A Jakarta beauty agency with 120 creators on its books does not usually lose deals to a better offer. It loses them to a direct message that sat unread for four days. The offer was fine. The channel was wrong. Those same creators open WhatsApp before they open anything else, and every agency in Indonesia already knows it, which is why so much creator outreach happens from a personal number with a backlog of 300 unread chats.
That habit is also why numbers get restricted. Send one message to eighty contacts before lunch, collect twenty replies asking who you are, take fifteen blocks, and the number that carried your whole roster goes quiet for a week while somebody drives to a shop for a replacement SIM. The roster does not care that the offer was good.
This is written for the agency or in-house team managing 50 to 200 beauty creators in Indonesia, where the roster is too large to message by hand and too valuable to lose to a restriction. It covers how broadcast lists differ from groups, what actually triggers limits, how to structure a list that survives a growing roster, and where a personal number should stop and a business number should take over.
Why WhatsApp wins in Indonesia, and what that costs you
The channel advantage is real and it is not subtle. A creator checking TikTok DMs is in work mode, filtering brand pitches between videos. A creator checking WhatsApp is in personal mode, answering family and friends. Message placement inside a personal inbox is why reply rates differ by an order of magnitude, and it is also why an unwanted message there feels like an intrusion rather than an ad.
That asymmetry cuts both ways. The same personal context that makes a reply likely makes a block likely. On TikTok, ignoring a brand costs one tap and no memory. On WhatsApp, ignoring means the message sits in a list with your mother and your landlord, and the fastest way to clear it is to block the sender. You are borrowing intimacy you did not earn, and the repayment is measured in blocks.
So the operating assumption should be uncomfortable: every batch you send is spending a reputational budget, not just a messaging quota. A 6% reply rate that comes with 4% blocks is a losing trade at 200 creators, because the number degrades before the roster converts.
Broadcast list versus group: the distinction that decides everything
Most teams start with a WhatsApp group because it is the button they already know. It is the wrong tool for outreach, and the reason is structural rather than stylistic.
| Broadcast list | Group | Business API send | |
|---|---|---|---|
| Can recipients see each other | No | Yes, all of them | No |
| Replies arrive as | One-to-one chat | Shared thread | One-to-one chat |
| Recipient must have your number saved | Yes | No | No |
| Practical ceiling | 256 contacts | 1,024 members | Rate-limited by provider |
| Block risk when used for cold outreach | Moderate | High | Low, with opt-in |
A group exposes your entire roster to each other. Two creators compare offers, a third sees that someone else got a higher rate, and by evening the group has become a negotiation forum you cannot control. Broadcast lists avoid that because each recipient receives the message as if it were sent only to them, and each reply lands in a private thread where you can actually negotiate.
The cost of that privacy is the saved-contact requirement. A broadcast list only reaches people who have your number in their address book. Cold outreach on a fresh number therefore does nothing at all, silently. This is the single most common reason a team concludes that “broadcast lists do not work” when in fact the messages were never delivered.
The three behaviours that get a personal number restricted
Nobody outside WhatsApp can tell you the exact thresholds, and anyone who quotes a number is guessing. What is observable is the pattern: restrictions cluster around three behaviours.
Sending to people who have not saved you. A batch where a large share of recipients have never interacted with the number reads as bulk messaging. The fix is not a better message, it is a different first contact: get the creator to save your number before the first send, or accept that this contact is not ready for WhatsApp yet.
Volume without response. Sending 60 messages and receiving zero replies is a stronger negative signal than sending 20 and receiving 8. Platforms measure engagement, and one-way blasting is the shape of spam regardless of intent.
Reports, not blocks. A block removes you quietly. A report tells the platform something was wrong. Fifteen blocks in a day is survivable; three reports is not, because reports escalate.
The practical implication is batch discipline: keep each send small enough that a bad batch cannot damage the number, and watch reply rate between batches rather than after the month closes. A team sending in batches of twenty can afford one bad message. A team sending in batches of eighty cannot.
Building a list that survives a 200-creator roster
The naive approach is one list called “creators” containing everyone. It fails at about 40 contacts, because the people in it are at wildly different stages and the only message that fits all of them is a generic one, which is the message that gets blocked.
Split by state, not by size. Four lists cover almost every team:
| List | Who is in it | What gets sent | Target reply rate |
|---|---|---|---|
| Saved, never contacted | Creators who added your number after a TikTok or email exchange | One short introduction naming the specific product and rate | 20%+ |
| Replied, negotiating | Anyone who answered in the last 30 days | Follow-up on the open question only, no new pitch | 40%+ |
| Sampled, no content | Product delivered, nothing posted | Deadline check with one concrete ask | 25%+ |
| Dormant 60 days | No reply and no content, no block | Nothing. Move to the next quarter | n/a |
The dormant list is the one teams skip, and it is the one that protects the number. A creator who has ignored two messages is not a prospect, and including them in a third batch spends block risk for no expected return. Pruning is not giving up on the roster; it is keeping the channel alive for the 60 people who will still answer.
