CC048 companion worksheets — fictional planning examples
Prepared: 8 October 2026

Files
- cc048-number-source-map.csv: five symbolic number aliases. No aliases are callable. Expand each real pool to one inventory row per actual telephone number in a restricted operational copy.
- cc048-call-outcomes-fictional.csv: twelve fictional call events. No personal caller information, actual recordings or transcripts.
- cc048-test-matrix.csv: ten proposed acceptance tests. Every test is not_run; no provider setup was tested.

Use
Import CSV as UTF-8. Keep ID and timestamp columns as text. The example offset is -04:00 for the fictional dates; it is not a Canada-wide time-zone setting. Adapt approved business definitions before operational use. Column names are vendor-neutral internal names, not provider API fields.

Number map
Document actual number, provider/account control, routes, pool members, recording state, effective dates, capacity review, completed tests and retirement date in an access-controlled operational inventory. Placeholder routes are not working phone destinations. Voice tracking does not establish SMS support or portability.

Outcome rules
One row represents one event. call_id is unique in this fictional file. Production identifiers must include the provider/account namespace where needed.
enquiry_id groups related events. parent_call_id links a repeat or callback to an earlier call. Repeated inbound calls remain separate events.
connection is the business mapping of call handling. Preserve the unmodified provider status separately in production.
qualification is the event's reviewed assessment. Use the latest reviewed enquiry assessment at or before the cut-off for enquiry-level reporting. This flat sample stores one assessment per event. Keep timestamped assessment history in production if earlier cut-offs must remain reproducible after later edits.
The two fictional voicemail scenarios establish new-business intent, while full qualification remains incomplete.
source and evidence_method describe the event. F004 has callback/unknown; the enquiry retains F003's paid_search_autumn acquisition source.
No telephone number alone should be assumed to uniquely identify a person.
revenue_cad is blank unless confirmed in this worksheet. Blank is not zero. F012's CAD 750 is fictional confirmed revenue, not a default lead value.
Do not sum duplicated enquiry revenue across calls. Production revenue should be held once per transaction or enquiry.
is_test is true or false. F009 is a fictional test scenario, not evidence that any actual test ran.
crm_match_status=matched describes the fictional scenario only. No real CRM was accessed.
This is not a Google Ads or other provider upload template.

Expected reconciliation at 2026-10-06T16:00:00-04:00
12 events = 1 test + 1 non-test outbound callback + 10 non-test inbound calls.
10 non-test inbound calls = 7 new-business events + 1 spam + 1 existing-customer call + 1 wrong number.
7 new-business inbound events group into 6 enquiries because Q001 appears twice.
Latest reviewed enquiry status: Q001 qualified/booked; Q002 qualified/quoted; Q003 not_qualified; Q004 qualified/booked; Q005 pending; Q006 qualified/won.
6 enquiries = 4 qualified + 1 not qualified + 1 pending.
2 missed human connections: F003 and F011. Q002 is recovered through F004. Q005 remains unresolved.
Resolved-qualification coverage: 5/6 = 83.3% rounded. The five have qualified or not_qualified decisions.
Qualified share among those with resolved qualification: 4/5 = 80%.
All six enquiries have a reviewed record; Q005 remains pending after review. Review coverage and resolved-qualification coverage are different metrics.
Confirmed qualified share of all six: 4/6 = 66.7% rounded.
Independent fictional spend input for paid_search_autumn: CAD 600 in media spending over 5-6 October 2026, excluding tax, agency fees and tracking costs. Q001 and Q002 qualify. CAD 600/2 = CAD 300 in media spending per qualified phone enquiry. The sample represents one fictional Toronto operation. The cut-off is 6 October 2026 at 16:00, UTC-04:00.
This tiny dataset teaches reconciliation and is not a performance benchmark.

Proposed decision rules
Do not recommend a budget change from this example.
Before using production data: verify source coverage, outcome coverage, deduplication, cohort maturity, currency and spend scope.
Maintain an exception queue for unclear enquiry matches, missing source evidence, missing owners, rejected integrations and inconsistent statuses.
Require actual test evidence before converting not_run to passed. Preserve failures and blocked cases.

Privacy
Review applicable laws and vendor processing before collecting operational data. Tracking metadata can be personal information. Recording, voicemail storage, transcription and automated analysis need a separate appropriate review.
Use only fictional data in public training materials and sample exports.
