There Are Two Kinds of Planning in Construction. Most of Us Were Taught the Wrong One

Most project managers are taught to plan for the argument, not the finish.

It is rarely said out loud, but it is how the industry behaves. We build the programme, then we spend the job defending our position with it. Notices go in. Delays get logged. Records get kept, not so the job runs better, but so that when it goes wrong we can prove whose fault it was. An entire profession of programmers, surveyors and claims consultants exists to win that fight after the event, and they are very good at it. Adjudication referrals in the UK have just hit the highest level on record.

Here is the part that should stop us. Even when we are not planning to fight, the planning we do barely works. On the average site, only about 54 percent of the work planned for a given week actually gets done. That figure comes from the research behind the Last Planner System, and it has held up for years. Think about what it means. The weekly plan, the thing the whole job is supposed to run on, is roughly a coin toss. We are not planning to finish. We are planning to be half right, and then to argue about the other half.

Last week I wrote that the industry barely teaches the craft of planning a job to finish on time. This is the other half of that story. Because the problem is not only that we fail to teach the right kind of planning. It is that we have spent decades getting very good at the wrong kind.

The Plan Was a Fiction Before You Started

Defensive planning looks irrational until you see where the programme came from. On a lot of jobs, the date was never real in the first place.

This is not a complaint, it is government policy. The Treasury’s own Green Book, the rulebook for how public money is appraised, formally instructs people to inflate their programmes, because estimates are, in its words, systematically too optimistic. For a standard civil engineering job it tells appraisers to add up to 20 percent to the duration. For a complex one, more. Read that again. The government has written down, as official guidance, that the date you were first given is probably wrong, and wrong in the same direction every single time.

There is a name for this. Psychologists call it the planning fallacy, the well-documented habit of underestimating how long our own work will take, even when we know perfectly well that similar work has always run long. In one famous study, students predicted they would finish their thesis in about 34 days. They took 55. Only a third finished inside their own estimate. We are not bad people. We are optimists with deadlines.

And then there is the less innocent version. Bent Flyvbjerg, who has studied project overruns longer than almost anyone, separates honest optimism from what he calls strategic misrepresentation, the deadline quietly understated to win the work or get the scheme approved. His own description of the difference is blunt. Both are deception, he writes, but the second one is intentional. It is lying. Anyone who has watched a tender programme get trimmed by a fortnight in the last meeting before submission knows exactly what he means.

So picture the project manager on day one. They have been handed a programme they had no part in setting, built on a number the Treasury would tell them not to believe, sometimes shaved further to win the job. In that world, planning defensively is the logical move. If the plan was always going to be missed, the sensible thing is to document the miss carefully. We have built a system that sets the date up to fail, then trains everyone to keep good notes while it does.

The Strange Thing is, the Contract is on Your Side

What makes this such a waste is that the rules of the game were not written to be played this way. The instruments we fight with were designed to stop the fight.

The Society of Construction Law’s Delay and Disruption Protocol, the document the whole claims world treats as scripture, says in its own opening that its purpose is to avoid disputes. It describes the programme first as a tool for managing the works, and only second as a way of working out entitlement. The order matters. Records are meant to be the by-product of running the job well, not the reason for the plan.

The NEC contract goes further. Its early warning system, the thing that is supposed to catch problems before they bite, comes with an explicit instruction that early warnings should not be used to allocate blame. The emphasis, in NEC’s own words, should be on agreeing the steps to fix the matter rather than on who is at fault. The president of the NEC users’ group put it plainly. The confusion, he said, comes from people who grew up in an adversarial culture, and that culture should be discarded when using the contract. The contract is almost begging us not to plan for the argument.

Planning to Finish

The other kind of planning starts from a completely different belief about what a programme is for. A programme is not a document. It is an agreement between everyone who will build the job. The labourer, the supervisor, the site manager, the designer, the customer. If they have not genuinely signed up to it, you do not have a programme. You have a picture on a wall and a list of people to blame later. Winning that agreement is the project manager’s first real product, and everything else follows from it.

Let me show you what it looks like, with the job that taught it to me best. A few years ago I ran a 22 million pound, four storey concrete frame secondary school, built on a live school site, with seven sectional completions to hit along the way. Our own senior leadership would not sign the programme off at first because they thought it was too tight. We handed it over six weeks and one day early, and gave the ground floor back twelve weeks early.

