The Filing System Nobody Can Search

Weaknesses:

1. Findability depends on personal memory. "Stuff" and "final FINAL v2" are not descriptions; when the two leavers go, that knowledge leaves with them.
2. No index to the paper contracts, so a three-year-old contract cannot be located reliably or promptly.
3. No single classification scheme - the same document type sits in three places with three logics (year, person, date).
4. No version control, so "final" is unknowable and an outdated copy may be relied on.
5. Retention is not applied, so records are kept too long or destroyed too early, both of which create risk.

Six-week priorities (in order):

1. Answer the two live requests first - locate the contract and the disputed invoice, and record where they were found. This tests the system and buys goodwill.
2. Build a simple index of the paper contracts: reference, parties, signature date, expiry date, location (drawer and folder), retention-until date. A spreadsheet is enough.
3. Agree one folder structure and naming rule (type – party – year – short description, with version numbers), and migrate the shared drive into it.
4. Interview the two leavers and write down where things are and what the conventions are.
5. Apply retention: list what is past its date, get written authorisation before destroying anything, and record what was destroyed and when.

Do not attempt a perfect system in six weeks; aim for one that a stranger can use and an auditor can follow.

Naming rule: [Document type] – [Party] – [Year] – [Short description] – v[n] Example: Contract – Brightwell Packaging – 2022 – supply agreement – v2.

Prevents: the "final FINAL v2" problem. The highest version number is the current one, and the description tells a stranger what the file is without opening it.

Location rule: every record lives in exactly one place - the shared drive is the master, paper originals are indexed to it, and a copy is never the record of reference.

Prevents: the three-places problem, where staff cannot tell which copy is authoritative.

Retention rule: each record carries a retention-until date calculated from the rule - contracts seven years after expiry, general correspondence three years - recorded in the index at the time of filing, not worked out later.

Prevents: both premature destruction (the disputed invoice) and indefinite hoarding that makes searching harder and creates data-protection exposure.

Destruction rule: no record is destroyed without written authorisation from a named manager, and a destruction log records reference, description, date and authoriser.

Prevents: quiet loss of records that may be needed for a dispute or an audit, and gives the auditor evidence that retention is actually managed.

Effect on the response:

• The invoice itself is probably findable, but the supporting evidence - the order, the delivery note, correspondence about a variation - sits in the date-ordered correspondence shelf and the unindexed contracts, so assembling the file will be slow and may be incomplete.
• If the delivery note was signed without a count, or the order was placed by phone, the business may be unable to prove what was agreed. That weakness is in the record, not in the argument.

What to do:

1. Assemble the file before replying: purchase order, quotation, delivery notes, invoices, and all correspondence in the period. Note gaps explicitly.
2. Reply within a stated time, confirming what the business holds and asking the supplier to state precisely what they dispute and on what basis. Do not argue from memory.
3. Do not concede or deny on the phone. Put the position in writing so both sides have the same record.
4. If the evidence is genuinely missing, say what is missing and propose a resolution (e.g. a joint review of both parties' records) rather than bluffing.
5. Feed the lesson back: the reason this is hard is that delivery notes were signed unverified and orders were not always confirmed in writing. Those are the controls to fix.