Don’t make something unless it is both necessary and useful; but if it is both necessary and useful, don’t hesitate to make it beautiful.

-Shaker Philosophy

In yesterday’s post, I started looking at the question of “How much should I model in ArchiCAD?” If you haven’t read that yet, start there. Let’s continue with that discussion, but this time focusing on “What Should I Model” and “What Should I not Model?” And again, these lists aren’t exhaustive, but are meant to get us all thinking about the complexity of this fundamental question.

What Should I Model?

  1. An old rule of thumb from gurus on the forum is this: When deciding what to model, don’t base that decision on size, complexity, or drawing scale but on how many times you see the element in question. The more times it is seen, the higher the probability it should be modeled. In general if some element is shown two or more times, model it. If it is once or zero, it’s debatable (more on that later). Remember, when doing this analysis of whether to model an element based on number of times seen, an element might show up in a plan, section, elevation, 3D view, interior elevation, schedule, detail, etc.
  2. The answer to this question varies by project type. Baseboard is a great example. For high end custom residential, I always model it. For some other project types, say hospitals or airports, I might not. You might be thinking, isn’t that bad BIM? Doesn’t the hospital or airport contractor need that information just as much as the custom residential team? Of course. Or maybe more so. But in a project with hundreds of rooms, is there some other way that information can be documented? For instance, instead of modeling all the baseboard, could Zone Perimeter be used as a good proxy? Perhaps. The point here isn’t to find the right answer, but to remember that data doesn’t have to also mean 3D representation. Often we can just add text based data to less complex model elements to convey as much or more (this is one of the reasons I am so interested in IFC properties these days).
  3. The answer to this question varies by output type. What are you producing in the end? Traditional documentation? Something more advanced? A few schedules? Tons of schedules? A BIMx model? Is there one piece of documentation that is the driving factor? If your primary point of communication with your client is a BIMx Docs model, then model for that first. If your goal is quantity take-offs make sure you model in a way that supports that, even if it means less glamorous 3D views or “uglier” 2D views (not that either of those are a given). Depending on your primary output, visual 3D perfection might not be the goal.
  4. The answer to this question varies by collaboration requirements. Are you exchanging IFC with your engineers or other team members? Are you just doing DWGS, PDFS, or smoke signals? Depending on what information you are sharing with others, and in what format, you might decide you need to model more or less. For instance if you are working with a structural firm that will be modeling their work, then you shouldn’t be modeling any of that. Of course if you are going down that path, you’ll need to spend time COMMUNICATING those decisions and making sure the data exchange works. Here’s some advice for that.There was a reason to model all of this

What Should I not Model?

