The variation arrives as a one-line email. The basement waterproofing the contractor priced at £17,000 needs a structural revision the surveyor flagged in week three. The new figure is £23,000. The client reads it twice, opens the original quote, and reads that twice. The numbers do not contradict each other. The numbers describe two different jobs. The first job was the one quoted at tender. The second job is the one the building is asking the trades to do. The signatures on the contract belong to the first job. The invoices arriving on the second-of-the-month belong to the second. By the end of the project there will be eleven of these emails. Each one is technically defensible. Each one is delivered on a Friday.
The cost plan was finalised before the structural engineer’s report came back two months later. The two documents describe a building with different load paths. Nobody is wrong. The cost plan was prepared on the drawings that existed at the time, and those drawings did not yet account for the steel that the engineer would later specify on the second floor.
The architect has produced a Stage 4 drawing pack. The pack is detailed enough for a contractor to begin work. The pack is not detailed enough for a contractor to price every element of the work. The contractor prices what is on the drawings. The contractor cannot price what is not.
A fixed-price tender returns at £88,000. The contract is signed. Eighteen months later the final account settles at £105,600. The movement was not waste. It was not negligence. It was the cumulative cost of every detail that became a question once the trades arrived on site and discovered the question was real.
Somewhere in the project file there is a contingency line. The QS set it at 10%. The actual contingency the project needed sat between 10 and 18%. The contingency was spent before the second-floor stripout began. After that the variations were paid out of the homeowner’s reserve, then out of the homeowner’s planned overrun budget, then out of conversations with the bank.
None of these failures is a failure of competence. The architect is RIBA-registered and has done eleven projects in the same postcode. The QS is MRICS and used the BCIS rates current to the quarter. The contractor has been trading for fourteen years and priced the job in good faith. The client read every page of every contract before signing.
Each professional executed against the brief in front of them. The brief in front of the architect at Stage 4 was to produce a coordinated design package. The brief in front of the QS was to produce a cost plan from that package. The brief in front of the contractor was to price the package as drawn. None of those briefs contained the words “and discover what is missing before pricing it.”
The 20% overrun is the name everyone gives to the gap between what the documents described and what the building actually contained. The gap was always going to be 20%. The contract was not designed to discover the gap before signing. The contract was designed to allocate the gap once it appeared.
This is the part where the homeowner stops trusting the industry. Not because anyone defrauded them. Because every professional in the chain handed off a defensible deliverable, and the building still cost a fifth more than the document with the homeowner’s signature said it would. The trust collapses on the gap.
The failure is sequence. The contract is signed before the design is complete enough to price.
A Pre-Construction Services Agreement, abbreviated PCSA, is a fixed-fee contract under which a contractor performs defined pre-construction services before any commitment to build. The Joint Contracts Tribunal publishes a standard form for it: the JCT Pre-Construction Services Agreement (General Contractor). It is the formal mechanism for what the industry calls a two-stage tender. Stage 1 is the PCSA itself. The contractor is paid to do the investigative work that turns an incomplete design into a buildable, priceable scope. Stage 2 is the main construction contract, signed only when Stage 1 has produced a cost plan the client and the contractor both stand behind. The PCSA is the price the project pays to find out what the project actually is.
What a PCSA actually does
A PCSA produces six things, in roughly this order: a detailed cost plan, a construction programme, a buildability review, a risk register, trade engagement, and pre-contract documentation. None of these is unique to PCSA. They are deliverables a competent contractor would produce eventually anyway. What PCSA changes is when they exist. The CIOB documented this discipline in a 2025 Technical Information Sheet on pre-construction services.
The cost plan is a line-by-line projection of project cost, built from the drawings, the engineer’s report, the surveyor’s findings, and the actual prices returned by the trades the contractor has spoken to. It is not a quote. It is a forecast with the workings shown. Where a number is uncertain, the cost plan names the uncertainty and carries a provisional sum for it, resolved once the design pins down.
The buildability review is the contractor reading the architect’s drawings and naming everything in them that will not build the way it is drawn. The beam that conflicts with the soil pipe. The window head that will not pass thermal regulations. The flooring transition the joiner will charge an extra week to make work. The review goes back to the architect. The architect revises. The cost plan is updated against the revised drawings.
The risk register lists every known unknown. The trade engagement schedule confirms which subcontractors have priced which packages and when the prices expire. The pre-contract documentation is the bound output of the other exercises: the document the main contract references when signed.
The work happens before the main contract. The contractor is paid a fixed fee to do it. The contractor is not racing to win a fixed-price quote with assumptions inside it.
