Decoding Data Trust: Building Confidence in Clean Energy Management Software
By Will Troppe, Director of Product at Power Factors
You can’t use a data platform you don’t trust. Data drives decisions in the renewables industry, and the ability to trust the information and insights being delivered to you can make or break operational success.
Clean energy management software is designed to enable smarter, faster decision-making by providing clarity and insights into asset performance. Untrustworthy software is worse than useless — it’s a liability.
Data trust challenges are particularly vexing in the clean energy industry due to its distributed nature, variety of data types and sources, and large volumes of high-velocity data of questionable veracity.
In Power Factors’ recent “Decoding Data Trust” video series, I break down data trust: what it is, why it’s important, and how software vendors can build (or break) it.
Trust Is Assessed, Not Assumed
Trust is hard to build and easy to break. It’s built over time, through observation and experience, and can break in an instant. When it comes to data trust, the burden of proof is on the data platform. You know you have a trustworthy data platform if you’re confidently acting on higher-level insights, metrics, and reports rather than being forced to return to raw data.
You know you have a trustworthy data platform if you observe:
- Users operating by exception, with an assumption of high data quality unless explicitly instructed otherwise, because users should be calm, not on edge
- Data consistency across different data access methods (UI, API, BI, and beyond), because internal inconsistency is the best way to erode trust
- Automatic handling of late and changed data, because remote renewable energy data connections are fickle and data intermediaries plentiful
- Visibility into logic and data processes, because understanding cause and effect visually builds trust
Watch part one of the Decoding Data Trust video series “Why It Matters in Clean Energy” to learn more about why trust in clean energy data systems is so critical and what to look for in a trustworthy system.
Part 1: Decoding Data Trust: Why It Matters in Clean Energy
The Four Principles of Building Data Trust
Different users build trust differently. Technical analysts, executives, and operations teams each have varied expectations when approaching renewable energy management software. Still, building data trust with any user adheres to some fundamental principles.
When choosing a software vendor, look for a provider that understands — and practices — the four principles of building data trust:
- Traceability: Seek providers who prioritize transparency in their data handling, allowing you to follow along from raw data to recommendation, understanding each step in the data transfer and calculation pipeline.
- Visualization: Solutions should provide clear visual validation of conclusions and recommendations, supporting advanced analytics insights by delivering visualizations that allow you to validate and understand the results.
- Repeatable Reliability: Trust is built through repeatedly delivering insights that lead to positive outcomes. When a system’s recommendation leads to a faster resolution of an issue, you’re more likely to trust it the next time; when you trust it the next time, you’re rewarded again. It’s a positive feedback loop that helps your business and your career.
- Consistency: You need a solution that delivers consistent, reliable performance. Inconsistent platform performance can lead to differences between expectation and reality and false negatives, which erode trust.
And another important piece of building data trust? Documentation.
Good documentation begins with an intuitive, easy-to-use product that reduces its need in the first place. (When did you last consult the user manual of your favorite social media app?) Then, necessary documentation should be delivered in-workflow. Clean energy professionals are constantly context-switching and working at capacity. When you’re busy and need an answer, you shouldn’t have to look far. Worst-case, relevant details should be a hyperlink away.
In the second part of our “Decoding Data Trust series, “Building Trust Through Transparency,” I talk about why traceability, visualization, repeatable reliability, and consistency are so important to building trust.
Part 2: Decoding Data Trust: Building Trust Through Transparency
How Time, Workflows, and Architecture Influence Trust
In addition to prioritizing the four principles of data trust (plus good documentation), there are a few other factors that play a key role in building your trust in your data platform:
Time: Your world operates in real-time, so your data platform must, too. If not, your cloud truth will differ from your ground truth, which will quickly erode trust.
Workflows: No news isn’t necessarily good news. Workflows must deliver proactive updates, even if the update is, “there is no update.” Proactive “System OK” messages help you feel confident that the system will communicate to you when things are fine and when they’re not — mitigating disruption caused by the potential for false negatives.
Architecture: Owners and operators bet their business on the reliability and scalability of their software platform. Knowing that the architecture behind that system is microservices-based and cloud-hosted builds trust in the system’s ability to deliver now and into the future. A racecar driver must know enough about their racecar to know how to push it to the limit; so, too, must a clean energy operator know enough about their data platform.
Watch part three of the Decoding Data Trust video series, “A Framework for Reliable Data” to learn more about how time, workflows, and architecture influence data trust.
Part 3: Decoding Data Trust: A Framework for Reliable Data
Power Factors’ Approach to Building Trustworthy Systems
Time, workflows, and architecture play important roles in engineering a data platform worthy of trust. How you approach logic systems built on the platform — the baselines, KPIs, and analytics that power performance analysis — builds trust, too, by prioritizing simplicity, transparency, and configurability, in that order.
- Simple: Logic should be as simple as possible for the use case.
- Transparent: Users should have easy visibility into the configurations and formulas that are applied.
- Configurable: Users should be empowered to understand and manage their configurations themselves.
Done out of order, you’ll end up with a system that is configurable in theory but falls short in practice.
There is a direct relationship between a site’s data and the analytics it can support. All sites aren’t equal when it comes to data quality, the number of signals they provide, and the quality of metadata, so why would you assume they are capable of leveraging analytics equally? Claiming you can deliver high-quality analytics regardless of data inputs erodes trust. Still, vendors should design analytics so they can deliver the most value possible from the most imperfect data possible.
The way to build trust in the results of AI algorithms is similar. You build trust in AI-powered analytics the same way you build trust in other parts of renewable asset management software: through visual human validation and repeatable reliability. This is especially true for “black box” AI algorithms that lack clear cause-and-effect relationships.
Watch part four of the Decoding Data Trust video series, “Building Trust in AI Algorithms,” for more about Power Factors’ approach to building trustworthy systems. I also address the question on everyone’s mind: “How do you build trust in AI algorithms?”
Part 4: Decoding Data Trust: Building Trust in AI Algorithms
Will Troppe is Power Factors’ Director of Product. Will has been with Power Factors since 2015 and leads our asset performance management product line. Follow him on LinkedIn!
Interested in learning more about Power Factors’ approach to trusted data? Get in touch!