  1. Just because you can, doesn’t mean you should. Or need to. Will not modeling something make your life easier? Stefan Boeykens wrote a post a few years back entitled Improve usage of BIM during early design phases. All his advice is great (and pertinent to this topic), but it’s his sixth point that I want to highlight: “Model as little as possible…” What if we turn things around and look at the “How Much I Should Model” question from the reverse. “How Little Can I Model?” When trying to decide what to model, assume nothing should be modeled and find proper justification FOR modeling. Some things are easy (Walls, Windows, and Doors in every project of every scale). But as we dig deeper, the answer clearly becomes more complex.
  2. If 3D is slower, unnecessary, or harder to manage, don’t model it. There are things in much of my typical work that aren’t shown more than once. Most electrical, some structural, and a lot of site plan info, for instance. In those cases I do whatever is fastest—sometimes using a tool that results in 3D data is faster, sometimes it’s slower. But while I say this, I also know my work is going through a transition. I used to use 2D lines and groups of elements to show most of my electrical info. Now, even if I’m turning off the 3D model, I try to use Objects that have a 3D component for all my switches, outlets, and light fixtures. I am preparing to show this information. Or at least to be able to turn it on if I need it. Or if I’m not using 3D elements, I’m drafting everything with Fills because Fills can have IDs and can be scheduled. That is a forthcoming post. This technique allows me to ignore the 3D geometry but still get the benefits of model once, show multiple times (in this case showing a light fixture in both plan and schedule). The important point here is 3D isn’t always better. It usually is, but it isn’t always true.
  3. When you have run out of time, money, or knowledge. There are instances where you can model up to a point, but then need to leave a void. Two years ago I was working on a townhouse in the outskirts of Houston. I’ve shown the model before (here and here and here and elsewhere). It’s a hyper detailed model. We modeled things I’d never recommend modeling (when a client is also an architect and a friend, this sort of thing happens). When it came to deal with the curving, grand staircase, I hit the edge of both my ability and the capabilities of ArchiCAD 16. I started modeling the stair and it was awesome. But then I hit glitches and it was going slow because I didn’t know the Morph Tool well. And I was finding aspects of the design that I just wasn’t sure about. So I stopped modeling. I built the treads out of slabs because that was easy (and valuable) and just faked the rest. Some images are below. The rest of the section is all modeled (except for the framing, which was easy to do with the 2D Lumber Object). But that stair is a huge lie. It doesn’t matter. Our drawings weren’t going to be the basis for the construction of that stair anyways. It was going to be built in situ by experienced craftsmen who had done many stairs like it before. We’d modeled the rest of the building well enough and did the math on the stair to know it’d work fine. So I just faked the drawings and moved on with my life. It sadly meant the 3D model couldn’t be used for walk-throughs without explaining the incomplete stair, but that wasn’t the end of the world. And to be honest, drafting that stair felt so good. It reminded me of happy moments as a young intern, in my early ArchiCAD (and pre-ArchiCAD days), when creating a stair section meant making a half-lie thing of beauty out of lines where not really knowing (or needing to know) the ramifications of every single one of my actions was okay.
  4. When no one cares. Let’s be honest. Sometimes the only person who wants to see something modeled is the person sitting in front of ArchiCAD. Sometimes the reason we want to model a certain thing (or add a piece of extra data) is because it’d be fun or a good challenge. But that’s no always valuable to the client, the project, or the budget.  It’s up to us to remember to ask the painful question of “am I fussing over this little bit of model for my own pleasure?” If the answer is yes, think about the benefits of not modeling.

Faked Stair Section Model of Stair

What Else?

I feel like these two posts covered a lot of this topic, but there’s still so much more. What do you think? What did I miss? How do you decide what to model or what not to model? What examples do you have from your own experience with this question?

Are you following Graphisoft North America on Twitter? Click Here to keep track of all the latest ArchiCAD News in North America (and beyond).


  1. Eduardo

    Some other modeling rules are:
    • Have I done this before? If not then model it, you might be surprised at what you learn.
    • Is this an object critical to the project (for example your stair in another situation). You will be surprised how many times that 1 drawing becomes 10.
    • Anytime you say “F**K it Fake it and forget about it” or “Leave it as is and we will fix it later” always come back and model it as soon as possible. If not you will find 6 months down the road saying “if I had only model that you would not be having the current problem on the project”
    On another note;
    • ALWAYS organize your model so that you can figure out what you did 6 months or years later when you open the file.

  2. Jared Banks

    -Great advice. I really like “have I done this before?” Brings this concept back to a design perspective. A big reason to model is that it helps design. So if you haven’t done something before, it probably means you need to focus on it a little closer.

    -Is this critical to the project: totally. But here’s a corollary: is ALL of this critical to the project? My judgement with my crazy stair was that the curved underside wasn’t critical (as the exact curve wasn’t going to affect anything else). But that was a judgement call based on knowledge of the project, the builder, lack of ability and lack of time. So yeah it was a bit of a risk at the end of the project. I definitely don’t want that example to be construed as advocating finding reasons not to model something. More that it’s okay if the time/cost gets out of whack to walk away from perfection, if you are making the decision to do so through strength and understanding.

    -“Fake it! Leave it and fix it later” are traps! I probably should write a third post entitled something like:

    If you catch you or your coworkers saying this, step back and ask if you are being stupid.

    “It doesn’t matter if it’s modeled accurately”
    “Yeah that should be out a 1/2 inch but…”
    “We ignored the trim”
    “I don’t like modeling finish floor”
    “to get the numbers right we had to…”
    “I know it’s modeled wrong, but…”

    • Jared Banks

      Oh and yes, yes, yes, yes. An organized model (layers, views, attributes, etc.) makes a world of difference (why I am Template crazy). Again a post unto itself!

  3. Eduardo

    …and when opening an old project Never, Never, Never! change something that looks like a mistake until you understand or remember Why it was done that way.
    • This is based on the assumption that you always model correctly. I have had 2 projects were I committed this mistake, the first one were I learned the rule and the second one were I had to kill an underling for forgetting it.

