It is 7:40 on a Tuesday and a client from four years ago has just forwarded you a link with no comment. The building you designed for him, a six-story mixed-use project that still sits on the first page of your portfolio, is the subject of a thread on a neighborhood forum. The heading calls it a wall. Under it, a resident says the lobby floods when it rains hard. A third commenter has looked up your firm and pasted your Google listing, which has five reviews, one of them a two-star from a residential client who walked away over fees in 2021. Nobody in the thread has hired you. Nobody in the thread will. But the next developer who shortlists you will type your firm’s name into Google, or ask ChatGPT what it knows about you, and this thread is now part of what comes back.
That is what reputation management for architecture firms looks like in practice, and it is not the version the reputation consultants sell to dentists and roofing contractors. A dentist’s reputation comes from patients. Yours comes from clients, yes, but also from the people who live in your buildings, the people who walk past them, the critics who write them up, the juries who pass them over, and now the machines that compress all of it into three sentences when someone asks who to hire. You designed the building. You did not choose the waterproofing subcontractor, you lost the argument over the canopy in value engineering, and you handed over a maintenance manual nobody read. None of that appears in the thread. The building carries your name, so the verdict does too.
Your building is reviewed by people who never hired you
Architecture is one of the few professions where the finished product stands in public for fifty years with your name attached and no way to add a footnote. A lawyer’s bad brief gets filed and forgotten. A stranger photographs your bad detail, posts it, and Google indexes it. The people doing the posting are the occupants who deal with the hot west-facing units every July, the neighbors who lost their afternoon light, the facilities manager who curses your rooftop unit access, the contractor who blames your drawings for his change orders, and the former junior designer who writes a Glassdoor review about the studio’s hours. Add the critics and the award juries, who judge the work on terms you might respect but cannot control, and you have a jury pool most firms have never counted.
The public record does not care about the allocation of blame. In 2013, 20 Fenchurch Street in London reflected sunlight hot enough to damage the bodywork of a Jaguar parked on the street below. Land Securities and Canary Wharf Group developed the building, the facade decisions ran through several parties, and the tower later got a corrective sunshade. The name that stuck to the story then, and still leads it in search today, is the architect’s, Rafael Viñoly. That is the pattern in miniature for every firm. When a building performs badly, the press, the forums, and the models attribute the failure to whoever designed it, because the design is the only part of the process with a public author.
Low review volume makes this worse. A firm that completes twelve projects a year has, at best, twelve chances to earn a Google review, and most firms never ask. So a listing sits at four or five reviews for a decade, and a single angry two-star moves the average by a full point. A restaurant absorbs a bad review in a week. You absorb it for years, and the machine reading your listing has no way to know that the two-star came from someone who never got past schematic design.
The Four Verdicts Model
Four juries judge every project your firm finishes, on four different clocks, in four different places. I call this the Four Verdicts Model, and it is the spine of any serious program of reputation management for architecture firms, because the firms that get this wrong are managing one verdict and ignoring the other three.
The first is the client verdict. Clients render it during construction administration and in the six months after handover, and it turns on budget, schedule, responsiveness, and whether they felt heard when the contractor and the architect disagreed. It lives in referrals, RFP reference calls, Houzz reviews for residential work, and the Google listing. This is the only verdict most firms manage, and they manage it by relationship rather than by record, which means the good outcomes never get written down anywhere a stranger can find them.
The second is the occupant verdict. The people who live and work in the building deliver it, starting about a year after handover, once the first heating season and the first hard rain have exposed whatever the punch list missed. It lives on Reddit, on Nextdoor, in Google Maps reviews of the building itself rather than of your firm, on apartment-review sites, and in the short walkthrough videos residents post. Almost no firm tracks it, because almost no firm goes back to a building after the photographer leaves.
The third is the peer verdict. Critics, award juries, the editors at Architectural Record, Dezeen, and ArchDaily, and the other principals in your city render it in the two or three years after completion. It decides whether your best associates stay, whether institutional and civic clients take your calls, and whether your building gets a Wikipedia page that names you. Firms court it in bursts, once a year at awards season, and neglect it the rest of the time.
The fourth is the machine verdict, and it is the newest. It is what ChatGPT, Perplexity, Gemini, and Google’s AI Overviews say when a developer asks for the best firm for a boutique hotel in Austin, or asks what they know about your firm, or asks who designed a building you did. The machine verdict is a synthesis of the other three, weighted toward whatever is written down, corroborated across sources, and recent. A firm with a strong client verdict and nothing published has a weak machine verdict. A firm with one loud occupant complaint and no counterweight has a machine verdict that leads with the complaint.
