Why Hasn’t Building Simulation Scaled?

Part 2 – Defining the problem – the 3 T’s (Technology, Talent, and Time)

Since building simulation carries so much value for the lifecycle of a building, why hasn’t building simulation become more widely adopted?  

I’m far from the only one asking this question.  Over the years, I’ve heard plenty of answers: too complicated, too much time, too difficult to learn, it’s not required, it costs too much, my client isn’t asking for it, and so on. 

I believe each of the myriad of roadblocks to adopting building simulation more broadly fit into three (alliterative) categories:

• Technology

• Talent

• Time

They aren’t in any particular order; like the legs of that three-legged stool, they all matter and depend on one another.  Scaling simulation requires addressing all three.

These three categories are the root causes of a commonly cited reason for not doing simulation: clients don’t see the value.  And if clients don’t see the value, then simulation won’t scale.  

Over the course of this blog series, I’ll address each category individually at first, then discuss the critical interrelationship among the three, and propose a solution for addressing all three.

Symptoms that arise from the underlying causes

“My client just isn’t asking for it.” (and therefore it doesn’t have value)

This gets framed in a few ways – but this is probably the most common framing I hear.  But is something only valuable on a project when the client asks for it?

While I’m not an architect, I’ve worked with many who spend a lot of time guiding their clients toward a different decision than the one they initially thought they wanted.  That includes suggesting features in a building the client didn’t initially ask for.  And honestly, when has a client not asking for something ever stopped a good architect (or good engineer) from proposing a great idea? 

Value of simulation in design (and operation)

Value can be defined as talented people using the right technology to deliver critical insights at the right time.  While that “equation” might not show up as a neat line item on a fee proposal or balance sheet, being able to do exactly that is how many building professionals sell their services.  And scaling value is something every client should want for their projects and every design team should want for their clients.  

In one study, modeling fees were compared to the energy cost savings achieved on a host of projects varying in size and type.  For most projects, the modeling fee paid for itself in one or two months.  The energy efficiency measures themselves typically take years to pay back.  By this measure, the value of building simulation is readily apparent.  And that was with the “traditional” modeling tools and practices of the time. 

Fixing it all is harder – and there’s no trendy buzzword for it

This isn’t as elegant as saying things like “democratize energy modeling” or pointing at the highly siloed software tools and associated learning curves as the only things holding simulation back.  Solving only one of the 3 T’s doesn’t solve the whole problem, and can actually be counterproductive.

Simpler tools aren’t a cure-all, either.  Tools with limited inputs are often harder to troubleshoot because they expose fewer of their assumptions to scrutiny.  Getting back to George Box, if modelers can’t see and understand how the assumptions are working in a model, no good modeler would ever trust the results.  It’s a part of every good modeler’s evaluation of a software tool: how well can I understand the inputs and other associated limitations?

“Democratizing” building simulation

To paraphrase Socrates: When you’re sick, do you “democratize” your health care by asking the masses to vote for a treatment or do you seek the advice of a doctor and other medical professionals?

While you could let the masses vote for your treatment, a medical professional is much more likely to get you a positive outcome than the “democratized” masses.^[1]

The point isn’t whether democracy is inherently good, bad, or otherwise; the point is that wisdom and learning have value when it comes to making investment decisions that will last for decades.  Bypassing that wisdom just to add capacity isn’t a wise investment or a sustainable path forward.  Using any tool, simple or complex, without knowing how to judge its results just adds uncertainty.  Adding more uncertainty to the process has negative value.

Another way to think about “democratization” in the name of more throughput: if there aren’t enough architects to design buildings, should we “democratize” architecture by making it so everyone is an architect?  Or maybe just do away with the Architect Registration Examination (ARE) so it’s easier to have more architects?  Wouldn’t that inherently undermine the value of architecture?

In Part 3, we’ll unpack the 3 T’s a bit deeper, starting with Technology.

^[1] Levine, D.M., Tuwani, R., Kompa, B., et al. (2024). “The diagnostic and triage accuracy of the GPT-3 artificial intelligence model: an observational study.” The Lancet Digital Health, 6(8), e555–e561. Across 48 case vignettes, practicing physicians listed the correct diagnosis in their top three 96% of the time, versus 54% for a nationally representative sample of 5,000 lay people who could search the internet. https://doi.org/10.1016/S2589-7500(24)00097-9