You wrote the plan. Welcome, first real action, check-in, nudge. It works beautifully for the eight customers who signed up last week. Then a ninth signs up on Wednesday, a tenth on Friday, and you are suddenly running three clocks at once with nothing written down about which customer is where.
That is the part nobody writes about. The message plan is solved. The operating model for a continuous intake is not.
Why does a WhatsApp onboarding sequence break the moment you have a second customer?
A day-by-day plan assumes everybody starts on the same day. On a rolling intake they don't. On any given Tuesday one customer is on day 0 and another is on day 7, and the plan you wrote has no opinion about that. The plan is fine. The enrollment model underneath it is the thing you have not designed.
The message shape is settled ground and I am not going to rewrite it here. If you need the four-touch build (welcome, first action, check-in, nudge) with the exact per-step setup, that lives in how to set up a WhatsApp automation sequence. If you want the longer post-purchase spine that runs out to day 30, that is the drip sequence guide. Both are done. Neither tells you what to do when customers keep arriving.
Here is the failure in practice. You launch your sequence on Monday for last week's signups. Wednesday, three more customers sign up. You have two bad options: add them to the running sequence and hope, or start a second sequence and now maintain two. By week four you are maintaining seven, none of them are in the same state, and the whole thing has quietly turned back into the to-do list you built the automated WhatsApp messages sequence to escape.
The fix is not a better message plan. It is deciding, explicitly, how people enter and leave. Everything below is that decision.
Does an onboarding drip start a personal clock for each customer, or move everyone together?
Everyone moves together. A Blueticks drip runs over a shared audience, and the gap before the next step is anchored on the previous step's completion for the whole group, not on each recipient's signup date. There is no per-contact "their day 3". A mid-sequence joiner never receives step 0.
This is the single most misread thing about the product, so let me be exact about the mechanics.
A drip is a list of ordered steps over one shared audience. Each step is dispatched as a real campaign tagged with the drip and the step index, which is why steps inherit campaign pacing and deduplication rather than reinventing them. The gap attached to a step is a value plus a unit, and the unit set is minute, hour, day, week, month, year. Minute is the floor. There is no seconds option, and the gap value itself is just an unbounded non-negative number, so there is no maximum delay either. year is a unit, not a ceiling. The final step carries no gap, because nothing follows it.
When a step's campaign finishes its send wave, the orchestrator takes that completion timestamp, adds the gap, and creates the next step's campaign due at that moment. Read that again: the anchor is the previous step's completion, not anyone's signup. If step 1 takes six hours to finish sending because you have a large audience and an 8-second default pause between messages, the day-3 message lands three days after that finish, for everyone, at the same time.
The second consequence is the one nobody has written down. Every step re-resolves the audience when its campaign is created. So a contact you add to the audience board mid-sequence is picked up by the next step and never receives the earlier ones. They join at step 4 with no welcome. That is not a bug you can configure away. It is what shared-audience enrollment means.
Treat it as a design constraint, not a defect. Cohort anchoring is what makes one drip serve fifty people with zero per-person bookkeeping. It costs you per-person timing precision. Decide which one you actually need before you build.
Should you run weekly cohorts or one sequence per customer?
Two models work. Run a fresh drip per intake window (a weekly cohort) when volume is steady and a day of timing drift is harmless. Schedule messages individually per customer when volume is tiny or the exact date carries consequences. The deciding question is whether the specific day matters, not how many customers you have.
| Weekly cohort drip | One sequence per customer | |
|---|---|---|
| How you enroll | Add this week's new customers to one audience, launch one drip | Schedule each message by hand against that person's own dates |
| Timing anchor | Previous step's completion, shared by the cohort | Whatever date you type |
| Timing precision | An individual's "day 3" drifts by up to the length of your intake window | Exact to the person |
| Work per extra customer | Effectively zero once the audience exists | Every message, every time |
| Mid-sequence joiners | Join at the next step, never get step 0 | Not applicable |
| Best for | Setup nudges, feature walkthroughs, a whatsapp nurture sequence that tolerates drift | Trial-expiry warnings, renewal dates, anything with a legal or billing deadline |
My rule of thumb after running both: hand-scheduling stops being tolerable somewhere around five to ten new customers a week. That is a judgement from operating it, not a measured threshold. The real trigger is emotional. The first week you skip a follow-up because you couldn't face building it, switch to cohorts.
The hybrid is what most people land on and it is fine. Cohort drip for the qualitative touches, one hand-scheduled message per customer for the single date-critical one. Two systems, but each is doing the job it is actually good at.
Run next week's intake as a sequence instead of a to-do list. Drop this week's new customers into one audience, set the steps once, and let them fire on schedule from your own number, with no per-message fees and no API setup. Start with Blueticks free.
What does an onboarding operator's week actually look like?
Onboarding is a weekly job, not a day-0 task. The cadence that holds up: build and launch Monday, read the exit list midweek, prune Friday. Three touches from you per week, regardless of whether four customers arrived or forty. The whole point of the cohort model is that this cadence does not grow.
Monday, build and launch. Take everyone who signed up since last Monday, put them in one audience, and launch the drip with a first-step due time. That is the entire enrollment ritual. Ten minutes.
Wednesday, read the exits. This is the part that makes the week worth having. When a recipient replies and the drip has stop-on-reply enabled, they are recorded as an exit on that drip, with the reason, the step they had reached when they exited, a timestamp, and a best-effort snapshot of what they sent. Best-effort is doing real work in that sentence: the snapshot may carry the text, or just a media type, or a small thumbnail for an image reply, or nothing useful at all. Do not build a process that assumes you can always read what they said. Assume you get a name and a step, and open the chat for the rest.
