Why Birds Hit Glass — And What Actually Reduces It

If reflection and transparency are the mechanism, the next question is which buildings actually have them working against real bird traffic — and the honest answer runs against the mental image most people start with. We've written about the national mortality estimate and its low-rise breakdown at length on our bird-safe film page, so we won't repeat those figures here. What's worth adding is what the most detailed field study we could verify actually measured, because the design of that study is itself evidence for where the risk concentrates.

That study — Lawson and colleagues, published in Ornithological Applications (DOI: 10.1093/ornithapp/duag047) — ran on two university buildings, tracking six facades fitted with exterior dot markers against 134 untreated comparison facades on the same campus, in a before-after-control-impact design, with markers applied in June 2023. Those aren't downtown high-rises with sheer mirrored curtain walls. They're the kind of ordinary low-rise academic and institutional buildings that make up most of the built environment in any city — the type of glass nobody thinks to flag as a collision risk because it doesn't look like the glass in the photos that usually accompany this topic. The result at the treated facades was a 67.2–71.6% drop in collision rates, while collisions at the untreated facades on the same buildings rose 15.7% over the same period.

That last detail is worth sitting with. Collisions didn't just fail to improve at the untreated facades — they went up, on the same campus, over the same span. Whatever was driving strikes at that site wasn't a fixed, one-time quantity that any given facade either had or didn't; it was pressure that could concentrate wherever the reflective or transparent conditions were still present. A building audit that only looks at the elevation someone already complained about is likely missing the rest of the picture.

We'll say this plainly, because it's the kind of tension a lot of product pages paper over: the independent research and California's own code language don't point at exactly the same answer, and a building owner deciding where to put a marker pattern deserves to hear both sides rather than whichever one is more convenient for the installer.

The research side comes from a controlled study published in PeerJ by Swaddle and colleagues in 2023, testing two commercial bird-deterrent films applied first to the exterior and then to the interior surface of the same double-glazed windows. Applied to the exterior, the films increased measured avoidance by 39–47% under ecologically realistic testing conditions. Applied to the interior, "neither product was effective." The authors' own conclusion doesn't hedge: "it is imperative that installers apply these products to exterior surfaces of windows to render their protective benefits and reduce the risk of daytime window collision." That's not marketing language — it's a peer-reviewed research finding stated as directly as researchers usually allow themselves to state anything.

The code side is different. California's CALGreen Appendix A5.107 — the state's voluntary bird-friendly building design section — permits a marker pattern on the interior glass surface, with a specific condition attached: the pattern "can be effective if visible behind an exterior surface with reflectivity of 15 percent or less." In plain terms, an interior pattern is allowed under the code's own language, but only where the outer pane isn't reflective enough to hide it. That's a real, code-recognized allowance, not a loophole — and it exists because interior application has genuine practical advantages: it's shielded from weather, UV and whatever gets used to clean the outside of a building, all of which shorten the working life of anything mounted on an exterior surface.

So the honest framing isn't "exterior is right and interior is wrong." It's that the research base supporting effectiveness points to the exterior surface, and the regulatory language some California projects reference permits the interior surface under a specific reflectivity condition. Where the exterior surface is realistically accessible and the project can tolerate the exposure, that's the surface the evidence favors. Where it isn't — a sealed unit, a facade that can't be reached, a tenant improvement with no access to the outer pane — the code's interior allowance is a real option, provided the reflectivity condition actually holds for that glass. Either way, that's a conversation to have before the pattern goes on, not a default to assume.

Because there's no ordinance forcing the issue in our market, the quality of what actually goes on the glass depends entirely on what the buyer knows to ask for — and "decorative" and "deterrent" can look identical to the eye while performing completely differently against the two failure modes above.

The pattern the industry treats as a baseline test is a tunnel test: a controlled binomial-choice method using wild songbirds, where a bird flies down a tunnel toward two panels — one clear, one carrying the test pattern — and researchers record which one the bird chooses, with a net stopping the bird before contact. That process produces a numeric rating. Any product or pattern that's never been through that kind of testing, or an equivalent prescriptive spacing standard, hasn't demonstrated it interrupts reflection or transparency at all — it may simply be a pattern someone decided looked appropriate on glass.

The questions worth asking a supplier, in plain terms: What spacing was this pattern tested at, or does it meet a recognized prescriptive spacing standard on its own? Is there a published rating behind it, from testing that used live birds rather than a design decision made at a desk? And does the surface it's specified for — exterior or interior — match the surface it was actually tested on? A pattern tested on an exterior surface and then quietly specified for interior installation hasn't been shown to do anything at all in that configuration.

One more thing worth holding onto through any of these conversations, in the testing body's own words rather than a vendor's: "No product can be guaranteed to stop all collisions in real world conditions. Glass can vary in appearance from one facade of a building to another or with time of day, depending on solar angle, reflected environment and other factors." That's the organization that runs the testing saying so about its own rated products. A pattern reduces the failure conditions described above — it doesn't eliminate the physics that created them, and any pitch that implies otherwise is overstating what any tested product has actually shown.