Article summary
Set reporting thresholds, unknown disclosure and cross-border phone rules for small Monaco +377 demographic samples.
In a small Monaco +377 list, one or two rows can materially change an age or gender percentage. “Easy to inspect manually” does not mean suitable for profile claims. Small samples require stricter reporting thresholds.
Set minimum reporting rules in advance
| Condition | Action |
|---|---|
| Sample below the internal threshold | Do not publish percentages |
| Unknown rate too high | Report quality only |
| Provenance too concentrated | Do not generalise to the market |
| Individual identification risk | Aggregate further or delete |
Mark cross-border phones separately
Phones serving Monaco customers may belong to other plans. +377 supports formatting and does not establish residence, citizenship or customer value.
Suppress small cross-tab cells
Combining age band, gender, source and place can make individuals easier to identify. Suppress small cells rather than disclose them for richer-looking content.
Put unknown in the main chart
Do not remove unobserved rows and task exceptions from the denominator. Usable count, missing count and causes matter more than one percentage.
Do not make individual decisions from observations
Prohibit use for eligibility, price, denial of service or priority. Personalisation should follow explicit customer choice and actual interaction.
Isolate input and output
TXT contains one phone per line and no identity or location labels. Excel lands in controlled staging; only an aggregate meeting the threshold may leave it.
Run a sensitivity analysis
Move one record from one age band to another and observe the percentage shift. If the conclusion reverses easily, label it unstable.
Apply automatic expiry
Timestamp every observation and delete detail after the small-sample analysis fulfils its purpose. Reassess next cycle rather than accumulate indefinitely.
The Monaco decision
A professional result is sometimes “insufficient evidence.” When sample and coverage do not support a conclusion, stopping publication is more valuable than forcing a profile.
