#30 - More Rules for Great Facade Design
A quick design tune-up on a real project
In This Issue
More Rules for Great Facade Design
About a year ago, I wrote a piece called Three Rules For Great Facade Design. That piece was intended to provide a few design heuristics that could be applied at any scale (though all of my examples focused on larger scale buildings). It laid out the following principles for designing a nice-looking building:
Compose your building with a base, a middle, and a top
Make your building easy to sketch
Keep the “Big Moves” simple
This month we will implement similar principles on a real project. A friend recently sent me some schematic drawings for their missing middle apartment building in Toronto, and we were chatting about strategies for improving the design. With my friend’s permission, I’ll share this example with you below.
The principles I’m going to suggest are not firm rules. You can break some (or even all) of these rules and still create beautiful buildings. But if you follow these rules, you will avoid some of the most common mistakes among buildings which are not beautiful.
One note before we start: this is a very early design, and I assume the architect hasn’t really started tuning the facade yet. That said, early is exactly when you want to get the big organizing moves right — because once the plans and and big moves are set, “fixing it later” gets more complex and expensive.
Here are a couple of images showing the preliminary design (click the gallery to expand):
We’ll focus on the bottom of the building — the ground floor retail and the cantilevered second floor. This cantilever exists because city staff have asked for a setback at grade. Pulling the entire building back would make the units a bit too shallow, so we will take this constraint as a design challenge; our job is to find a nice way to address an otherwise awkward condition.
Here’s a closer elevation showing the area we’re looking at.
Make It Easy To Sketch
My first impression is that the various facade elements are not lining up, which makes it hard to grasp the organizing principle behind the design. To put it another way, it is not easy to sketch; if I asked you to memorize the design, you would probably struggle to draw it accurately from memory.
How might we address this? First, here’s a diagram highlighting a few things that jump out at me.
My orange rectangles at the second floor are equally wide — you’ll notice that the middle pier in the original drawing is a bit bigger than the others. This usually happens when the dimensions of the land aren’t easily divisible, and the architect had to put the extra building width somewhere.
My general process is to use facade dimensions that are divisible by 100mm or 4”. This is a module which works with most building components off the shelf (like bricks and windows). It’s also small enough that field deviations can be dealt with relatively easily. If there is an odd dimensional remainder, I prefer to split the difference and allocate it to the piers at the sides, which will usually benefit from the extra dimension more than the middle of the building will.
The next thing you’ll notice is that the window openings have inconsistent widths. I would make them all the same, which makes the facade a little more stately, less jittery.
Finally, my biggest reaction to the existing elevation is that there’s no alignment between the elements at the ground floor and those on the floors above. The red rectangles in my diagram represent where the brick piers would fall if they were consistent with the second floor. I find it easier to start with rigorously aligned elements, and to then deviate when absolutely necessary, rather than starting with an arbitrary composition and trying to make it look tidy later.
Here’s a new diagram reflecting those changes. The piers are all the same width, and continue down to the ground. This still leaves plenty of glass for the retail space (we will examine how to increase the glazing, if desired, a little later).
And here’s a sketch of what the new facade looks like. I’ve also lifted the sill of the second floor window up from the floor slab and added some soffit thickness to account for the floor thickness and insulation below, since the bottom of the cantilever is exposed to the elements.
This version reads much tighter, with less visual clutter. Everything lines up, and so every element (pier, window, spandrel) feels more intentional. You could sketch this from memory — four equal windows separated by five piers of equal dimension.
Base, Middle, Top
Next, let’s move to the treatment of the retail podium and the cantilever.
The existing design doesn’t make a big visual distinction between the retail and the apartments above, except with differing glazing dimensions. The cantilever offers a slight visual break between the two, but it’s almost as though the architect doesn’t want you to notice it — there is no purposeful edge or banding to tell you the retail is a different thing from the apartments. As drawn, this would probably be a steel shelf angle and a painted soffit underneath, which is structurally fine but very a utilitarian detail.
