HarborLane Atelier is an architecture practice working on residential, hospitality, and small commercial projects across the Tampa Bay area. Every project publishes our design-stage cost estimate alongside the tender returns it actually received.
A client is told a project will cost around $1.4 million at concept stage. They proceed, spend a year on design, and tenders come back at $1.9 million. The architect explains market movement, scope creep, and material costs. All three are real, and the client still made a year of decisions based on a number that was 35% wrong.
Nobody tracks architects' estimating accuracy. The estimate is a conversation, not a deliverable, and by the time tenders arrive it has been superseded by so many revisions that comparison feels unfair. So estimating never improves and clients keep discovering the real number at the worst possible moment.
So every project we tender publishes our design-stage estimate against the tender returns. Our median variance is 8.4% over, and two projects have come back more than 25% over. Both are in the record with what we misjudged.
Illustrative projects. Bar shows variance from our design-stage estimate; centre line is on-estimate.
Straightforward. Local market, familiar detailing.
We underestimated the mechanical requirement in a historic shell. Our error.
Well within range.
Structural condition worse than survey suggested, and we should have tested more.
Close. Repeat client, repeat contractor.
Under estimate, which happens roughly one project in seven.
Seven architects and technologists in Tampa.
New homes and substantial renovations, with estimate accuracy published.
Restaurants, bars, and small hotels, including historic shells.
Offices and mixed-use up to about 30,000 sq ft.
Early-stage feasibility with a cost range rather than a single figure.
Site administration through to completion.
A fixed-fee read of whether your budget and your brief can coexist.
We re-estimate at concept, schematic, and construction documents, and publish the movement. A single concept number that is never revisited is how a client makes a year of decisions on a wrong figure.
Feasibility comes as a range with assumptions, not a single number, because a single number will be quoted back for twelve months.
A third of reality checks conclude the budget and brief cannot coexist. We say which one has to move rather than starting design and discovering it later.
Budget and brief tested for compatibility before design begins.
Cost range with assumptions at concept, re-estimated at each stage.
Returns compared against our published estimate.
Variance recorded per project, including the ones we got badly wrong.
Private clients, restaurant and hospitality operators, small developers, and owner-occupiers across Tampa, St. Petersburg, Sarasota, and the Gulf coast.
Fees on a percentage or fixed basis, stated at appointment, with the reality check available as a fixed fee first.
We decline roughly one enquiry in four, usually where a budget cannot deliver the brief and the client is not willing to move either.
Founded the practice in 2015. Publishes the estimate variance record.
Sixteen years in coastal and flood-zone residential work.
Thirteen years, including the historic shells that taught us hard lessons.
Eleven years on construction documents and cost checkpoints.
Their record shows a project that came in 33% over their estimate and says it was their misjudgement. I hired them because of that page.
The reality check told us our budget and our brief could not coexist. It cost a small fee and saved a year.
They re-estimated at each stage instead of letting the concept number ride. The tender was within 4%.
And what did the tenders actually come back at?