We Were Still Quoting From a Plant We Never Built

For a long stretch of last year, the number at the bottom of our quotations was descended from a plant that does not exist. We had costed it properly once, for a job in Manila that was designed, priced and then never built. The file was tidy, the arithmetic was sound, and because it was tidy and sound we kept opening it. Every new building got priced by taking that spreadsheet, nudging the volumes, and sending it out. Nobody had ever checked a single line of it against an invoice, because there were no invoices. The plant was never poured.

I have written about our costing before from a different end. The post about our design fee took apart the man-days behind a drawing set, and the one about reading a rival quotation line by line was about somebody else's document. The order we work in before a price appears was about sequence. This one is narrower and more uncomfortable: where the numbers inside our own quotation come from, and what happened when we found out that most of them came from nowhere.

An estimate that is reused stops being an estimate and becomes a habit

There is nothing wrong with estimating. You have to, at the beginning. Almost nobody has an influent result when they first call us, so the strength is assumed; the flow is often reconstructed from a water bill rather than measured. Early on, every number in a quotation is somebody's best reading of a situation.

What went wrong was not the estimating. It was that the estimate never got closed. A costing is supposed to be a question — is this what the work costs? — and the answer arrives later in the form of a supplier invoice, a crew's timesheet and a delivery note. If the job is built, you get the answer whether you want it or not. If the job is never built, the question sits there unanswered and looking exactly like an answer.

So the file aged in the wrong direction. Each time we used it, it felt more authoritative, because it had been used before. That is not evidence. That is repetition.

We re-costed the same plant at rates we had actually paid, and it came out materially cheaper

The thing that forced the issue was losing a job we should have won. We had quoted a commercial building off the inherited file. We were beaten on price, and not narrowly. Afterwards I did something I should have done a year earlier: I took the same plant, the same chamber sizes, the same equipment list, and re-priced every line using rates we had genuinely invoiced on finished work. Not catalogue prices. Not allowances. Rates attached to jobs that are in the ground and have been paid for.

Like for like, the same plant came out substantially cheaper. I am not going to publish the gap, because it is an internal number and it describes our own costing rather than a price anybody was offered. But it was not a rounding error, and it was comfortably larger than the amount by which we had lost the job. We did not lose on our cost base. We lost on a document.

That is the part worth carrying away from this. A company can be genuinely competitive and still quote as though it is not, because the quotation is built from a memory of what things cost rather than a record of what it has paid.

Every line in our model now says where its number came from

The fix was not clever. We rebuilt the cost model off a plant we had actually delivered, and we tagged each line with its source: a supplier invoice, our own crew rate, a unit we fabricate ourselves scaled from one we have built and sold. Where a line has no such source, it says so, and it is treated as the weakest number in the document rather than the most confident one.

Some of those rates are already public because we quote them openly. Our concrete benchmark is roughly fourteen thousand pesos per cubic metre of tank volume at 3,000 psi — about 1.2 million pesos for a 70,000-litre three-chamber tank. That number survives scrutiny for one reason only: we have paid it, repeatedly, through our own Cebu civil network, and we have watched what moves it. How the tanks are drawn moves it far more than any equipment choice; whether they go in the ground or on it moves it again; and we stopped letting the clarifier detail move it at all by settling on one detail and reusing it. Prices vary with filtration level, volume and site conditions, so the figure is ours and not yours — but it is at least a figure somebody has handed over money for.

The discipline has a side effect I did not expect. When a line is tagged with its origin, you can see immediately which parts of the plant you understand and which you have been guessing at for years. On our sheet the civil works, the fixed-film media and the installation labour were well evidenced. The instrumentation and the control panel were not. That told us where to go and ask.

Three lines were not in the old costing at all

Re-costing from invoices turned up something more serious than wrong rates. Three scopes of work had no line in the old model whatsoever, which means they were being delivered and absorbed, or not delivered at all.

The first is the regulatory package. Establishing the receiving water and its class, supporting the discharge permit, and setting up the self-monitoring report in the format the regulator expects are all real work by a real person over real weeks. We do that work at our own cost before quoting because an owner should not have to buy the answer to the receiving-water question twice, and the class decides the plant — but doing it for nothing and failing to cost it are different things. When a client is later asked for a submission, as happened on a job where a corrective action plan was requested before anybody had been cited, the hours are substantial and somebody is paying for them.

The second is odour control. We learned the hard way that a smell complaint inside an occupied building is usually an air-path problem rather than a biology problem, and that the vents and their terminations have to be drawn and located. Drawn and located means fabricated and installed, which means costed. Leaving it out of the model does not leave it out of the job.

The third is confined space and safety provision, and that one deserves its own paragraph.

Confined-space safety is the one line we do not trade

Every other line in a quotation is available for discussion. An owner can ask us to take out the concrete and price it three ways. He can ask us to halve the tank that treats nothing, and we will tell him what he is giving up and then let him decide. He can take the treatment level down to what he actually needs and we will agree with him.

