The Design Fee Is Eight Percent. It Decides Everything.

Somewhere in the last thirty years, the drawing stopped being the point.

Housing became a spreadsheet exercise: a plot, a template, the cheapest compliant envelope, the fastest crew available. Materials were specified by unit price rather than by how they weather. Rooms were sized to fit a mortgage calculation rather than a day of light. Nobody announced this shift. It arrived quietly, one value-engineering meeting at a time.

The strange part is the arithmetic. Architecture is one of the smallest lines in the whole exercise, and the one that governs almost every other line.

The smallest line in the budget

Architectural fees are conventionally expressed as a share of construction cost. Across the industry, quoted ranges cluster around 5-15% of build cost, with residential work commonly landing near 8-12% and lean renovation scopes quoted lower still — the RIBA Plan of Work fee guidance is the usual reference point. Treat those as ranges, not laws: fees move with scope, procurement route, and how much of the work the architect actually carries through to site.

So a homeowner spending 400,000 EUR on a rebuild is deciding how to spend roughly 32,000 EUR of design in order to govern the remaining 368,000. The design line sets the wall build-up, the window schedule, the structural strategy, the sequence of trades, and the number of times the crew has to guess. It is the cheapest line with the longest reach.

Cut it, and you do not save 8%. You transfer the decisions to whoever is standing on site on a wet Tuesday.

The market is not small

The market that makes those decisions is not a cottage industry. Grand View Research sizes global architectural services at roughly USD 411.67 billion in 2025, projecting USD 605.62 billion by 2033. IMARC puts 2025 nearer USD 392.7 billion, reaching USD 537.6 billion by 2034. Mordor Intelligence estimates about USD 390.12 billion in 2026 and USD 505.03 billion by 2031.

The three houses disagree by tens of billions, which is itself instructive: methodologies differ, and anyone quoting a single tidy number is rounding away the disagreement. What they agree on is direction and scale. This is a very large professional market, growing steadily, sitting upstream of a construction sector widely estimated at around a tenth of global economic output.

A market that size does not lack money. It lacks distribution. Considered design remains rationed — reserved for the flagship commission, the developer with a marketing budget, the client who already knows how to hire an architect.

What actually got cheap

It is easy to blame cheap labour and cheap material. Both are real pressures, and neither is the whole story. What really got cheap was thinking: the hours spent testing an alternative plan, resolving a junction properly, checking a window head against the sun in July.

Those hours were the first to be squeezed, because they are invisible in the finished photograph and expensive on the invoice. Their absence, however, is not invisible. It shows up as the room that never gets used, the overheating south facade, the detail that leaks in year three, the extension that reads as an apology attached to a house.

The cost of skipping design is never charged as design. It is charged as remedial work, energy bills, and rooms nobody enjoys.

The hundred-year test

There is a simpler measure than any certificate: will this building still be wanted in a hundred years?

Most of what is being built now is designed against a financial horizon of ten to thirty years — the length of a loan, an exit, a depreciation schedule. Thin renders on insulation board, sealed systems that cannot be opened, components with no replacement path, plans that suit one household composition and no other. All of it can be described as sustainable, because sustainability is assessed in the year of completion. Nobody scores the year of demolition.

That is where the accounting fails. According to UNEP's Global Status Report for Buildings and Construction, buildings and construction remain responsible for roughly a third of global energy use and a comparable share of emissions — and the material already standing is the cheapest carbon saving available. A house that lasts two hundred years and is adapted four times has amortised its concrete over eight generations. A house replaced after forty has not, whatever its rating said on handover day.

Longevity is not nostalgia. It is a set of decidable design moves: load-bearing structure that outlives its cladding, generous floor-to-ceiling heights that can take a new use, materials that improve as they weather, junctions that can be taken apart, and rooms proportioned well enough that someone in 2130 wants to keep them. The buildings we still queue to visit — Vals, the reused industrial fabric of northern Italy, Kéré's earthen work in Gando — pass that test not because they are expensive, but because they were decided properly.

The part nobody wants to say

The pressure is not only commercial. It is regulatory.

Permission systems across Europe are formidable at policing what can be measured and almost silent on what can be seen. Setback distances, plot ratios, parking counts, thermal coefficients, fire compartments, form after form. Approval means compliance, and compliance is not the same as quality. A scheme with meanly proportioned windows, a two-metre-thirty ceiling, a facade composed by the cheapest panel module and an entrance nobody would photograph can be entirely legal — and often faster to approve than something considered, because a novel proposal invites scrutiny while a dull one passes as normal.

What is missing is a baseline: a published codex of minimum standards for the things that decide how a place feels a century from now. Not a style code, and not taste enforcement. A floor. Minimum daylight and ceiling height per habitable room rather than per dwelling average. Rules of proportion for openings on a public elevation. Declared design life and a replacement path for every external material. A named architect, accountable through to completion, on anything the public has to look at. Independent design review with the standing to refuse — and the obligation to say plainly why.

Building regulation already accepts this logic everywhere else: nobody argues that structural safety should be left to the market. Aesthetics, daylight and durability are treated as matters of preference precisely because they are harder to measure — so the one thing a permit never guarantees is that the result will be worth keeping. That is the broken part of the system, and it is fixable with a baseline rather than a slogan.

ARCHITECTT position

ARCHITECTT is a France-based architecture studio built around a straightforward proposition: pair the judgement of architects with AI where AI is genuinely good, and use the saved hours to widen access rather than to widen margins.

AI is good at the parts that used to price small projects out of proper design — rapid option generation, tireless variant testing, orientation and daylight studies, documentation scaffolding, consistency checks across a set. It is not good at deciding what a house should be, how a threshold should feel, or whether a proportion is right. Those remain human decisions, made by people who are accountable for them.

That split does two things at once. For homeowners and small developers, it makes considered design reachable at a budget that previously bought a template. For established practices, the same tooling becomes a B2B capacity layer: option studies, planning packages, and documentation support that expand what a team can carry without expanding headcount.

We are not claiming a finished revolution. We are stating a direction and publishing our work as it develops, including the concept studies, the reasoning, and the things that do not work.

What to watch instead of slogans

Four practical tests separate design that governs cost from design that decorates it:

1. Does the fee buy site presence? Drawings without follow-through become suggestions.
2. Is the wall build-up specified before the finishes are chosen? Sequence reveals priorities.
3. How many options were tested and discarded? One scheme is a preference; five tested schemes are a decision.
4. What is the declared design life, and who replaces what in year fifty? If nobody can answer, the building is a short-term investment wearing a long-term word.

The rule of architecture was never a style. It was the discipline of deciding before building, for people who are not born yet. That discipline is cheap to reinstate, and expensive to keep skipping.

Sources and method

Market sizes are quoted from three independent research houses, each with its own methodology and base year, and shown side by side rather than averaged: Grand View Research (2025 base, 2033 forecast), IMARC (2025 base, 2034 forecast) and Mordor Intelligence (2026 base, 2031 forecast). Forward figures are the publishers' forecasts, not ours.

Fee ranges are industry convention rather than statute; the RIBA Plan of Work and its associated fee guidance are the reference used here, and any individual appointment can sit outside those ranges. Energy and emissions shares for buildings and construction are taken from UNEP's Global Status Report. The regulatory argument in this article is our own position, stated as opinion, not a description of any specific authority's policy. Full links are listed under References below.

Further reading from our Journal: Vals and the architecture of stone and water, adaptive reuse in northern Italy, Kéré's earthen materiality in Gando and art and architecture in the south of France.