. This creates a tension between the speed of software development and the safety requirements of energy systems. For a data person, the strategic insight is understanding a company as either a laggard or a leader, since each needs different expertise. Laggards need someone to help them change their approach and build foundational capabilities. Leaders need someone who understands both data security and architecture to push forward. ## What separates digital leaders: Mostly expectations, some strategy When asked what makes them strong in digitalisation, Digital Leaders cited the following (Figure 3.3): ![[strengths-for-digit.png]] Figure: Top factors Digital Leaders attribute to their digitalisation success. The top 3 seem like corporate templates to me. What does "culture of innovation" actually mean operationally? But points 4 and 5 are concrete: you need to pick technology that fits your constraints (not just the fanciest), and you need to work closely with vendors who understand your domain. It's striking that digital literacy ranked 7th at 21%, well below "culture of innovation" at 44%. Digital literacy is far more actionable and concrete - you can measure and teach it - whereas a "culture of innovation" is vague to assess and harder to build. ## Key technologies mentioned When the report talks about digitalisation, it means: - Sensors, IoT, and remote communications: Getting data from the physical world - Digital twins: Virtual models of physical systems (more on this below) - AI and machine learning: Making sense of all that data - Digital automation and control networks: Letting systems respond to data automatically - Remote sensing and witnessing technologies: Monitoring things from a distance Digital twins deserve a note: This term is thrown around a lot. In the report, it's used to mean: taking data from multiple sources (sensors, SCADA systems, etc.) and combining it with the actual network topology to create a model where you can run scenarios. Example: "What happens to this grid if we lose this generator and have a heat wave?" You can test that digitally before it happens in reality. This is different from "a digital copy of something". It's specifically useful because you overlay the topology. ## What remains unclear (and worth investigating in the future) ### How much of "data" is self-collected vs. purchased? - When companies say they're being data-driven, are they building data collection infrastructure (sensors, systems) or mostly working with data from vendors and standard systems? - A utility might buy grid data from equipment manufacturers. That's different from instrumenting their own operations. ### What are the actual technical barriers? - The report lists "data quality/management issues" as a barrier, but doesn't dig into what that means - Are we talking about schema inconsistency? Missing data? Incorrect data? Slow data pipelines? These require very different solutions ### How do you actually move from "we have a lot of data" to "this data is useful"? - The report gives one example (grid topology overlay) but that's domain-specific - What's the general pattern? What should someone know who's facing this challenge in their organisation? ### What's the economic return? - The report says AI could reduce clean energy generation costs by $1.3 trillion globally by 2050, but doesn't break down what organisations are actually seeing today - Are Digital Leaders realising 10% margin improvement? 5%? Still figuring it out? ## Where a data career matters in this transition Based on the report, here's what I see: - Infrastructure: Building the systems to get data out of silos and making it accessible. (Data engineering, data architecture) - Foundation work: Making sure data is clean, documented, and trustworthy enough for decisions and AI. (Data quality, data governance) - Analysis and insight: Understanding what patterns exist in operational data and what they mean. (Analytics, domain expertise) - Bridges: Working between the systems engineering teams and the decision-makers to make sure data tools actually solve real problems. (This is harder to name but critical) The gap in the report: It doesn't explain how the Leaders got there. It just shows that they did. What does a successful digitalisation effort actually look like over 3-5 years? That is analysis for another day. [^1]: See my early motivations [[My First Brush with Sustainability - How I Launched (and Killed) a Tetra Pak Collection Drive]]