Cheesecake Math, Revisited

In an article about the Pike St. site of the former Cheesecake Factory, I did some arithmetic on a napkin and published it rather quickly because I wanted to make a hot-dog reference on National Hot Dog Day, which tells you what kind of operation this is. The claim was that the SCC Arch doesn't draw a steady enough crowd to make the corner across the street a viable site for a large-scale restaurant. I still believe it. But this Dispatch runs on an invitation — read it with your own AI, check my work — and the artificial intelligentsia pushed back that I had asserted a standard deviation and never shown where it came from. So I redid the math and found I'd been too easy on the convention center.

I missed it the first time because of confirmation bias: I walk that block, and Pike Street by the convention center is a commercial dead zone most days of the year — empty sidewalk, for-lease signs, a hot-dog stand that's closed most days and not much else. So when the numbers came back saying the corner draws almost nobody most days, they matched what my own eyes had already told me, and I filed them and moved on.

The problem was in a single word: attendance. The convention center's calendar lists an "expected attendance" for every event, and that number means two different things depending on the show — sometimes the crowd on a given day, sometimes the total who pass through across the whole run — and nobody tells you which. For a medical conference it barely matters: four thousand doctors register, the same four thousand are in town all four days, so four thousand is both the total and the daily crowd. The Northwest Flower & Garden Festival lists sixty thousand, but that's a five-day figure for a crowd that mostly turns over from one afternoon to the next — maybe twelve to fifteen thousand on the floor at once, not sixty thousand. Same word, wildly different meaning.

My first pass divided every number by the days it ran, as if every crowd rotated like the garden show's. That badly shortchanges a convention — a lot of Comic Con's badge-holders hold weekend passes and are on the block Thursday and still there Sunday. Count the full crowd every day and you make the opposite mistake, since plenty do buy a single day. The truth sits somewhere in the middle. And here's what matters: it doesn't change the answer. Count the crowds the stingiest way — everyone comes once — and the day-to-day swing is still 2.4 times the average; count them the most generous way, everyone every day, and it's 2.1. Wherever the real number falls, the swing runs past twice the mean. So I used the generous count — the most favorable reading the convention center could ask for — and the corner still can't hold a Cheesecake Factory.

The visitor draw averages about twenty-two hundred people a day. The standard deviation — the size of the daily swing — is about forty-eight hundred, more than twice the average itself. More than half the year, the public calendar gives a restaurant nothing to plan around. Then a comic con drops twenty-five thousand people onto one block. That is not a number you staff a factory against. It's a number you meet with a hot-dog cart. Shout out to DEEZ DOGZ, which can move five thousand dogs on a Comic Con day.

You cannot lease your way out of a standard deviation like that. But that is the pitch on the convention center's own empty storefront: "Prominent Foot Traffic. Prime Visibility. Promising Potential." Prime visibility, sure. The foot traffic is the variable in question.

None of this means Pike Street can't hold a restaurant. Two blocks up, at 9th and Pike, the smaller-format Daawat does fine. It's the format that fails, not the block — 7th and Pike, the Arch's own frontage, closest to the hotels and the downtown core and emptiest of all, is exactly where a big box starves, because the crowd it lives or dies on isn't steady downtown trade, it's the convention spike, in town maybe a third of the year. Daawat is sized to a neighborhood. The Cheesecake Factory was sized to a crowd that only really shows up for Comic Con.

The notice in the window thanks Seattle for many years of loyal patronage and then points you to where the format still works: Alderwood Mall. Bellevue Square. Southcenter Mall. Three suburban malls — steady, captive, a crowd that never spikes to twenty-five thousand and never drops to zero.

So what's the answer? Do we turn the Arch into a mall with wall-to-wall retail to try enticing people to Shop at the Arch, achieving critical mass for a large-format restaurant to reopen? Is that plausible? Is that our future? It sure looks like the plan. (See The Campus Master Plan: What's Been Decided.)

We already see the cost of the status quo: empty streets, vacant storefronts, and expensive architecture sitting idle.

But we can do better.

Imagine every day a "golden day" with 5,000 to 7,000 visitors coming to the Arch, whether for events on the public calendar, the private calendar, or the space-available calendar.

What would it cost to program the Arch in every empty space on every empty day?

I don't know, and I'm out of napkins.