QL-D is a product, not a printout
A utility records search that arrives as a zip of unrelated PDFs is not a PAS 128 desktop study. QL-D work means the returns are compared, conflicts are noted, and the supposed apparatus is plotted in the same coordinate system the rest of the project will use. Lines that exist only because a statutory owner said so are labelled QL-D. That honesty is the value. Pretending a record is a surveyed position is how strikes happen.
PAS 128:2022 treats Type D as a prerequisite to Types C, B and A. You can stop after QL-D if the decision is “do not buy this site”. You should not skip it because a detection crew is already in the van. The CICES quality-level note describes Type D as the foundation, including the option for a client to supply robust existing data.
A good study also records what did not come back. An unanswered owner is a residual risk, not a blank that can be filled with optimism. We write those gaps down so a designer can decide whether to wait, to assume a protection zone, or to move straight to detection.
What we ask you to send
A red-line boundary, a site address, and any drawings you already hold. If a previous consultant collected returns last year, send those too; we will say whether they are still usable. If the site is a live highway or a campus with a private network, name the people who can authorise access to their archives. That is often slower than the statutory process and more important.
UK settings where records are unusually fragmented
The Clyde waterfront in Glasgow has seen dock, road, civic and later commercial schemes on the same ground. Records do not live in one drawer. Cardiff Bay has a similar pattern: reclamation, later infrastructure, and public-realm work that cut through earlier corridors. A desktop study in those places is an editorial job as much as a request letter.
London adds volume. Multiple statutory owners, Transport for London assets, and private basements mean the Type D plot is busy even before a surveyor lifts a cover. We still start there, because a PAS 128 survey that ignores the paper trail wastes the detection day confirming things a record already showed — or worse, missing the one owner who never returned a plan.
What happens after QL-D
Site reconnaissance (Type C) tests the plot against the street. Type B detection then looks for the apparatus the records implied and the apparatus they never mentioned. Trial holes are reserved for the crossings that still control the design. That sequence is how you spend money in the right order.
If you are writing a brief this week, read when to specify a desktop study and keep HSG47 in the same pack. Records are planning. They are not location.
What “current enough” means in practice
A return from three years ago on a quiet rural lane may still be useful. A return from last spring on a Glasgow Clyde waterfront plot that has seen a later diversion is already a historical document. We will say which is which. The test is not the date stamp alone; it is whether the asset owner and the surveyor still consider the information robust, which is the language PAS 128 uses for client-supplied Type D data.
Private networks need introductions. District heating, landlord electricity, private water and campus ducts are often absent from statutory letters. A desktop study that ignores them is complete only in a bureaucratic sense. Name the estates manager in the brief. One email from you is worth a week of us guessing who holds the PDF.
Conflicts should be drawn, not averaged. If two owners show a crossing in different places, the QL-D plot should show both and say so. Averaging them into a tidy single line is how Type B crews start in the wrong bay. Cardiff Bay parcels are a frequent example: reclamation-era corridors and later infrastructure drawn by different hands.
Build the study into the critical path. Statutory clocks do not care that your planning committee is on Thursday. Starting the records search the week you wanted to detect is how programmes slip while everyone stares at an empty inbox.
When the plot exists, decide deliberately whether to stop. Some purchases die at QL-D, and that is a successful study. Some sites need Type B on a small working area only. A desktop study that automatically upsells a full detection of the entire red line is a sales process, not a specification.