The people behind the collections.

Keith works at the point where a collection stops being a storage problem and starts being a catalogue with commercial potential. Most of his week is spent on two kinds of conversation: an organization describing shelves of material it can no longer play, and a production describing a scene it needs and cannot find.

On the intake side he walks owners through what is realistic — which formats are urgent, which can wait, what documentation will be needed later, and how the intended use should shape the transfer decisions made at the outset. On the licensing side he handles research requests, advises on what is and is not available for a given use, and coordinates clearance and delivery with the technical team.

He came to archive work from distribution and sales, which shapes how he reads a library: not as a list of tapes, but as a set of stories someone is already looking for.

Marguerite's work begins before any machine is switched on. Sticky-shed binder, warped reels, shedding oxide and unlabelled cassettes each demand a different sequence, and choosing the wrong one can cost the recording permanently. She documents condition on arrival, prescribes baking, cleaning or mechanical rebuilding where it is warranted, and only then commits to a capture.

In the bay she maintains the playback machines and the alignment discipline that makes a transfer defensible: known settings, recorded decisions, and files that can be re-examined years later without guesswork. She treats the capture as the last chance to get it right, because for a great deal of material it is.

Her background is in post-production engineering, and she is direct with clients about what restoration can and cannot recover — a useful antidote to the assumption that any damage can be fixed later in software.

Desmond's premise is that a digital file left unmanaged is only marginally safer than the tape it came from. He plans ingest into the asset-management platform a client already works in, sets the naming and folder logic, and separates preservation masters from the derivatives that editors, researchers and licensees actually touch.

He specifies LTO archiving and the retrieval procedures around it, then tests them — because an archive nobody can get material back out of has not been built yet. Where a client's operation is spread across departments, he maps the handoffs so material does not stall between them.

He works closely with the licensing desk, since the difference between a library that can be sold from and one that cannot is usually an infrastructure decision made years earlier.

Ilse treats metadata as the part of the job that determines whether the rest of it pays off. She designs description schemes that fit the collection in hand — a late-night comedy library, a run of ceremony footage and an institutional research archive each need different fields and different levels of granularity.

Her work covers transcription and closed captioning, timecoded logging of people, places and subjects, and the consistency checks that keep a growing catalogue searchable rather than merely large. She also records what is known about provenance and rights alongside the content itself, so that a promising clip does not turn into a dead end at the moment someone wants to license it.

Trained in library and information practice, she is the person most likely to ask what a researcher will type into the search box five years from now.