Confined space is different, and the reason is not commercial. A sewage chamber with a sludge bed in it generates hydrogen sulphide, and the crust in a grease trap sits on top of exactly that. The gas is heavier than air, it collects at the bottom of the thing a person is about to climb into, and it deadens the sense of smell before it does anything else, so the warning is removed before the hazard is. Gas detection, forced ventilation, a harness point and a second person at the hatch are not a premium tier. They are the condition under which the plant can be entered at all.

It is also why our drawings treat access as a design item rather than an afterthought. A hatch of a usable size, a lifting eye above the sump, somewhere to stand with a hose — everything you take out of the plant has to go somewhere, and the cheapest moment to make that possible is before the cover slab goes on. A plant that can only be serviced by a man lowering himself into a chamber on a rope is not a cheaper plant. It is a plant whose price has been moved onto somebody else.

Then I went through our own quotations and told my team the prices were bloated

The correction cut both ways, and the second half is less flattering still. A few months after rebuilding the model, a quotation for an advanced treatment train came to me and I said plainly that our prices were bloated — that the process in question is expensive, but not as expensive as we were making it out to be.

The cause was the same error in the other direction. Rather than price the equipment that job actually needed, somebody had lifted line items from a different plant's costing and scaled them. A model assembled from another building's components is no more defensible than an estimate from a building that was never built. Where we have no supplier price for something, the answer is to go and find real prices and size the unit to the actual load, not to borrow a number that once sat next to a similar word.

Both failures have one root. A cost model is only as good as the last time somebody checked it against reality, and in both cases nobody had. It is also why we are relaxed about an owner putting our design through an AI and checking it, and why we read a client's own drawing properly before pricing it. A number that cannot be checked was never a number.

Cebu is not Manila, and an imported estimate prices it as though it were

There is a geographical version of the same mistake. The retired file had been built for a Manila job, and a Manila cost base is not ours. We have no freight coming in, our fabrication network is local, the crew is our own, and we are in our home environmental region, so a permit enquiry is a conversation rather than a trip. Our entry number here should be lower than a Manila entry number for the same plant, and the old file quietly insisted otherwise.

That matters to an owner more than it sounds. It means a quotation can be high for a reason that has nothing to do with the plant, the technology or the standard being met, and everything to do with whose cost structure the spreadsheet remembers. A supplier genuinely importing equipment has a different structure and may be right to quote differently. A supplier who is local and quoting as though he is not is simply carrying an old file.

What an owner can actually do with this

I am not going to tell you to interrogate your supplier, because that is not how these conversations go and a short quotation on a small job is not a failure of transparency. But there is one question that is fair at any size, and it is the question I had failed to ask myself: has this number been paid before, by anybody?

A rate that has been invoiced on a finished plant behaves differently from an allowance. It survives being asked where it came from. It comes with a job you can go and look at. And it tends to be accompanied by the three scopes above, because a company that has actually run the job knows the regulatory hours, the vent work and the safety provision exist. A quotation with no permit support, no odour provision and no safety line is not necessarily dishonest, but it is almost certainly a document written before anyone had delivered the work.

The other half of it is the running cost, which the purchase price does not show and which the design decides long before the commercial page does. Building the plant is the easy part; the drawing is where the money is settled, and which parts we can still support in five years is a costing question dressed as a specification question.

Where this stops being transferable

Everything above is the correction of one company's own bad habit, and a habit is not a method. A firm with a formal procurement department and a bill of quantities on every job cannot drift this way, because the format forces each line to be priced from something. Our exposure came precisely from being small enough to quote off a spreadsheet, which is also what lets us turn a building around quickly.

The rates themselves transfer to nobody. They are one Philippine company's concrete, labour and fabrication, measured in one region, and the parameters that govern your industry, your discharge class, your soil and your access will move them. A plant on a cramped site with no truck access does not get our access-and-haulage line; a plant with a long programme because the concrete is somebody else's does not get our civil rate at all.

And the specific conclusion — that re-costing made us cheaper — is not a law. It made us cheaper because the retired file was built in a more expensive city and carried allowances that were generous. Somebody else re-costing from invoices may discover the opposite: that they have been quoting below what the work costs them, which is a far more dangerous place to be and much harder to see, because the market rewards it for a while. The part I would hold to anywhere is the narrow one. Price from rates you have paid. If nobody has ever paid a number, do not let its age make it true.

If you are holding a quotation and you cannot tell which of its numbers have ever been tested, our STP price estimator and design generator will give you an independent figure to read it against, and the treatment services and contracting pages set out what we do and do not include. If the plant already exists and the question is what it costs to keep running, that is maintenance and operation and support on a plant somebody else built. Otherwise, book a free 15-minute consultation and bring the document.