Before any of this, the roster itself has to be worth messaging. A beauty agency in Jakarta picking from an eight-million creator pool needs to cut on follower tier and category long before the WhatsApp step, because sending a well-built broadcast about a skincare set to a home-decor creator wastes the one thing you cannot buy back. DAMI’s creator filtering sets follower tiers across six bands and drops the list into a named resource group, which is the point where a WhatsApp batch becomes a segmented send instead of a blast.
Where a personal number stops and a business number starts
A personal number is the right tool up to roughly a hundred active conversations. Past that, three things break at once: the inbox becomes unscannable, replies get lost, and the number accumulates exactly the kind of one-way volume that triggers restrictions.
The honest limit of the personal route is that it does not scale and it is not recoverable. When the number goes, the conversations go with it, because there is no export and no backup. Teams that have run this way for a year usually have a story about losing four months of negotiation history to one restriction.
The business route trades that risk for setup friction. A verified business account sends at volume without the same block exposure, but it requires verification, and messages arrive with a business label that some creators treat as less personal. For a first contact with a creator who has never heard of you, the personal number still outperforms. For the follow-up sequence across 200 people, it does not.
The split that works is by stage: personal number for the introduction and the negotiation, business account for the follow-up cadence and the sample reminders. DAMI runs both, with a personal-number bulk send for the warm portion of the list and a verified WhatsApp API send for the follow-up where volume is the whole point. The switch is not a philosophy, it is a threshold: when your unreplied follow-up queue passes a hundred names, the personal number has already stopped being the cheaper option.
The message shape that survives a personal inbox
A creator reading WhatsApp on a phone, between two other conversations, gives a brand message about two seconds before deciding. Everything follows from that.
Lead with who you are and what the rate is, in the first line. Not a greeting, not a compliment about their content, not a question. “I run creator partnerships for [brand], we pay [rate]% on [product], sample ships this week” survives the two seconds because it contains the only three facts a creator needs to decide whether to keep reading.
Second line: one specific detail that proves you looked at their account. Not “love your content” — that is the line every agency sends. A specific video, a specific product category they clearly favour, a specific format they use well.
Third line: one question with a yes-shaped answer. “Want the sample?” converts better than “let me know if interested” because it is answerable in one word, and one-word replies are what keep reply rate high enough to protect the number.
What to leave out: attachments on the first message, links, and anything that looks forwarded. A first message carrying an image and two links reads as a template even when it was written by hand.
Reading the results without fooling yourself
Reply rate alone is the wrong metric, because a batch of twenty with eighteen replies and twelve blocks looks excellent on reply rate and is a disaster. Track three numbers per batch: replies, blocks, and opt-outs where a creator explicitly asks to be removed.
A healthy batch at this roster size looks roughly like: 25 to 35% reply, under 2% block, and near-zero opt-out. A batch that returns 10% reply and 8% block is telling you the list segment was wrong, not that the message needs another rewrite. That distinction matters because the usual response to low replies is to send more, which is the exact behaviour that finishes the number off.
The measurement problem is attribution: WhatsApp replies land in a private thread, and nobody remembers which batch a conversation came from three weeks later. Tag as you go, not afterwards. DAMI’s chat view keeps roughly the last 30 days of conversation per creator with live translation, so an Indonesian reply can be read and answered without a translator in the loop, and the thread history stays attached to the creator record instead of living in a phone that might get restricted tomorrow.
FAQ
How many contacts can one broadcast list hold?
256, which is above the practical limit for outreach. The constraint that matters is not the list ceiling but the number’s tolerance: keep individual sends to twenty or thirty recipients and the list size stops being the risk factor.
Does the creator need to save my number first?
Yes, for broadcast lists. If they have not saved it, the message is not delivered and you get no error. Test with one contact you control before assuming a batch went out.
Can I use a personal number for cold outreach at all?
You can, and it is how most Indonesian agencies start. The cost is block rate and number lifespan. If the plan is 200 creators rather than 20, the personal number is a starting tool, not the system.
What happens to conversations if the number gets restricted?
They are gone from your side, and there is no export path. This is the strongest argument for keeping the negotiation record somewhere other than the phone.
Is the business API worth it below 100 creators?
Usually not. Verification takes time and the business label can reduce first-contact reply rate. It starts paying at the point where follow-up volume, not first contact, is what you are actually managing.
What to do before the next batch
Three decisions, in this order. First, cut the roster: anyone dormant for 60 days comes off the send list, permanently. Second, split by state rather than by size, so the message fits the people receiving it. Third, fix the first line of the message so it carries the rate and the product, because everything else in this article only works if the message earns the second second.
If the roster is the part that is fuzzy — you know you have around 150 beauty creators across two markets and cannot say which tier each is in — fix that before touching WhatsApp. The channel comparison and the multi-channel outreach structure both assume you already know who you are talking to. Getting the list right first is what makes a small send safe; sending carefully to the wrong people is still sending to the wrong people.
Start in the DAMI workspace by filtering the pool to your market and follower tier, tagging the creators already in negotiation, and only then running the first broadcast to the saved-contact segment. A hundred-person warm list is worth more than a two-hundred-person blast, and it is the only version of this that is still running in six months.
