The questions buyers actually ask us second

A description of the material and a rough volume. Formats and sizes, bound or loose, condition, whether it is Arabic, English or mixed, what has to be searchable afterwards, and where the output is going. For anything substantial we would rather walk the storage area than work from a spreadsheet, because condition is what moves a price.

No, and almost nobody does. Volume is estimated from shelf metres, box counts or filing cabinet drawers during the survey, then confirmed against real page counts once the first batches run. Estimates from linear measurement are usually within a workable range, and the contract can be structured so you pay against verified counts rather than the estimate.

Not usually, and it is often counterproductive. Preparation is a specialist step with its own method, and material re-sorted informally beforehand tends to lose the original filing order that the index will depend on. What genuinely helps is telling us which series are which, and flagging anything fragile, confidential or actively in use.

Yes, and it is usually the sensible route. A single department or record series gives you a real cost per page, a real throughput figure and delivered output your team can test in the target system. That result then makes the case for the rest of the programme far better than a proposal document does.

Fewer people than you expect, but the right ones. Someone who knows how the records are filed, someone who can approve the index fields, and someone from IT if the output is going into a live system. Where retention is in scope you also need whoever owns that decision, because it changes the capture scope.

There is a practical floor rather than a published one. Very small volumes carry the same setup cost as large ones: survey, preparation method, index design, quality standard and delivery configuration all have to be built regardless. Below a certain size that overhead dominates the per-page figure, and we will say so rather than quietly load it into the rate.

Normally against delivered and accepted volume at agreed milestones, rather than a single payment on completion. That keeps the incentive on steady delivery and gives you a checkpoint where output can be rejected before more of it is produced. Setup activity and any equipment mobilisation for on-site work are usually identified separately.

Condition discoveries, mostly. Heavier stapling than the survey suggested, damage needing repair, unexpected oversized sheets in a routine batch, duplicates, or index fields that turn out not to exist on a portion of the documents. Any of those is raised as a variation before the work is done, not applied silently to the final invoice.

Usually yes, per page. Equipment has to be mobilised into a room you provide, throughput is lower than a production floor, and the team is working within your access hours. It is the right choice where records genuinely cannot leave the building. Many projects split, keeping restricted material on-site and sending the rest off-site.

We will work to yours. Client-drafted confidentiality terms are normal in this industry and there is nothing unusual about a records project being governed by the client's paper. Ask separately whether individual operators are covered, because a company-level agreement alone leaves a gap that a per-person undertaking closes.

Yes, by arrangement, and it is a reasonable thing to ask for. A visit while work is in progress tells you more than a pre-contract tour of an empty floor: you can see custody handling, how batches are staged, and whether access controls are actually applied. Some clients write scheduled audit rights into the agreement.

By limiting who touches it and recording who did. Access is scoped to the engagement rather than granted across the production floor, personal devices are excluded from production areas, working copies are removed once delivery is accepted, and any residual project data is deleted on an agreed schedule. Your own obligations as data controller stay with you.

Yes, and it should be designed before capture rather than retrofitted. Security marking is treated as part of classification, so it becomes an index value at capture and maps to permission sets in the target repository. Retrofitting permissions across a delivered archive means reclassifying files one at a time, which nobody enjoys paying for twice.

No, and be wary of anyone who says it will. Recognition performs well on clean printed originals and degrades on faint photocopies, stamped or annotated pages, carbon copies and handwriting. A representative sample is tested before the approach is fixed, and documents that will not recognise reliably are identified up front so they can be indexed by key fields instead.

They are set per record class, not once for the whole project. Resolution, bit depth and colour mode are chosen against what has to remain legible and what recognition needs downstream, then proved on a sample batch before volume production. Coloured forms, pencil annotation and faint carbon copies all behave differently, and the sample is where that surfaces.

Yes, through batch separation. Barcoded separator sheets, patch codes or blank pages are inserted during preparation to tell the scanner where one document ends and the next begins. Barcoded separators can also carry the index values for the document that follows, which removes a keying step. Without separation a tray arrives as one enormous file.

Whatever is agreed before capture, driven by the classification scheme rather than invented at delivery. If output is going into a repository, the structure is built to that system's import requirements instead. We also deliver an index file mapping every image to its metadata, which is what makes a bulk load into a future system possible.

Yes, where the rule for what must be removed is defined by you. Redaction is applied so the underlying content is destroyed in the delivered copy rather than covered by a graphic that can be lifted off. It adds handling time per page, so it is scoped as its own stream and priced separately.

A named point of contact for the project rather than a shared mailbox. For anything running over several weeks you should also know who supervises the production team handling your material, and be able to reach them directly when something needs a decision. Ask for both names before contract, not after the first problem.

It goes back through capture at our cost, against the quality standard you signed off on the pilot. That standard is the reference point, which is exactly why it is agreed in writing before volume production rather than debated afterwards. Rejections are worth raising early: the same fault usually affects everything produced under the same settings.

For on-site work, yes, subject to your building access and security arrangements. Evening or weekend running is common where a records room is in daily use or where a department cannot lose access during working hours. It needs agreeing early because it affects staffing, cost and how your own security cover is arranged.

Delivery is the milestone, but the useful part is what comes after: verifying the load into your repository, correcting index issues found in real use, and handing over the conventions so your team can keep new records consistent. Ongoing capture, such as a digital mailroom or day-forward scanning, is scoped separately.