The line you hear at conferences is that most delay happens in the
gaps between contractors, consultants and statutory bodies rather than
inside any one of them. It is a good line. The best UK evidence does not
say it. The correction now circulating in its place is also wrong, and
it is wrong in a more interesting way.
What the
evidence measures, and what it does not
The Get It Right Initiative’s research report, published in April
2016 and prepared by Expedition Engineering, is the one substantial UK
dataset on why construction goes wrong. Nineteen companies. A grounded
theory analysis of what practitioners reported. An expert panel ranking
causes over two anonymous rounds. Then an online survey of 143 people,
run specifically to test whether the panel spoke for the wider
industry.
Start with scope, because it governs everything below. GIRI
does not measure delay. It measures error and the cost of
error. The word delay appears three times in the whole report, twice
inside worked examples. Any sentence beginning “GIRI found that delay”
is already in trouble.
So if you want a UK figure for the share of programme delay that
originates at interfaces, there is not one. Not in GIRI, and not in
anything else I could open. The figure gets quoted anyway.
The ranking
everyone cites is the wrong table
GIRI ranked 17 root causes of error by financial impact twice, using
two different instruments, and got two different answers.
The expert panel put “poor interface management and design”
joint seventh of 17 across the industry. The online
survey put it 13th. The 13th is the number currently
doing the rounds, usually attributed to GIRI’s experts. It is not
theirs. It comes from the survey table on page 40, and the report is
explicit about what that table is: those respondents “were asked the
same question that the experts were asked during the Delphi Method
Analysis”. Same question, different people, different answer.
The panel’s own table sits on page 29. Almost nobody quoting GIRI on
interfaces has opened it.
The
top of the ranking is coordination failure under another name
Here is the expert panel’s ranking across the industry, from the top.
Inadequate planning from task through to project level, first. Late
design changes, second. Poorly communicated design information, third.
Then poor culture in relation to quality and poorly coordinated and
incorrect design information, joint fourth. Inadequate attention paid in
the design to construction, sixth. Then three at joint seventh:
excessive commercial pressures, poor interface management and design,
and ineffective communication between team members.
Ties mean the top seven ranks hold nine causes. Count the ones that
are failures between parties rather than inside one, and you get five:
poorly communicated design information, poorly coordinated and incorrect
design information, inadequate attention paid in the design to
construction (which is a designer not thinking about a builder),
ineffective communication between team members, and interface management
itself.
Arguably six. On planning, the first-ranked cause of all, the
report’s own respondents said “planning skills are not as good as
necessary particularly when planning involves co-ordination across
trades”.
That is the finding. Coordination between parties dominates the
causes of error, and it almost never arrives wearing the word
coordination. It arrives as a design problem, a planning problem or a
communication problem, each of which already has a department.
The definition is the
argument
The strongest evidence is not in the rankings. It is in the
terminology section near the front, where the report defines what it is
counting:
a defect is any failure to meet the project requirements at a
handover. A handover may be from one contractor to another contractor on
completion of a package, or a handover may be from a Tier 1 Contractor
to a Client on completion of a project.
The unit of analysis is the handover. Error becomes visible at the
moment work crosses between parties. That is not incidental framing, it
is the foundation the study is built on.
Then the report explains indirect cost, and does the rest of the work
for us:
if the steel frame was completed late due to correcting an error the
cladding contractor would have costs overcoming delays. Regardless of
who picks up the contractual liability there is still a cost to the
project.
The party that makes the error is not the party that pays. Keep that
last sentence. Liability is a separate question from cost, and the
project carries the cost whatever the answer.
Your
estate is buildings, and buildings are not civils
The industry-wide figure hides the split that matters to you.
In Buildings, both instruments agree: poor interface management and
design ranks sixth of 17. In Civil Engineering, the
panel put it eighth and the survey put it 14th. The across-industry 13th
is an average of two sectors that do not agree with each other.
The report also records that the survey link was circulated through
steering group members’ supply chains and by the Civil Engineering
Contractors Association to its members. It draws no conclusion from
that, and neither will I.
If your estate is buildings, wards, teaching space, housing, offices,
the number that applies to you is sixth, not 13th.
What this means for your
organisation
Your organisation almost certainly has no owner for the gaps and no
line in the risk register for them. Every party owns its own scope,
competently and in good faith. The interfaces between scopes are owned
by whoever happens to be standing nearest when they fail, which is
usually you. Nobody was negligent. It is simply what happens when the
dominant cause of error has no name, no budget line and no seat at the
meeting.
The practical move is not a new workstream. It is to stop accepting
design, programme and communication problems as three unrelated
categories when the panel’s own ranking suggests they are frequently one
problem, observed from three desks.
The honest limit
GIRI is 10 years old, and the industry has had a decade to move. The
study is 19 companies, and it is expert opinion on financial impact
rather than measured causation. Its headline estimate that error costs
the industry around 21% of total spend is a modelled figure that rests
on an assumption stated in the report itself: “If we assume that the
direct cost of recorded error is around 5%”. Treat it as an estimate,
and attribute it, because it is not a measurement.
And GIRI is about error, not about statutory consents, not about
planning authorities, and not about the approval bodies the conference
line likes to blame. On those, this article has nothing to offer,
because the evidence does not exist. It is better to say that than to
supply a number.
