Plumbing and heating trade supplier with three branches in Staffordshire and Cheshire, selling to installers and builders on trade account; a printed catalogue goes out each quarter. 46 staff.
Verdict: The profile covers 12,010 records in 3 sources. Required fields are 87% filled and valid (gaps accepted on purpose left out); 8.7% of records are extra copies by the exact rules; 55% have a UK GDPR basis recorded; 64% of email addresses may lawfully be marketed to; and 318 invoiced customers are missing from the CRM. The clean-up plan has 24 problems over 10,271 records, 469 of them handed to the Email List Health Check: about 320 hours with the allowance (360 available), purging past-retention records first, then merging, then the UK GDPR basis. Prevention: 13 required-on-create rules, a duplicate check on create in 1 of 3 sources (1 planned) and a 30-minute monthly routine. 6 of the 6 catalogue products named will be unlocked by the clean-up (0 ready today, not counted). Measure again the same way on 11 Mar 2027. Costs, side by side: £9,483 once (team time with the allowance, and the fee); £1,872 a year to keep it clean; and bad data now costs £3,393 a year, counted from the records (estimates left out).
| Source | Records | Extra copies (rate) | Exported | Duplicate check on create |
|---|---|---|---|---|
| CRM Trade customers, prospects and their contacts; the catalogue mailing list | 5,480 | 600 (10.9%) | 5 Oct 2026 | Planned by 4 Dec |
| Accounting package Trade accounts: invoicing name, address, email and account number | 2,140 | 63 (2.9%) | 5 Oct 2026 | No |
| Mailing tool Email newsletter and offers subscribers (one per email address; no company records) | 4,390 | 0 (can't hold them) | 6 Oct 2026 | Yes |
The mailing tool keeps one record per email and holds no company records, so it can't hold either kind of copy and is left out of the duplicate rate. The accounting package has no duplicate check when an account is opened, which is how a trade counter came to open 63 second accounts; a closed account leaves the count.
Counts only, from the exports. 18 fields profiled; 13 required; 2,066 record gaps accepted on purpose: CRM prospects without an email (contacted by post), trade-counter records with no named buyer, sole traders trading under their own name, and the records being purged. Every other gap in a required field has a row in the rulebook. The weakest field is the UK GDPR basis in the CRM. 386 extra copies by the same email; 277 by the same company and postcode; 302 similar-name pairs flagged, about 248 the same person on the sample (an estimate). 1,250 with no activity in 3 years: 1,070 to delete, 180 to archive while still needed. 318 in the accounts but not the CRM (customers matched: 81% of 1,690 invoiced in 24 months); 145 marked Customer in the CRM with no account.
| Source and channel | Records | Company or LLP: basis / none | Sole trader or unknown: met / not | Can be marketed to |
|---|---|---|---|---|
| CRM · Email | 4,443 | 1,980 / 540 | 610 / 1,313 (soft opt-in 520) | 2,590 (58.3%) |
| CRM · Post | 4,727 | 2,150 / 880 | 540 / 1,157 | 2,690 (56.9%) |
| CRM · Phone | 3,732 | 1,720 / 610 | 430 / 972 | 2,150 (57.6%) screened 2,150 |
| Mailing tool · Email | 4,333 | 0 / 0 | 2,985 / 1,348 | 2,985 (68.9%) |
Legitimate interests, not Contract, is the basis for marketing to customers. Emails to companies and LLPs need no consent, but sole traders, partnerships and anyone whose legal form is unknown need consent or the soft opt-in: 520 trade-account customers qualify for the soft opt-in (details collected at account opening with an opt-out, similar products, an opt-out in every message). Nobody is emailed to ask for consent: that email is itself direct marketing (ICO). Every number used for marketing calls is screened against TPS and CTPS at least every 28 days. Not legal advice.