    • Jared Banks

      thanks for the laugh! How about…never make changes to an old file before making sure you have a safe, untouched copy of that file.

      Anytime I restart an old project (or make big changes), I save a copy in a folder labeled “DATE Project Before Changes”.

      • Eduardo

        Always make a copy and all of them are Name-Phase-Rxx where Phase is Design Dev, Construction Docs, etc. and XX is the new version number and every exported DWG, PDF or BimX starts with YY-MM-DD_Name.
        I teach my students that at a minimum every Monday they should always do a “Save As…” of their projects so that they can always travel back to a previous version when the Crit does not go their way.

  4. James

    This probably seems obvious, but something I used to have to remind myself. The whole model doesn’t have to have the same level of modelling at the same time, if an area is changing or not quite resolved or that area you’re about to have a meeting about, then model only enough to get the message across and discuss with, because I usually had to heavily modify it afterwards – so the more elements or more complex that area of the model was, the harder/slower it was to edit/change.

  5. Patrick May

    i remember modeling a winder stair similar to you images above before the morph too existed. It had a square riser face though, and it was for exactly the reason stated in the comments; because I hadn’t modeled one before.

  6. Jason Smith

    Hi Jared

    I agree with point 1.
    But I do get carried away sometimes, it may take longer to do but it will usually save me time by resolving the issue before it gets to site.

    Another consideration is, what is the end result. Does the builder need that view, schedule, elevation etc.

    Sometimes it has to be done to see if it can be done, That’s Experience.
    The element or elements or even the process to create it can be used again in less time. Templates with favourites is a way to get the mundane elements completed in quick time so that you can model the more complicated elements.

    There are too many ways of doing the same thing.
    What is the best way of modelling that element/s?

  7. tom pansing

    the better you get at modeling the more you can do each project.
    your skill set, template, pre-prepared modeling settings, etc. at some point will allow you to model (almost) everything in a timely way. better budgets, better projects, better clients also help to provide financial resources to build as much as you can.

  8. Daniel Lindahl

    Another thought: excessive modelling on a very simple project can put you in the awkward position where the client thinks “is this what we are paying you for?”

    That takes it back to only modelling what is necessary for the required output.

  9. Jared Banks

    So many awesome comments! Thanks everyone. Keep them coming.

    Tom, absolutely the better you get, and the better your template becomes, the more you can model. And the more you should.

    Daniel, oh man is that true! How much to model is definitely about managing client expectations! You never want the client to think you’re wasting their time/money. We have to get good at explaining that sometimes (especially as we get more experienced) more detail is faster and easier than less detail. But at the same time this is a great double standard. No client will ever say “why did you draw smoke coming out of the chimney and clouds in the sky?” when looking at a hand drawing. Perhaps it’s because clients have a concept of time spend on a drawing while they don’t for a model.

  10. Alexander Ståhl

    Might be something obvious but I find continous communication is essential in this matter. When working on larger projects using teamwork for example it is important to inform, remind and clearly explain in a pedagogical way to your coworkers about how to work with the answers to these questions.

  11. Ian Booker

    I’d like to throw in a googly here. Up until now, architects control what the client and/or builder sees by controlling the display of information (plans, sections, details,etc). We generate extra information when requested.
    BUT the time is fast approaching when builders will expect not drawings but 3D virtual buildings that they can slice and dice as they want. What then of the level of detail?
    The latest BIMx (unfortunately costly) version does this to a certain extent, as I have recently found out.
    Still, look on the bright side – no more time consuming annotations?
    * googly is a cricketing term

    • Jared Banks

      Ian, great question. I think, for me, the answer is more consistency and clarity. The models we give don’t need to be completely detailed, but it needs to be obvious where the level of detail stops. So show all insulation but no water proofing. Or windows but no nailing flanges. Whatever it takes to make the distinction between “not shown/modeled” and “not in the project”.

      To that end, I think the answer also lies in Building Materials. The more we model based on properly set up Building Materials (and subsequent Composites/Complex Profiles), the easier it’ll be to raise the level of detail in a model AND the consistency of that detail. I definitely know my models have become much more ‘real’ and accurate since switching to the building material paradigm.


Submit a Comment

Your email address will not be published. Required fields are marked *

Work from anywhere with secure, real-time collaboration, thanks to BIMcloud!   Find out more