The risk register: what infrastructure projects do differently, and why self-builders should borrow it

Most self-builders carry their risks in their head rather than on paper. "The ground might be bad." "The contractor might go under." "The neighbour might object to the party wall notice." These are real worries, and usually the right ones — but because they're never written down, scored or revisited, they compete for attention on the basis of whichever one feels loudest that week, not on how likely it actually is or how much damage it would do.

On a programme costing hundreds of millions, managing risk that way would be considered negligent. So a specific discipline has grown up around it: name each risk precisely, assign someone to own it, score how likely it is and how bad it would be, and record what's actually being done about it. None of that depends on scale. It works just as well applied to a single house, and it's considerably less effort than it sounds.

What a risk register actually is, and isn't

A risk register is not a list of worries. The difference is specificity and ownership. "Groundworks might be expensive" isn't a risk register entry — it's an anxiety. "If trial pits find made ground rather than natural strata, foundation design and cost will need revisiting before tender, and the groundworker's price will need renegotiating" is. The second version names the trigger, the consequence, and implies who needs to act and when. That's the entire difference, and it's what makes a register useful rather than just a longer way of worrying.

The categories people forget

Some risks get plenty of attention because they're obvious — weather, a contractor running late. Others get missed consistently, usually because they're awkward to think about or easy to assume away:

• Ground conditions that only become clear once trial pits are dug or foundations are actually excavated, not when a desk-based assessment was done
• A party wall negotiation that stalls, or a neighbour who simply doesn't respond, holding up a start date that assumed consent would be straightforward
• Services capacity — whether the grid connection, water main or sewer can actually take the new load, rather than assuming because the plot has a postcode it has capacity
• A contractor becoming insolvent mid-build, leaving work half-finished and a deposit or stage payment exposed
• Planning conditions that turn out not to be dischargeable on the timetable the programme assumed, because a survey or technical submission takes longer than expected
• A design that's still genuinely incomplete at tender, leaving contractor pricing qualified in ways that only surface once work starts
• A key individual — the architect, the contractor's site manager — leaving partway through, taking undocumented decisions and context with them
• A survey finding that gets quietly set aside because acting on it would mean admitting a delay, rather than because it's actually been resolved

Scoring honestly: likelihood and impact, not gut feel

On a formal register, each risk gets scored on likelihood and impact — often a simple 1-to-5 scale on each, multiplied together to rank them. The precision of the number matters much less than the exercise of making the comparison explicit. A low-likelihood, high-impact risk like contractor insolvency can still sit near the top of the list, because the consequence is severe even though it's unlikely — and that's exactly the kind of risk that gets under-weighted when everything is being managed by instinct instead of on paper.

Naming a risk early changes what you do about it

This is the actual payoff. Once a risk has a name, an owner and a score, it forces a decision: mitigate it, reduce its likelihood or impact, transfer it to someone better placed to bear it, or accept it and hold contingency against it. Ground conditions risk gets mitigated by commissioning a proper intrusive investigation before tender rather than discovering the problem after the contract is signed. Contractor insolvency risk gets reduced by checking trade references and financial standing before appointment, and partly transferred by staging payments against completed work rather than paying ahead of it. A risk that's just been worried about has no owner and no date attached to it — a risk on a register has both, which is the whole difference between hoping it doesn't happen and actually doing something about it.

What actually belongs on a self-build risk register

• The risk itself, written specifically enough to act on — not a vague category like "ground" or "money"
• An owner: the person actually responsible for watching it and acting if it starts to materialise
• Likelihood and impact, scored honestly rather than by whichever worry has been loudest that week
• The mitigation: a specific action with a date attached, not a hope that it won't happen
• The residual risk left over once the mitigation is in place — rarely zero, and worth stating rather than ignoring
• A review date, because risks change as a project moves — a register set up at the start and never revisited is barely better than not having one