What a PCSA costs
A PCSA fee is a fixed sum, paid to the contractor for the Stage 1 work, regardless of whether the project proceeds to Stage 2. It is structured as a percentage of estimated project value, typically two to three per cent on UK residential two-stage tenders. It is paid in stages tied to deliverables: cost plan submission, buildability review sign-off, pre-contract documentation handover. The structure of the PCSA fee varies by contractor and project type; the principles are constant.
It looks like an added cost. It is not. The same investigative work happens on every project that ever completes. On a single-stage tender it happens in the form of variations, change orders, contingency draws, and bank conversations. On a two-stage tender it happens in front of the contract, with the cost named on a line item rather than absorbed into surprise.
If the client does not proceed to Stage 2, the contractor keeps the PCSA fee and the client keeps the deliverables. The cost plan, the risk register, and the buildability review are the client’s. They can be issued to a different contractor for a Stage 2 tender. The PCSA does not commit the client to build with the contractor who priced it. It commits the client to find out what the building costs.
When PCSA is right and when it is not
PCSA is the right structure when the project carries design uncertainty, complex scope, heritage fabric, or value above £150,000. Listed Building Consent territory in RBKC, City of Westminster, or Camden is PCSA territory by default. Anything below ground floor is PCSA territory: basements, lower-ground excavations, party-wall complications. Anything that requires structural intervention to a Victorian or Georgian building, where the building’s behaviour is partly hypothesis until the trades are inside it, is PCSA territory.
PCSA is overkill on a £40,000 kitchen refurbishment with fixed scope, an in-stock product specification, and no structural work. The two-stage process costs more in time and fee than it returns in resolved uncertainty. A single-stage tender on a tight specification and a fixed-price contract is the right answer for that project.
The dividing line is design uncertainty, not project size, though the two correlate. A £200,000 kitchen-and-bathroom refit with bespoke joinery, structural steel, and a staircase inside a listed building carries more uncertainty than a £400,000 like-for-like refurbishment. The first wants a PCSA. The second probably does not. Naming the limit increases the credibility of the recommendation in the projects where the recommendation fits.
How Myrmex runs a PCSA
We run the PCSA on a six-zone framework. The zones are Scope, Technical, Governance, Financial, Programme, and Actions. The duration is four weeks. The fee is 2.5–3% of project value, fixed at the outset.
Scope locks what the project is and is not. Technical resolves the buildability questions and the design coordination gaps. Governance sets the decision authority, the change control process, and the contract structure under Construction Management. We operate as a Construction Manager, not as a Main Contractor, which means each trade is contracted directly to the client and we coordinate the chain. Financial produces the cost plan, the cashflow forecast, and the contingency framework. Programme produces the construction sequence with gated milestones. Actions is the open list of items that must close before the main contract is signed.
The framework exists in this form because we built Myrmex specifically to fix the gap that ate budgets at Construct & Furnish, project after project. Construct & Furnish ran two years and completed eighteen residential and restaurant renovations across London before it was formally dissolved. The systems failures inside it (quotes from memory, three messaging apps, contradicting spreadsheets, scope drift, variations issued in batches) are the failures the six zones are designed to make impossible.
The walk-through of the six zones in detail lives in a separate piece. This pillar names them. The cluster article works them.
Where the standards come from
The operational standard underneath our PCSA is consistency: the same product produced the same way by different people on different days, with no variation. The standard did not come from construction. It came from a cocktail bar.
We came to construction from London hospitality. Eight years each, both at Be at One. Artur was Denis’s instructor there before either of us was a contractor. The Be at One brand standard was simple and absolute: every cocktail at every venue tasted and looked exactly the same. A Pornstar Martini ordered at the Liverpool Street venue at 9pm on a Friday was the same drink as the Pornstar Martini ordered at the Soho venue at 11pm on a Tuesday. Same glass, same garnish, same temperature, same shake count. Different bartenders, different shifts, different locations, identical drink.
The standard, applied across thousands of drinks across years, produces a particular kind of operational discipline. Specifications are written down. Sequences are followed. Substitutions are escalated, not improvised. Quality is the floor, not the ceiling. The bartender who pours a slightly better Pornstar Martini is the bartender breaking the standard.
That discipline is what our PCSA is built on. The six zones are the project equivalent of a cocktail spec sheet. The four-week duration is the project equivalent of a 90-second build time. The fixed fee is the project equivalent of a fixed price per drink. The framework does not produce operational discipline. The framework records the discipline so that whoever runs the next project runs it the same way.
Myrmex is six weeks old. The discipline behind it is older than the firm. Our PCSA is the form the discipline takes when the product being delivered is a residential renovation between £150,000 and £1,000,000 in Central London.