Model a Maximo hierarchy change before you make it
A Maximo location hierarchy is cheap to change on paper and expensive to undo once work orders, PMs, routes and classifications hang off it. Model the before and after first, agree it in a room, then hand the Maximo team a structure everyone has seen.
Published 2026-10-09 · Theory only: general industry knowledge, no client data.
What this covers, and what it does not
Maximo is IBM’s product. The name belongs to its owner, appears here only so you can tell what this guide is about, and Strata is not affiliated with IBM.
Strata does not connect to Maximo or to a GIS, it does not migrate data, and it does not hold your records. It holds the design: the levels, the kinds of asset on them, what you record about each kind and why. That is where you agree a structure before anything changes in the live system. Menu names and screens differ between Maximo versions, so this guide sticks to ideas that hold across them: locations, assets, work orders, PMs and classifications.
Why a hierarchy change goes wrong
Work gets written at one level. The thing being maintained sits at another. In the Gas distribution template, a District regulator station contains a Regulator run, Overpressure protection and Station piping. A crew might write the work order against the station, while the part that is inspected, repaired and eventually replaced is the regulator run inside it. Whether that run is its own maintainable item, a location, or just a line of text in the work order decides what history you can report later.
That gap is what people argue about when they redesign a drilldown. The argument is always about a level: does this deserve a record of its own? It is expensive to settle after the fact, because work order history, PMs, routes and classifications all hang off the structure. Move a piece and you have to ask what happens to each of them. Multiply the answer by thousands of stations and a wrong choice becomes a project.
Most teams have this argument in a spreadsheet and find the problems after go-live. A better order is to model the before and the after, agree it in a meeting, and only then hand the Maximo team a structure everyone has seen.
Step 1: start from what you have
Export the current location or asset hierarchy from Maximo to a spreadsheet, one column per level. Copy the cells and paste them into Strata under Import, Text / Excel. Strata recognizes a wide sheet, merges rows that share a branch and fills blanks under a repeated parent, then shows you what it decided so you can correct it.
One caution. Strata models kinds of things, not every record. Paste one row for each kind of station and its parts, not seven thousand rows of station names, or you will get seven thousand classes. Record how many exist as a count on the class instead.
Before you change anything, export the current design as JSON and as an editable PowerPoint. That is your “before”. Strata keeps one working document, and it has no side-by-side comparison of two versions, so the saved files are how you keep both.
Step 2: build the proposed version
Now reshape it. Setup, Levels has a row for each level with free-text columns: a name, its function, what it corresponds to in Maximo, what it corresponds to in ArcGIS or the Utility Network, and the work performed at it. The last one is the column that settles the argument above, because it forces you to say where work is written. The gas template fills in the first two for each level, and two of them make a useful test for whether something earns its own level or record:
- Asset class: Assets alike enough to share a design life, an inspection regime and a cost model.
- Component: A part that is inspected, maintained or replaced on its own schedule.
If a candidate level cannot say which of those it is, it is probably a grouping and not a thing. Put data points where they apply to everything beneath: set one on a parent class and every class under it inherits it, so you are not redefining it across hundreds of locations.
Strata can also add up what a design implies. Set a count on a class, either a total or a number per parent, and the record projection sums the records at each level. Run it on the before and the after, and the cost of an extra level shows up as a number before it shows up as a data load. It is only as good as the counts you enter, so treat it as a sizing guide.
Step 3: put both in front of people
Walk the room through the design in Present mode, and export the editable PowerPoint for the review pack, with the before and the after as two decks. A share link works for the pre-read. It shows viewers the version you last published, and as the owner you can download earlier ones.
Take a short list of questions to every level:
- Do we write work orders at this level?
- Do PMs sit here, or on the parts beneath?
- Which classification applies, and what attributes does it need?
- What happens to existing history when this record moves?
The people who write the work orders usually see the problems first. Put them in the room.
Step 4: hand it over
Strata exports a nodes CSV, with an id, parent, level, purpose, count and path for every class, and an attributes CSV with each data point’s purpose, type, unit, current and desired system and the decisions it feeds. There is a JSON file too. Hand those to the Maximo team. Turning them into locations, assets and classifications, and loading them, is their work in Maximo’s own tools. Strata does not write to it.
A hierarchy that has been through this is easier to defend, because every level has a stated purpose and everyone has seen the alternative. For how an agreed hierarchy maps into the systems, read From asset hierarchy to EAM, CMMS and GIS. To try the exercise on a worked example, open the gas distribution template, which has the station classes used above.
Related guides
- What is an asset hierarchy? The pillar: levels, classes and instances, common patterns, and the mistakes to avoid.
- Hierarchy, EAM, CMMS and GIS Taking an agreed hierarchy into the systems, and keeping them aligned.
- Hierarchy vs. relationships Classification, connections and decisions are three different questions.
- Asset data glossary Plain definitions of the terms the templates use.