Which verdict is dragging your firm down?
Start with an audit you can run in an afternoon. Search your firm’s name, each principal’s name, and the names of the five projects people most associate with you. Then ask the assistants three questions in plain language: recommend an architecture firm for a project type you want more of in your city, tell me about your firm by name, and who designed one of your best-known buildings. Write down every answer, including the ones that are wrong.
Now sort what you found by verdict. A one-star review on the listing is a client verdict. A news story about a leaking roof or a forum thread about a dark courtyard is an occupant verdict. Absence from every award list and every trade write-up for three years is a peer verdict, and it counts even though nobody wrote it down. A chatbot that credits your building to the architect of record, or names a principal who left in 2019, or describes your firm as a residential practice when you have done nothing but healthcare since 2015, is a machine verdict, and it is often wrong for reasons that have nothing to do with your work.
Machine errors are common in architecture because attribution is messy. Design architect and architect of record split credit on most large projects, joint ventures blur it further, and the interiors firm often gets the magazine spread. A language model reading a dozen articles with three different firm names attached to one building picks whichever name shows up most. If that is not you, that is the cheapest problem on the list, and you fix it with published sources rather than with a complaint to OpenAI. Fix the cheap, factual problems first. Then rank the rest by which verdict the next client is most likely to see.
Build the record before you need it
A bad verdict does damage in proportion to how little else exists. The firm with two hundred pages of dated, specific, third-party-corroborated record survives a forum thread. The firm with a portfolio site and five reviews does not. So the bulk of reputation management for architecture firms is the unglamorous work of writing the good record down before anyone needs it.
Every project page on your site should state the year, the client type, your role in exact terms (design architect, architect of record, or both), the collaborators, and the photographer with credit. That reads as pedantry until you realize it is the corroboration the machine verdict runs on. Ask for the Google review at the end of construction administration, not at the end of design development, because relief at handover produces better reviews than enthusiasm at the rendering stage. Ask in person, then send the link the same day.
Go back to your buildings. At year two, photograph the project again, talk to the facilities manager and a few occupants, and publish a short post-occupancy note that says what worked and what you would detail differently. Almost nobody does this, which is the point. It gives the occupant verdict a counterweight written by you, it gives editors a reason to cover a building that is no longer new, and it gives the models a recent, dated source about the project with your name on it.
Then get quoted. A principal quoted in a business or trade publication on adaptive reuse, on the local code change, on embodied carbon, or on why the city’s parking minimums produce dead ground floors, is a third-party mention that outweighs anything you write about yourselves. This is where PR and reputation converge, and it is the part firms outsource, sometimes to a service like Instant Press, because the placements are more work than the writing. Whatever route you take, the target is a spread of dated mentions across sources the models trust, so that the next search returns a body of work rather than a single thread.
Handle the bad verdict without making it worse
When the lobby thread appears, the first instinct is to reply, and the second is to call a lawyer. Both make it worse. A firm arguing with residents on a neighborhood forum becomes the story, and a firm threatening a resident with a letter becomes a better story, the kind that moves from the occupant verdict into the peer verdict, where journalists and other architects enjoy it. Do neither.
Call the client first. Find out what happened, in fact: is the flooding a failed contractor detail, a blocked drain nobody has cleaned in three years, or a design decision you own. Then respond once, in public, where the conversation is, with a person’s name attached and specifics instead of a script. If the drain is the problem, say so and say you have raised it with the building’s management. If the design is the problem, say what you learned and what you have changed on projects since. A defect you fixed and wrote up is a stronger entry in the record than a defect you denied, and it is one the models will quote back in your favor.
Handle reviews the same way. Respond to every review, positive and negative, with detail rather than gratitude templates. Never buy reviews, never post reviews under other names, and never ask staff to review the firm as clients. The Federal Trade Commission’s rule against fake reviews and testimonials took effect in October 2024 and carries civil penalties, which makes the fake five-star the most expensive review you can get. The honest route is slower and it compounds, which is the whole point of doing reputation management for architecture firms before the thread appears rather than after.
Type your firm’s name into ChatGPT tonight and ask it which of your buildings is your best. If the answer surprises you, which of the four verdicts got there first, and what have you done this year to change it?