You’ve probably been in that meeting. The CEO points at a shiny, rotating 3D model on the screen and says, “This is our digital ecosystem.” Meanwhile, the underlying data is a rotting mess of mismatched codes, broken APIs, and unreadable IoT feeds.
We all want the glory of the ecosystem. It sounds great in a pitch deck. But as the data experts at VOYAH (the electric vehicle brand) discovered, building a digital ecosystem from the top down is a massive trap. An ecosystem isn’t something you build; it’s something that grows. And it can’t grow in a swamp of dirty data.
Most product managers treat data scaling as a technology problem. You buy a bigger data lake, you upgrade to a faster warehouse, and you assume the ecosystem will magically form. But the real unlock isn’t the architecture—it’s behavioral.
If your users still treat data as a byproduct of their daily work, your grand ecosystem is already dead. Here is the unglamorous, stage-gated reality of how data capabilities actually evolve, and why you cannot skip a single step.
Stage 1: Stop Building Models. Start Cleaning Data.
In digital twin projects, we see this all the time: the client gets a platform, and the only thing they want to do with it is look at a 3D model. It looks cool when leadership visits, but it gathers dust the rest of the week.
Why? Because we failed to build the one thing that matters: the cognitive habit that data is an asset, not a byproduct.
VOYAH’s first step wasn’t to build a platform. It was to get the data clean. It was the brutal, boring work of hooking BIM models to business codes, standardizing IoT protocols, and unifying space-time benchmarks. Data isn’t a byproduct of your system; it’s the only reason the system exists. Until your team believes that, you’re just playing with expensive 3D toys.
Stage 2: Break the Silos (The Boring Magic)
Once the data is clean and trusted, you can finally connect the value chain. VOYAH built a data middle platform to connect sales, production, materials, and procurement. They found the core value data and made it flow.
In a digital twin project, this is the moment the platform stops being a “cool visualization tool” and becomes a “management tool.” Progress data drives model changes. Alarm data triggers response workflows. The silos collapse, and the data forms a loop.
The flashier the dashboard, the deeper the grave it’s digging if the underlying data is broken. This stage is all about making sure the dashboard is actually telling the truth.
Stage 3: The Ecosystem (If You Have to Force It, You’ve Failed)
Only after you have trusted data assets and a connected value chain can you even whisper the word “ecosystem.” VOYAH is now extending its data capabilities upstream and downstream, collaborating with partners to form an industrial IoT platform.
Notice the sequence: the ecosystem came last. It was the natural result of the first two stages.
If you try to build the ecosystem first, your foundation will crack. The data quality will fail, the APIs will be unstable, and the moment the bottom layer breaks, the entire flashy superstructure will collapse.
Data capability isn’t bought; it’s grown. Stop trying to buy the top-tier architecture before you’ve earned it through the unglamorous work of data plumbing. Start with the plumbing, or prepare for the flood.
FAQ
Q: Isn't modern data architecture (like data lakes) designed to handle messy data so we don't have to clean it first?
A: No. Throwing dirty data into a data lake just gives you a polluted lake. The architecture doesn't fix human habits or broken encoding. If the data isn't trusted at the entry point, no amount of storage capacity will make it useful.
Q: How do I know if we're ready to move from Stage 1 (Data Assets) to Stage 2 (Value Chain Connection)?
A: You're ready when users stop treating data as a byproduct and start relying on it for daily decisions. If your team still prefers their old spreadsheets over your cleaned data models, you haven't earned the right to connect the chain yet.
Q: What's the contrarian take on digital twins?
A: Most digital twins are just expensive 3D models built to impress the board. They are practically useless until they are connected to live, trusted operational data. If your twin doesn't trigger a workflow, it's just a video game.