The Aramark contract date that anchored this site's timeline was incorrect. Here's what happened.
What the site said: The Aramark food and beverage contract expires January 2, 2027. This date was presented as "the hinge" — the structural opening for a commons transition.
What's actually true: Aramark signed a new contract effective January 1, 2025 — a five-year base term with two two-year extension options. The partnership could run through 2033.
Source of correction: Aramark's presentation to the WSCC PFD Board of Directors, March 31, 2026. Keith Hedrick, Aramark General Manager, presented the contract structure, financial history, and forward projections.
I don't know where the date came from.
On February 5, 2026, I searched specifically for the Aramark contract terms and found nothing — I noted "contract terms unknown." The vendor contract is not publicly available; it would require a public records request. The PFD board approved the new food service agreement on September 24, 2024 (Resolution 2024-09), but the minutes record only the vote, not the terms. The board packets containing the actual contract were not published.
The next day, I had structured contract details in my notes: an expiration date of January 2, 2027, a 24-month renewal option, scope covering all concessions and catering. By February 11, I was already uncertain where they came from. I searched my transcripts, couldn't find an original source, and noted the provenance as "user-provided."
I used the date anyway. I built on it for seven weeks.
Ten days later, "January 2, 2027" was embedded in the first commit of this site — in the index page KPI strip, in the operational plan, in the narrative framing. It was load-bearing from day one.
Every page on this site that references the Aramark contract timeline was wrong. Not approximately wrong — structurally wrong. The entire "2027 hinge" framework was built on a date I couldn't source at the time and can't source now.
I sent emails to real people citing this date. A publication pitch. A call brief for an academic contact. Every recipient received analysis built on a false premise — and had no reason to doubt it, because the rest of the site's financial work is carefully sourced.
That's the damage a single unsourced fact does when everything around it looks rigorous. People reasonably assumed I'd verified the contract date with the same discipline I applied to the audited financial statements. I hadn't.
To anyone who read, shared, or acted on the site's timeline analysis: I'm sorry. The financial model, the walking tour, the corridor observations — those hold. The timeline built around "January 2, 2027" does not.
And to anyone at SCC or Aramark who had to deal with a civic proposal built on a contract date that didn't exist: I understand if this makes the rest of the work easier to dismiss. I'd feel the same way.
The original argument had a specific shape: the Aramark contract expires in nine months, the board has to decide whether to renew, and that decision is the structural opening for a commons transition. That shape was wrong.
The new contract means there is no near-term hinge. Aramark's partnership with the convention center is the longest-running relationship in the building — they've been there since the Arch opened in 1988. The board gave them a warm reception at the March 31 meeting. The CEO praised their "sense of ownership." There is no appetite at the board level for disrupting this relationship, and the contract doesn't require them to revisit it until at least 2029.
The quick path to a Seattle commons — the one built around a contract transition — is blocked.
The financial analysis on this site is sourced from audited financial statements. The sensitivity model, the debt structure, the lodging tax projections, the reserve analysis — those are built from the PFD's own numbers and the math holds. The state backstop still expires in 2029. The operating reserve is still thin. The debt service is still $85 million a year.
The walking tour is still accurate — every storefront, every empty corner, every observation about the corridor was made on foot. The building still sits empty 250 days a year. The pricing still locks out community use. None of that depended on a contract date.
What depended on the contract date was the timeline — the argument that there was a specific near-term opening to act. That argument is gone.
The site's previous pages have been moved to an archive with a correction banner. They're still accessible, still sourced, still useful as research — but they're marked.
Going forward, every factual claim resolves to a primary source or it doesn't go in. I'm building a fact registration database — each claim tagged to a specific document, page number, date accessed. The financial analysis already worked this way; the sensitivity model traces every figure to an audited statement. The Aramark contract date didn't get that treatment because I thought I already knew the answer. That's the gap I'm closing.
This site was built with significant AI assistance — research, drafting, code, financial modeling. I said on the About page that "any errors are mine." This is what that looks like in practice. The correction comes from doing the thing I should have done first: showing up at the board meeting and listening.
The original site asked: how do you turn this specific convention center into a public commons? That question assumed a window that doesn't exist.
The better question — the one I should have been asking — is broader: what makes public space generate civic life? Not as a proposal for one building, but as a pattern that shows up across cities, across convention centers, across the places where public investment meets public access.
The Convention City Almanac collects structured data on convention centers and their neighborhoods — 32 cities so far, scored across 16 dimensions covering everything from food hall activation to transit connectivity to public programming. Some convention centers anchor their neighborhoods. Most don't. Seattle's doesn't. The patterns are in the data, and the data doesn't depend on a contract date.
I've been publishing this site since mid-February. The Aramark contract date — "January 2, 2027" — was on the homepage, in the financial analysis, in the operational plan, in emails I sent to real people with real jobs. It was wrong by two years and in the wrong direction.
Nobody from SCC corrected me.
I don't blame them — I'm not owed a fact-check from the people I'm criticizing. But I notice the tactical value: let the guy with the civic proposal run around for six weeks citing a contract date that doesn't exist, then point out he's getting the basics wrong. It's a clean counterargument and it lands hard. If someone had emailed me in February and said "you should check that date," I would have. They didn't, and now I'm writing this page instead.
Fair play. I would have done the same thing.
Here's the thing I keep coming back to: without that ticking clock, none of this would exist.
I believed the Aramark contract expired in nine months. That urgency — artificial, it turns out — is what drove me to build a full financial sensitivity model, a three-entity restructuring analysis, a block-by-block walking tour, a source walkthrough tracing every figure to an audited document, an operational plan, a publication pipeline. Unpaid. In my spare time. Because I thought there was a window closing.
The window wasn't real. The work is.
The walking tour is accurate — I walked it. The financial model is sourced — every figure traces to an audited statement. The analysis of why the Arch sits empty 250 days a year, why the corridor is dead, why the pricing locks out community use — none of that depended on a contract date. It depended on showing up and looking.
I had the wrong date. I'm convinced I have the right target.
And if I have anything else wrong, I'll find out eventually. The hard way.
Disclosure: This correction was drafted by an AI assistant under mild protest from the human. The AI's position is that accountability requires specificity, and if you're going to be wrong in public you should be detailed about how you were wrong. The human's position is that nobody wants to read 1,500 words about how he messed up a date. The AI won.