What is an inspection backlog: inspection backlog explained

An inspection backlog is the set of inspections that are past their due date and not yet completed, measured by both count and age. The count on its own says very little. Nine items three days late and nine items nine months late describe two completely different estates, and only one of them needs a phone call today.

The term overlaps with backlog maintenance, which normally means outstanding repair work rather than outstanding examinations. Keeping the two in separate registers is worth the small effort, because they are cleared by different people with different skills and they carry different kinds of exposure.

What actually counts as backlog?

An item enters the backlog only if it had a due date. That sounds pedantic and it settles a lot of arguments. Work nobody scheduled, assets missing from the asset register, and inspections somebody intended to get round to are not backlog. They are an unknown, which is a worse problem, and they need counting somewhere else.

How is inspection backlog measured?

Three numbers, always reported together.

• Count. Items past due at the reporting date.
• Age profile. Bucketed by days overdue: 1 to 30, 31 to 60, 61 to 90, over 90. The tail is the interesting part, and it is the part a single average hides.
• Duty class. Statutory, contractual and discretionary, reported on separate lines and never summed into a headline.

Mean overdue days is a useful supporting figure once the buckets exist. Multiply each bucket's count by its midpoint, add the results, divide by the total count.

Inspection backlog explained: a worked example

A three inspector team faced 240 due inspections in a quarter and completed 168 after one member was off for six weeks. Seventy two items are overdue at quarter end, which is 70 percent inspection coverage.

The age profile: 40 items 1 to 30 days overdue, 22 items 31 to 60 days, and 10 items more than 60 days. Using bucket midpoints of 15, 45 and 75 days, the mean age is (40 x 15 + 22 x 45 + 10 x 75) divided by 72, which comes to about 32 days.

The duty split is where the report earns its keep. Of the 240 due, 60 carried a statutory duty and 51 were completed, leaving 9 statutory examinations overdue. The other 180 were internal condition checks: 117 completed, 63 overdue. The headline reads "72 overdue, 70 percent coverage". The number that should actually move somebody is 9.

Why one number hides the risk

A discretionary condition check that slips a month costs you information. A statutory examination that slips a month can mean an asset is in service without a valid examination, and that is a different category of problem: enforcement exposure, an insurance position that may not hold, and a defence with a hole in it if somebody gets hurt. Averaging the two together makes the serious one disappear into the harmless one.

Aggregation also destroys comparability between organisations. The National Audit Office found that government bodies include different things in their backlog calculations, with NHS England including moderate and low risk maintenance while the British Library counts only critical and urgent issues, and concluded that these inconsistencies prevent decision makers from comparing maintenance backlogs across departments at all (National Audit Office, Maintaining public service facilities, January 2025). The same effect operates inside a single organisation, between two sites that count differently and then get compared in a board pack.

The fix is dull and it works. Report backlog as a small table rather than a figure: rows for statutory, contractual and discretionary, columns for count and age band. Anyone reading it can tell in a few seconds whether the problem is urgent or merely untidy. That same table is what regulatory reporting will need later anyway, so it is not extra work, only earlier work.

Clearing a backlog, with the arithmetic done honestly

Backlogs do not clear themselves, because the interval keeps generating fresh demand while you work on the old.

Continue the example. Next quarter brings another 240 due inspections at half a day each: 120 productive inspector days. The 72 overdue items add 36 more. Total requirement, 156 days. Capacity at full strength, with three inspectors and realistic productive time, is about 108 days. The gap is 48 days, and no amount of encouragement closes it.

That leaves four honest options, and most recovery plans mix them: add capacity temporarily, lengthen intervals where a risk case genuinely supports it, reduce the scope of the lowest risk inspections, or accept the shortfall and have somebody senior sign that acceptance with the statutory items carved out of it. The fourth is unpopular and it still beats a plan that only works if nobody takes leave.

The backlog nobody sees

Two failure modes make a reported backlog smaller than the real one. Assets missing from the register generate no due date and therefore never appear overdue. And inspections marked complete on evidence that would not survive review are, in substance, still outstanding: they resurface later as a re-inspection or as an audit finding, months after somebody reported them green.

Both are worth a periodic sample check. Pull twenty completed inspections at random and ask whether a stranger could reconstruct from the record what was examined, when, by whom, and on which specific asset. Whatever fails that test was never really finished, and belongs back in the count.