Article summary
Freeze age bands, unknown states and minimum-cell rules before reviewing Sri Lanka +94 WhatsApp profile results.
The common analytical bias in a Sri Lanka +94 WhatsApp Gender and Age report is changing age bands after seeing the output. Moving boundaries to make one segment look stronger removes reproducibility.
The analysis plan precedes Excel
| Item | Pre-register |
|---|---|
| Age bands | Boundaries and business reason |
| Unknown | Independent category |
| Minimum cell | Suppression or merge threshold |
| Provenance strata | Forms, partners, legacy CRM |
Remove +94 format exceptions from the profile denominator
Preserve phone_raw and country evidence and create candidates separately. Report exception counts, but do not turn them into missing age or gender.
Task fields are not self-reported data
Age, gender, avatar and mapped phone are limited dated observations. They do not establish legal age, identity or real-world gender.
Do not complete unknown from a name
Names in any script, avatars and language cannot fill age or gender blanks. Unknown is more credible than fabricated completeness.
Small-cell protection limits re-identification
When crossing region, source, age and gender leaves too few records, merge bands or stop output. Do not lower the threshold after seeing results.
Keep reference attributes outside TXT
Upload one phone per line. Order age, language and permission stay internal. Excel reconnects through batch_row and never overwrites self-reported fields.
Publish two denominators
Successfully checked phones form the task denominator; phones with usable profile observation form the analytical denominator. Show the difference.
Prohibit high-risk uses
Do not use the output for identity, minor determination, employment, credit, insurance or differential pricing. Profile data does not create marketing permission.
The Sri Lanka conclusion
Freeze the analysis rule before opening Excel. Age bands then serve the business question rather than the desired chart.