| # | Problem and source | Records | Decision | Who checks | Hours |
|---|---|---|---|---|---|
| 1 | Opted out but still mailed CRM | 37 | Suppress | Sales office manager | 0.6 |
| 2 | Not touched in the retention period CRM | 1,070 | Delete | Sales office manager | 1.8 |
| 3 | Not touched in the retention period CRM | 180 | Archive | Sales office manager | 0.6 |
| 4 | Duplicate: same email CRM | 386 | Merge | Sales office manager | 12.9 |
| 5 | Duplicate: same company and postcode CRM | 214 | Merge | Sales office manager | 10.7 |
| 6 | Duplicate: same company and postcode Accounting package | 63 | Check by hand | Finance manager | 4.2 |
| 7 | Duplicate: similar name CRM | 302 | Check by hand | Sales office manager | 20.1 |
| 8 | No UK GDPR basis recorded CRM | 2,380 | Check by hand | Sales office manager | 79.3 |
| 9 | No UK GDPR basis recorded Mailing tool | 1,370 | Check by hand | Marketing assistant | 22.8 |
| 10 | Gone away CRM | 268 | Correct | Sales office manager | 35.7 |
| 11 | In accounts not in CRM Accounting package | 318 | Add to CRM | Sales office manager | 10.6 |
| 12 | Customer in CRM not in accounts CRM | 145 | Correct | Finance manager | 2.4 |
| 13 | Invalid format (email) CRM | 212 | Correct | Sales office manager | 5.3 |
| 14 | Invalid format (email) Accounting package | 41 | Correct | Finance manager | 1.0 |
| 15 | Invalid format (postcode) CRM | 143 | Correct | Sales office manager | 2.4 |
| 16 | Invalid format (postcode) Accounting package | 27 | Correct | Finance manager | 0.5 |
| 17 | Missing key field (legal form) CRM | 920 | Enrich | Sales office manager | 23.0 |
| 18 | Missing key field (customer or prospect) CRM | 520 | Correct | Sales office manager | 2.6 |
| 19 | Missing key field (email) Accounting package | 278 | Enrich | Finance manager | 9.3 |
| 20 | Missing key field (postcode) CRM | 610 | Enrich | Sales office manager | 15.2 |
| 21 | Missing key field (email) CRM | 296 | Enrich | Branch managers | 4.9 |
| 22 | Missing key field (postcode) Accounting package | 22 | Correct | Finance manager | 0.7 |
| 23 | Gone away Mailing tool | 412 | Hand to another product | Marketing assistant | 0.0 |
| 24 | Invalid format (email) Mailing tool | 57 | Hand to another product | Marketing assistant | 0.0 |
Opt-outs are suppressed first; then past-retention records are purged (deleted unless still needed: 180 lapsed customers with an open warranty claim or disputed balance are archived until it closes); then duplicates are merged; then the UK GDPR basis is recorded. Bounces and invalid mailing-tool addresses go to the Email List Health Check. Similar names are checked pair by pair: the sample confirmed only 82%, under the 95% needed to merge in bulk. Ledger accounts are never merged. Hours include realistic times for tracing gone-aways (8 minutes) and basis decisions (2 minutes), and the allowance covers exports and merge logs.
| Cost of bad data | Counted | Per item | How the cost was found | £ a year |
|---|---|---|---|---|
| Returned post Catalogues returned undelivered | 1,148 in 12 months | £1.32 | Invoice or receipt | £1,515 |
| Duplicate mailing Second copies of the catalogue to the same firm and address | 1,064 in 12 months | £1.32 | Invoice or receipt | £1,404 |
| Bounced email Hard bounces on the newsletter | 412 in 12 months | £0.00 | Supplier charge | £0 |
| Misdirected invoice Invoices sent to someone who had left, re-issued | 74 in 12 months | £6.40 | Timed | £474 |
| Other Sales staff looking up phone numbers that are wrong in the CRM | 520 in 12 months | £4.00 | Estimate | left out (£2,080) |
The one-off cost is 320 hours at £24 plus the £1,800 fee; the yearly cost is the monthly routine (6 hours a year) and branch time (72 hours a year). Every counted line of bad data rests on an invoice, a supplier's charge or a timed task; the guess at time lost to wrong phone numbers is shown but not added. The three are shown side by side, not set against each other. Returned post: 5.6% of items posted (1,148 of 20,600); a change smaller than 0.7 points by the re-measure could be chance.
