Operators repeat the same mistake: change copy when they need new proxies, or scrub the list and swap numbers when the offer is the problem. Wrong diagnosis costs time, money, and numbers. Here's how four funnel statuses pinpoint where to look.
Before diagnosing, split three independent zones:
Account and infrastructure - proxies, IP, number reputation, platform limits, Quality Rating.
List and segmentation - number validity, audience relevance, lead quality, send timing - see list segmentation.
Copy and offer - wording, length, question, CTA, offer fit for segment.
Each zone shows in its part of the funnel. You can't diagnose everything from one metric - see metrics dashboard.
Each message status points to its zone:
| Status | What it shows | Zone |
|---|---|---|
| Sent (one tick) → stuck, no delivered | Blocked account, rate limiting, recipient phone off | Account / infra |
| Delivered (two ticks) → low % | Dead or invalid list, bad proxies | List |
| Read (blue ticks) → low % | Weak preview, bad first screen | Copy |
| High read → no replies | Weak offer, no clear CTA, wrong segment | Copy / segmentation |
If bulk messages sit in sent (one gray tick) and don't move to delivered for hours - not a copy problem. Meta accepted the packet; something blocks delivery.
Check: proxy state, IP reputation, account limits, number status - see ban and limit mechanics. Rewriting copy changes nothing here.
Note: stuck on sent can also mean recipient phone off or no internet. Mass pattern = account. Isolated cases = normal.
95–98% delivery is the guide for quality API infrastructure. 70–85% read within 2 hours of send. Delivered Rate below 80% = red flag for list and infra - see benchmarks and red flags.
First hypotheses: unvalidated list (many numbers without WhatsApp), Bounce Rate above 2%, compromised proxies. Complaint Rate should stay ≤ 0.1%, Quality Rating > 0.80 - if below, start with account.
Rewriting copy when delivery drops is wasted time.
Most common and least obvious pattern. Delivery 95%+, blue ticks flowing - Response Rate near zero.
Main hypothesis: copy problem. Not the only one - wrong segment, bad send time, or weak offer unrelated to wording - see read no reply and first-message structure.
Check copy: clear question? simple yes/no reply? 150–200 characters? first screen overloaded? 15–25% disable read receipts - real Read Rate is always slightly higher than CRM shows.
B2B company broadcast to procurement list. 190 of 200 stuck on one gray tick. Marketer assumed text blocked for stop-words.
Used personal phone outside automation, manually sent same text to 5 contacts. Instant double ticks and reads.
Diagnosis: copy clean, broadcast account reputation the issue. After mobile proxies and SIM swap, broadcast ran normally.
Beauty salon WABA broadcast. Delivery 97%, blue ticks everywhere - zero replies in 2 hours. Marketer blamed "API glitch" and kept sending to full 3000 list.
An hour later Meta disabled the template. Aggressive long copy with 50% discount upfront annoyed people - they hit Spam not Reply - see WABA and templates.
Perfect delivery said: account clean. Zero replies with high reads screamed copy. Diagnosis was obvious - ignored.
Both cases - forum practice, no public methodology verification.
Send same text manually from personal number to 5–10 contacts - quick helper test. If it delivers and reads - delivery problem isn't copy.
Limit: contacts are loyal. Numbers likely in sender's address book, boosting session trust and hiding spam buttons. They don't model cold user behavior - see warm vs cold base.
Cleaner test - micro batch to 20–50 real cold sample numbers and watch Response Rate - or validate via A/B testing. Signal closer to reality.
Before changing anything, walk through:
Only then change something.
| Symptom | Change |
|---|---|
| Delivery < 80% | List, proxy, number status |
| Bounce Rate > 2% | List quality and validation |
| Complaint Rate > 0.1% | Content, segmentation, relevance |
| High Read Rate, no replies | Copy, CTA, offer, segment |
| Template Disabled | Template copy, offer structure |
| Sent not → Delivered | Infrastructure, account limits |
Take your last failed broadcast and run the algorithm: Delivered Rate first, then Read Rate, then Response Rate. Find where the funnel breaks - only then hypothesize the cause.
Practical rule:
Don't change copy until you know messages deliver. Don't change the number until you know people read and stay silent.