One way to improve this is to visually break the composition into a distinct retail base, using a different architectural language. A classic tactic is to add a decorative horizontal band or cornice. In this example, the first thing I’d do is wrap the soffit in a different material (blue in the sketch below), and use it to create a visual border at the bottom of the second floor.
A legible base, middle, and top is one of the easiest ways to makes a new building feel well-composed.
Big Moves = Simple. Details = Rich.
So far our design is, except for the cantilever, very simple. Essentially a box. This is important for reducing construction cost, and focusing our spend where it counts.
Next, I would add a corbel or bracket at each pier. The purpose of this is mostly aesthetic. People have highly-tuned intuitions about what “makes sense” structurally, and a design will feel right when you take these into account. The corbels represent the structural elements that would support the cantilever (the real structure is hidden). There’s a whole academic debate about whether this kind of architectural expression is good or bad. I want to avoid re-litigating that debate here, because it doesn’t impact how the average person values a building. But we talked about it previously if you want to go down the rabbit hole.
In any case, these decorative elements provide a pleasing, human-scaled rhythm to the facade, which was absent in the original design. They can also be used as a place to run wire, install lighting, and hang signage for the retail below. It doesn’t need to be expensive — cement board and wood are both great materials for this kind of thing.
Now, in some cases you will want to increase the amount of glazing for the retail space. That’s easy, and it’s often cost-neutral: less masonry, more glass. If you’re going to vary the widths of openings at a base, the classic architectural move is to change the material of the base so it looks like a somewhat different architectural language.
Here are some real world examples from a recent trip to New York. Notice that all deviations from the windows occur in the stone base.
In our case, you can build the base’s piers with the same blue cement board instead of brick. Then, if you make the windows bigger, it still looks coherent, as shown in the comparison below. If you did the same thing in brick, the piers would probably look odd — much too skinny.
When you build this kind of retail frontage with wood or cement board or even metal, you normally don’t want to use a super minimalist style, because these materials come in pieces with fixed dimensions. When you assemble them, you will see seams and joints all over the place. Seams and joints need spaces to allow for material expansion, and those spaces like to be covered with trim, which is partly why “traditional” panelized storefront styles emerged. You can fight this fact and try to make materials behave in unnatural ways, or you can lean in and have some fun with it.
Here I’ve added precast sills to the apartment windows to protect the brick from deteriorating from water running down the facade at the jambs, and I’ve painted part of the storefront columns in a contrasting color, roughly showing a panel-and-trim design.
You can take this kind of exercise as far as you like. I did these sketches on my phone, so they aren’t very precise, and don’t get into much detail. But I think a few simple moves like this can really elevate a design.
Below is a before & after comparison.
This isn’t really about taste or style — the principles I am sharing here can be expressed in any number of styles and materials. It’s more about creating architectural legibility, which usually results in buildings that people like.
Reading List
Mainstreets Development / Ontario Building Code Study. This report is a fascinating time capsule. In 1995, people were already proposing many of today’s vanguard reforms to Ontario Building Code, such as single-stair egress. Thirty years later, we have made almost none of these changes, and the province is undergoing a brand new OBC study to look for potential reforms. Implementation is really underrated in politics!
When a Start Isn’t a Start: The Problem with Canada’s Housing Data, from the Missing Middle Initiative. Canada counts a “housing start” unusually late (after excavation is complete and the building structure is back up to grade), which makes starts a very lagging indicator and muddies the public story about future supply.
Many Victorian Cities Grew Tenfold in a Century, from Works in Progress. Two hundred years ago, cities expanded at a speed that now feels politically impossible. This piece is a great reminder that “we can’t build” is a choice.
That’s A Wrap.
I write this newsletter because I like to connect with smart people who are doing interesting things. Reach out by replying to this email or commenting below.
Thank you for reading, and have a great month.