| Item | Time | Who | What you look at |
|---|---|---|---|
| Records created this month with a required field missing: fix or ask the branch | 7 min | Sales office manager | The CRM's 'created this month' view, filtered for blanks in required fields |
| Same-email and same company-and-postcode checks on records created this month: merge | 7 min | Sales office manager | The two duplicate reports, run the same way as in the clean-up |
| Returned post this month: correct or archive | 5 min | Sales office manager | The post room's returns tally |
| Opt-outs and objections this month: on the suppression list in every system | 4 min | Marketing assistant | The unsubscribe log against the suppression lists |
| Phone numbers due for marketing calls: screened against TPS and CTPS (at least every 28 days) | 3 min | Sales office manager | The screening date on the call list |
| Records past the retention period: delete, or archive only if still needed; log the count | 4 min | Sales office manager | Records with no activity in the retention period, and archived records whose reason has ended |
15 entry rules (13 required on create); required fields with no rule: 0; duplicate check on create in 1 of 3 sources (1 planned); monthly routine 30 minutes, First Tuesday of each month, 9.00 to 9.30. Deleted records leave the deletion export and the pre-clean backup on fixed dates: deletion export destroyed 10 Jan 2027; pre-clean backup destroyed 12 Mar 2027.
| ROI product and what it needs | Measure | Level it needs (assumed) | Today | Status |
|---|---|---|---|---|
| Target Account List One record per company, with postcode and customer or prospect marked, so the list isn't padded with copies | Duplicate rate | at most 1.0% | 8.7% | Needs the clean-up |
| Customer Retention Review Every invoiced customer in the CRM, so customers lost can be counted from one list | Customers matched | at least 98.0% | 81.2% | Needs the clean-up |
| Lapsed Customer Win-Back Plan Lapsed customers with a UK GDPR basis recorded for contacting them | UK GDPR basis recorded | at least 95.0% | 55.0% | Needs the clean-up |
| Lapsed Customer Win-Back Plan Current addresses, so the letters arrive | Returned post rate | at most 2.0% | 5.6% | Needs the clean-up |
| Automated Email and Text Flows Plan Contacts who may lawfully be emailed (consent or soft opt-in for sole traders) | Emailable lawfully | at least 80.0% | 63.5% | Needs the clean-up |
| Direct Mail Campaign Plan A postal list that arrives | Returned post rate | at most 2.0% | 5.6% | Needs the clean-up |
| KPI Dashboard Design Customer numbers that agree between the CRM and the accounts | Customers matched | at least 98.0% | 81.2% | Needs the clean-up |
| Measure later: how it's taken | Better | Now | Owner's target (assumed) |
|---|---|---|---|
| Duplicate rate Extra copies by the two exact rules (same email; same company and postcode) ÷ records in the sources that can hold those copies | Lower | 8.7% | 1.0% |
| Key-field completion Required fields filled and valid ÷ (records × required fields, less gaps accepted on purpose), from Profile | Higher | 87.2% | 95.0% |
| Returned post rate Items returned undelivered ÷ items posted, over the months stated | Lower | 5.6% | 2.0% |
| UK GDPR basis recorded Records with a UK GDPR basis recorded ÷ records, from the Profile rows for that field | Higher | 55.0% | 95.0% |
| Customers matched 1 − (customers in the accounts with no CRM record ÷ customers invoiced in the last 24 months) | Higher | 81.2% | 98.0% |
| Emailable lawfully Email addresses that may be marketed to (companies with a basis; others with consent or the soft opt-in) ÷ email addresses held, from Permissions | Higher | 63.5% | 80.0% |
6 named; 0 ready today (not counted); 6 will be unlocked by the clean-up. The levels and targets are the owner's judgement, not facts. The same counts are taken 90 days after the clean-up finishes, from the same sources and exports with the same rules; records added since are included, so the re-measure shows whether prevention is holding. No £ is claimed beyond what the records show.
Record counts, completion, invalid formats and permissions are counted from exports of 3 sources, by the rules stated on Plan; similar-name duplicates are an estimate from a sample checked by hand; minutes per record and the 20% allowance are estimates, so the 320 hours are built up, not measured. The retention period (3 years), the gaps accepted on purpose, the level each product needs and the targets are the owner's decisions or assumptions. The counted £ (£3,393 a year) comes only from the owner's records (invoices, price lists, supplier charges, timed tasks); estimates (£2,080) are shown and left out; the yearly cost uses the owner's estimate of branch time. The re-measure, 90 days after the finish, uses the same sources, exports and rules; it includes records added since, so it shows whether prevention is holding, not proof of why. Returned post is a rate with a give or take of 0.7 points; the other measures count every record. Retention, lawful basis, PECR and TPS screening follow ICO guidance; not legal advice.