We did not get lucky. We won that time before a spade was in the ground, and the way you do it is with a buffer. Not padding hidden in every task, a real, protected buffer, somewhere between 5 and 10 per cent, won deliberately and defended. The distinction matters, and Eliyahu Goldratt explained why decades ago. When you pad every individual task, the padding gets eaten. Work expands to fill the time it is given, and people start late because the slack looks like comfort. So you strip the padding out of the tasks, pool it into one buffer that protects the end date, and you guard that buffer like it is made of gold. Because it is.

Here is the detail almost nobody uses. The NEC contract already knows this. It requires the programme to show its float and its time risk allowance openly, and it treats the terminal float, the gap between when you plan to finish and the date in the contract, as belonging to the contractor and protects it. The buffer is not a trick you hide from the client. On a well run job it is a number you win in preconstruction, write into the programme, and defend in the open.

The second habit is that the plan is alive. It does not sit in a back office and get updated once a month into a tidy history book. It lives in the meeting, on the screen, re-sequenced with the supply chain in the room as the trades talk, owned by everyone from the trainee with a two week list to the MD. This is also how you move that 54 percent. When the people actually doing the work make the promises, and they only promise work that has genuinely been made ready, the plan starts coming true. Reliable sites run in the 80s, not the 50s. I ran a live programme in front of my supply chain, two and three times a week, for years. I came to call the way I do it the live programme method, for the simple reason that the programme lives in the room rather than in a drawer.

And when it does slip, because sometimes it will, the delivery planner does not reach for the contract. They get the trades around a table and rebuild the sequence together. I have taken an 11 million pound school that I joined eight weeks into delay, with seven months left, and handed it over on the contractual date. I have recovered a ten week delay from a structural frame failure back to two days. Neither was won with a better claim. They were won by rebuilding the plan with the people doing the work, fast, and in the open.

Two Jobs, Two Endings

You can see both kinds of planning written large in projects everyone knows.

When BAA built Heathrow Terminal 5, it did something unusual. Under the contract it used, the client held the risk, all of it, all of the time. The suppliers carried no risk premium, which meant they had nothing to protect by claiming and no reason to fight. Firms that were normally rivals were put in integrated teams and made to plan together. Terminal 5 was delivered on time and on its planned budget of around 4.3 billion pounds. The chaos everyone remembers from opening day was a baggage and staffing problem, an operational readiness failure after handover, not a failure of building the thing. The construction, planned without a reason to fight, finished.

Contrast that with a steelwork job at a paint plant in Northumberland that ended up in the Technology and Construction Court. The contract had started at under 2 million pounds and grown to more than 20. When it soured, the judge found that the client had devised a strategy to force its contractor into insolvency. He called the litigation bitterly fought. Two projects, same industry. One was set up so that nobody needed to fight, and it finished. The other was run as a fight, and it ended in court instead of in a finished plant.

You do not need a megaproject to see which kind of planner someone is. Just watch what they do on the Friday afternoon they discover a three week slip. The defensive planner closes the office door and opens the contract, working out whose fault it is and how to word the notice. The delivery planner gets the supply chain in a room first thing Monday and starts rebuilding the sequence. Same delay. Two instincts. Two endings. One of those jobs quietly finds its way back to programme. The other one lawyers up, and everyone on it spends the next year proving they were right instead of finishing the building.

What Comes Next

None of this is a personality type you are born with. Planning to finish is a discipline, and that is the hopeful part, because a discipline can be taught. The buffer won in preconstruction. The programme that lives in the room. The buy-in that turns a date into an agreement. The recovery that rebuilds instead of blames. None of it is magic. It is a craft, and like any craft it can be set down and learned.

That is exactly what I have spent the last two years doing, turning twenty years of delivery into something other people can actually learn. Next time, I am going to show you the programme I have built to teach it.

In the meantime, if any of this lands, I run a free workshop for contractors and tier two businesses who suspect that too many of their jobs are being planned to be defended rather than planned to be finished. No pitch, no slide deck. We look honestly at how your teams plan, and where the same fights keep starting. You can book directly at constructing-culture.co.uk.

A programme is not a document. It is an agreement between everyone who will build the job. Win that agreement, and most of the time the claim never comes.