TikTok's default window vs. your counting window

TikTok's default attribution is 7-day click + 1-day view. A user who saw your ad 7 days ago and messages today gets attributed to that ad.

Most advertisers count "today's" contacts and compare to "today's" TikTok conversions. Of course they don't match — different windows.

Fix: Align attribution windows. Either change TikTok's attribution to "1-day click" to match your daily counting, or switch your own accounting to rolling 7-day. Use the same window on both sides and the gap shrinks significantly.

Not all clicks are real users

TikTok ad traffic is not 100% clean. Some clicks come from:

• Bot traffic — especially severe in Southeast Asia campaigns
• Accidental taps — users scrolling and accidentally touching ads
• Competitor click fraud — intentionally burning your budget

These invalid clicks generate a ttclid in TikTok's system but produce no real WhatsApp conversation. TikTok records a "click," you get no matching contact.

Fix: Add validity checking in event reporting — only fire the conversion event when a user actually sends a message in WhatsApp, not just when they click. Huichuan's reporting logic triggers on "first message sent," automatically filtering invalid clicks.

Same user counted twice

If you're using both TikTok Pixel and Message Event Set, and both report the same conversion event, TikTok uses your event_id to deduplicate — but only if you're passing event_id correctly.

Many development teams forget to set event_id, or generate a new random ID on every call, causing TikTok to count the same conversion twice.

Another scenario: the same WhatsApp user starts multiple conversations (switched numbers, used a friend's phone), each counted as a new conversion.

Fix: Always pass a stable event_id (use conversation_id as the event_id). Huichuan handles deduplication automatically at the infrastructure level.

User converts on a different device

Common scenario: user sees the ad on mobile, notes down your WhatsApp number, then messages you on WhatsApp Web from their laptop that evening.

The original ttclid is tied to the mobile device. The WhatsApp Web message comes from a different IP and device — the two can't be linked, and attribution fails.

Fix: This is an inherent limitation of TikTok's attribution model. Mitigate it by extending ttclid retention (Huichuan defaults to 24 hours, configurable up to 72 hours) to capture more cross-device scenarios.

Sometimes "deduction" is deliberate — and forgotten

Experienced media buyers sometimes intentionally set reporting ratio below 100% — for example, reporting only 50% of conversions — to maintain higher CPA bids in TikTok's auction while keeping reported CPA lower.

This is an advanced strategy. But if the setting is forgotten or inconsistent across ad groups, your reconciliation reports become impossible to interpret.

Fix: Centralize reporting ratio management in Huichuan's dashboard, with full history. New accounts should start at 100% reporting and only adjust ratios after the algorithm stabilizes.