Who Should Own the Digital Twin Model?
A digital twin model needs three functional owners and one coordinator. Engineering owns the model logic: layout, routings, resources and rules. Operations owns the floor data: cycle times, changeovers and shifts, and reports every change on the floor. Finance owns the cost rates and assumptions that go into investment decisions. A named model steward keeps versions, updates the base model when the plant changes and decides whether a result still matches the plant well enough to be shown.
Why a model stops matching the plant when nobody owns it
A digital twin starts out matching the plant. It was built for a decision, the team checked it against the floor, and the result was approved. Then the project ends and the people move on.
In an illustrative case, a plant builds a model for a new hall. Over the next year, two machines are replaced, the shift pattern changes from two shifts to three, and a new product family enters the mix. Nobody updates the model, because nobody has that job. Twelve months later, finance quotes the output figure from the old model in a new investment request. Operations says the floor no longer works that way. Engineering opens a file nobody else recognizes.
No one did anything wrong. Without an owner, the model turned into a record of how the plant used to run.
Split the ownership into four roles
Each function owns the part of the model it knows best.
| Role | Owns | Decides | Typical person |
|---|---|---|---|
| Engineering owner | Layout, routings, resources, transport rules, model structure | When the structure of the model changes | Process or industrial engineering lead |
| Operations owner | Cycle times, changeovers, shifts, buffer rules as run | Whether the floor data in the model is still true | Production manager or shift coordinator |
| Finance owner | Cost rates, assumptions used in investment cases, which scenarios support an approval | Which results can go into a funding request | Controller or finance business partner |
| Model steward | Versions, base model updates after plant changes, scenario names, summaries for management | Whether a result still matches the plant as it runs today | Engineer with protected time for the model |
The split mirrors how the plant already works. Engineering stands behind the structure, operations stands behind the numbers from the floor, and finance stands behind the language in the investment memo. When a result is questioned, everyone knows whom to ask.
IT often runs the data connections but rarely knows whether a buffer rule or a cycle time is still right, and that judgment is what the owners above are for.
The model steward: the role you cannot skip
The three functional owners make sure their parts of the model match the plant. The steward keeps the base model, its versions and its scenarios in order. Their tasks:
- keep dated versions of the model and its inputs,
- run refreshes when the plant changes,
- name scenarios so people can refer to them in meetings,
- write short summaries of results for management,
- stop an outdated result from being presented as current.
The last task needs authority as well as time. If the steward cannot say "this result is from before the line change, we need to rerun it", old numbers will keep appearing in new decisions.
The steward also needs protected time. As an illustrative figure, plan half a day a week between projects and more during a project. A volunteer who works on the model "when there is time" is the most common reason ownership fails.
When to refresh the model
Testing an option does not change the base model, because each option runs as a scenario on a copy of it. The base model is updated when the plant itself changes in a way that could change which variant wins. Typical triggers:
- the bottleneck moved or split between two stations,
- a machine, robot or autonomous cart was added, removed or replaced,
- a routing, batch size or picking rule changed,
- the shift pattern changed,
- the product mix or order volume moved well beyond the volumes in the model,
- an investment project changed transport routes, buffers or storage.
One trigger is enough to schedule a refresh. Not every change needs a full rebuild, though. A renamed station needs a note. A single cycle time that changed slightly needs a quick rerun. A new routing or a new machine needs a structural update and a new baseline. After a large investment, update the whole base model to the plant as built. That version is the starting point for the next decision.
Setting up ownership for your first model?
Who in your plant could act as model steward, and how much time could they protect for it? Start there with us and we will sketch a role split for engineering, operations and finance around a DBR77 Digital Twin model.
Change rules that keep the model trustworthy
Ownership works when changes follow a few simple rules.
- Every change has an author and a date. If a cycle time changes, the record shows who changed it, when and why.
- The last good version is kept. Before a refresh, save the current model and its results with the date and the decision they supported.
- Operations approves floor data before an approval meeting. No cycle time or shift rule enters an investment case without the operations owner's check.
- Each refresh ends with a short note. What changed, which results moved, which earlier decisions should be looked at again.
- Plant changes include a model check. When a change request moves machines, buffers or routes, attach a before-and-after comparison from the model.
These rules sound like paperwork. In practice, each takes minutes. They save hours in meetings where people argue about which number is right.
Who maintains the data: by hand or by sensor
Ownership also depends on how the model gets its data. With static data, the operations owner supplies updated nominal cycle times and shifts, and the steward enters them by hand or brings them in by import (import scope to be confirmed). Nominal times are enough, since the simulation uses fixed times and the same input always gives the same result. Variants with different fixed values show how much a result depends on each one. For how this differs from averages, see Averages vs simulation over time. Live machine data from DBR77 IIoT is the next stage, once the model is used every week and should follow the plant between those updates. Someone still has to own the feed: fix missing signals and decide what the model uses when data is late. We explain when each approach fits in When to Connect a Digital Twin to Live Data.
Signs that ownership works
You can tell within a few months whether the split is working.
- Investment requests refer to a named, dated scenario, not to "the simulation".
- Operations hears about model changes in a short note before they reach a board slide.
- Finance recognizes the assumptions in the memo because they agreed them.
- When a result surprises someone, they know whom to call.
The opposite signs are just as clear: two versions of the model on different laptops, numbers nobody can trace, and a steward who learns about a new machine after it is installed. Once ownership works, the next step is to reuse the model across projects, with a scenario library and a shared assumption register, as described in How to Turn a Digital Twin Into a Repeatable Decision Process.
FAQ
Should IT own the digital twin?
Usually not as the main owner. IT can run data connections and infrastructure, but the model's content belongs to engineering, operations and finance, coordinated by a steward.
Can an external partner own the model?
An external partner can build the model and help with refreshes. The decisions about floor data, assumptions and publication should stay inside the plant, or the model leaves when the contract ends.
How often should the model be updated?
Update the base model when the plant changes, such as a new machine, a moved bottleneck or a new shift pattern, and check that it still matches the floor before a major decision. The options you test run on copies, so they leave the base model as it is. A fixed calendar review, for example quarterly, catches changes nobody reported.
What happens when the model steward leaves?
If versions, scenario names and change notes are kept as described above, a new steward can take over without starting from scratch. If the knowledge lives in one person's head, the model usually leaves with them.
Conclusion
Name the steward first and write down the hours they get for the model. Then sit engineering, operations and finance down once, let each take its rows from the role table, and agree the refresh triggers before the next project starts.
See how a model is updated between projects
A refresh is easier to trust when every version is scored on the same measures. The demo shows how DBR77 Digital Twin keeps that comparison steady, so you can see what moved between two versions of a model.
Sources
- DBR77 Digital Twin, Averages vs simulation over time
Want to test a decision from your plant?
Describe the decision to a DBR77 expert and we will show how DBR77 Digital Twin compares the options on output, waiting time and transport load.
Questions
Have a question about this topic?
Tell us what it is about. We will reply within one business day.
Digital Twin newsletter
Short notes on testing layout and flow decisions on a model, and news from DBR77 Digital Twin. You can unsubscribe at any time.