The deliverable is finished when the data-quality report, the rulebook and the routine are handed over; doing the clean-up is the team's work, in the order shown. Take a full export of every system before any merge or delete, and keep a log of what went. Not legal advice.
Fictional, continuing the example; the log is on the Agent sheet of the worked-example workbook, in counts only, and the months after add nothing to the figures above.
Stannerley took the guided route. At handover on 23 October 2026 the sales office manager and Geoff set up the agent together, in an AI assistant the business already uses, connected read-only to the CRM, the accounting package's customer list, the mailing tool and the unsubscribe log; Stannerley's IT made the connections beforehand. The company updated its privacy notice before the first run. Each Monday the agent checks the week's new and changed records against the entry rules on Prevention and the two exact duplicate rules, and puts what it finds on a fix list in one shared drafts document, with record numbers, for the person who looks after that system. It flags an entry rule when 5 or more new records break it in a week, and gives 7 days' warning of a TPS/CTPS screening, a destruction date or the re-measure (both the owner's settings on Inputs). Before each monthly routine it prepares the counts. It changes no record: the sales office manager, the finance manager and the marketing assistant approve and make every fix in their own systems, and the managing director approves anything deleted or destroyed. Nothing left the business, and no decision changed, without one of its own people approving it.
| Date | What the agent prepared or flagged, and the evidence | Kind | Approved or decided by | What happened |
|---|---|---|---|---|
| 26 Oct 2026 | Fix list for the CRM: 14 records created last week; 3 with no legal form, 2 copies of existing records by the same-email rule CRM records created 19 to 23 Oct; the entry rules on Prevention; the same-email rule on Plan | Draft | Sales office manager | Legal form added from the Companies House register for the 3; the 2 copies merged into the existing records on 27 Oct, both record numbers logged |
| 2 Nov 2026 | 2 customers on the unsubscribe log since 28 Oct still on the catalogue list in the CRM The unsubscribe log against the CRM's catalogue list and the mailing tool's suppression list | Flag | Sales office manager | Both suppressed in the CRM the same morning, kept on the suppression list, not deleted; already suppressed in the mailing tool |
| 3 Nov 2026 | Counts for the monthly routine, ready before 9.00: 9 October records missing a required field, 4 exact duplicates, 31 returned catalogues; numbers due for TPS/CTPS screening by 12 Nov CRM created-this-month view; the two duplicate reports; the post room's tally; the screening date on the call list | Draft | Sales office manager | Routine held in 25 minutes: the 9 records sent to the branches, the 4 copies merged, the returns worked through; the numbers screened on 10 Nov |
| 16 Nov 2026 | Entry rule not holding: 7 new records in a week marked Customer with no trade account number (the setting is 5) CRM records created 9 to 13 Nov; the entry rule for Customer or prospect on Prevention (enforced by a drop-down list, which can't check the number) | Flag | Sales office manager | All 7 re-marked Prospect until finance opens the accounts. The CRM's supplier asked to make the account number required when Customer is chosen; done on 2 Dec |
| 14 Dec 2026 | The deletion export from the purge on 19 Nov is due to be destroyed by 19 Dec (30 days, Inputs) The purge log (count and date); Inputs, days the deletion export is kept | Flag | Managing director | Destroyed on 18 Dec after checking nothing on it had been needed back; the date recorded in the purge log |
| 1 Mar 2027 | Re-measure due on 11 Mar (90 days after the 11 Dec finish); the pre-clean backup due to be destroyed on 12 Mar. Exports list prepared, the same sources and rules as in October Inputs (finish date, days to measure again, backup date); Sources | Draft | Sales office manager | Exports taken on 8 Mar the same way as in October and the re-measure columns filled in on 11 Mar; the managing director had the backup destroyed on 12 Mar |
At the 30-day check-in on 23 November, Geoff and the sales office manager read the first month's log together; his upkeep of the set-up ended there. The clean-up itself was the team's work, finished on 11 December as planned. What the agent added is that the new records stayed to the rules while it went on: copies were merged within the week they were made, two opt-outs were suppressed the morning they were found, and the one rule that wasn't holding was fixed at source. The monthly routine ran in its 30 minutes on prepared counts. Reading the weekly note took the sales office manager about 10 minutes a week (estimate). The re-measure on 11 March 2027 was taken by people, the same way, and its figures belong on the Value Case sheet; the months after add nothing to the costs or the counted £ above.