The step index is the signal to read first. Three people exiting at step 1 means your welcome is landing. Three exiting at step 4 with confused questions means step 2 didn't explain what it needed to.
Friday, prune. Remove anyone who has gone live and no longer needs the rest, and anyone who has gone genuinely cold. Exits are keyed on the recipient's WhatsApp identity and the set only grows, so once someone is out they stay out of every later step, even if they get re-added to the audience board later.
"I stopped trying to know where every customer was. I only need to know who left this week and at which step. That is a five-minute read, and it is the only onboarding metric that ever changed what I did next." — composite operator account, written by me from repeated support conversations, not a verbatim customer quote.
The gotcha: a paused drip does not advance. If a step's campaign completes while the drip is paused, nothing creates the next step, and there is no background sweep that will notice later. Resuming re-drives the transition and catches up. Pausing and forgetting for two weeks is a real way to strand a cohort.
What happens when an onboarding step gets no reply?
Nothing happens, and that is the honest answer. There is no no-reply branch. A recipient leaves the sequence for exactly three recorded reasons, and "did not reply" is not one of them. Silence is the default path. The sequence just continues to the next step on schedule.
Let me refuse the things you are probably about to look for. There is no conditional branching. There is no wait-until-condition step. There are no per-step reply rules. The exit reasons are exactly replied, opted-out, and manual. Opt-out always removes someone. replied removes them only when stop-on-reply is switched on, and that switch is per drip, not per step. You cannot let a reply to step 1 stop the sequence while a reply to step 3 does not. If you want the mechanics of stop-on-reply itself, the setup guide covers it.
The design consequence is the useful part. Every step after the first has to make sense to somebody who has said nothing, because most of them will have said nothing. That rules out a whole genre of follow-up copy. No "as I mentioned". No "since you didn't get back to me". No message whose first line only parses if the reader answered the last one. Each step stands alone, adds one new thing, and does not scold.
Where you get real branching is your Friday pass. The exit list is the automatic branch. Manual removal is the deliberate one. A whatsapp follow up sequence built this way is a straight line with a human reading the exits once a week, and honestly that is enough for onboarding.
Which onboarding messages count as utility and which count as marketing?
Only if you send through Meta's Cloud API. Onboarding straddles the category line: a setup instruction is utility, a day-7 upgrade nudge is marketing, and one sequence can contain both. If you send from your own number through WhatsApp Web, none of this applies to you. No categories, no approval queue, no per-message fee.
Start with the trap, because it is the expensive one. On the Cloud API, every template must be categorized as authentication, marketing, or utility, you pick the category at creation, and WhatsApp validates your pick against the template's contents. Get it wrong and the template is rejected, or an approved one gets automatically re-categorized later by Meta's recurring review of approved templates. Your day-1 "here's how to finish setup" is a good utility candidate. Your day-7 "upgrade for X" is marketing no matter how you word it.
The money direction matters and people get it backwards. Meta's pricing docs state that marketing templates are always charged on delivery, while "utility templates delivered within an open customer service window are free", and that volume tiers unlock lower rates for utility and authentication only. Marketing gets no volume discount.
The window is customer-initiated. A user messaging your business opens a 24-hour customer service window, and each new inbound message from them extends it. You cannot open it yourself.
On delivery risk: do not plan around a fixed marketing cap per person. Meta's per-user marketing template limits are personalised and adapt over time based on that individual's recent read rate and inbox volume, and a blocked send returns error 131049. There is no published number to design against, and retrying immediately just re-hits it and corrupts your own delivery reporting.
For the full category and pricing breakdown, see the 2026 pricing category guide. For the opt-in side, our collection guide covers consent capture.
How do you tell whether onboarding actually worked?
Measure per cohort, not per contact. The onboarding-specific number is cohort completion: of the people who arrived in one intake window, what share reached the first real win. Pair it with exit-reason mix as a quality signal. Generic delivery and reply rates are covered elsewhere and are not the onboarding job.
Be clear about the reporting boundary. Each step runs as its own campaign, so per-step delivery lives with that campaign, and the drip sequence guide already defines per-step drop-off. Delivery, read, and reply definitions belong to the campaign analytics guide. I am not redefining either.
What the drip itself records is the exit list: reason, step index, timestamp, and the best-effort reply snapshot. That is your quality signal. A cohort where most exits are replied at step 1 or 2 is working. A cohort where the exits are opted-out is telling you the sequence is unwanted, and you should shorten it before you send it again.
Cohort completion is not a surfaced number. Nothing in the product knows what your "first win" is, whether that is a connected account, a first invoice, or a first message sent. You track it in a sheet: one row per cohort, the size, how many hit the win, and the exit mix. Four rows in and the comparison between weeks starts telling you things a per-contact view never will, because you changed one step between cohort two and cohort three and can see it.
What can't the WhatsApp app do for a continuous intake?
The Business app has greeting messages and quick replies. It has no sequencer, no concept of a cohort, and no record of who exited where. For a one-off batch you can fake it with reminders. For a continuous intake it collapses into a spreadsheet job, which is exactly the job you were trying to delete.
A greeting message fires once, to everyone, with no step 2. Quick replies are canned text you still have to send by hand. Neither carries state, so neither can answer "who is on day 3". That is the gap a whatsapp sequence tool fills: an ordered set of steps over an audience, a gap after each step, reply-driven exits recorded with the step they happened on, and a drip status you can read at a glance (draft, active, paused, completed, aborted).
The caveat before you commit, and it is a real one. Browser-only sending needs WhatsApp Web open and the machine awake when each step comes due. A closed laptop at the moment a step fires does not lose the message. It lands whenever the machine comes back, which for a day-1 setup nudge can mean 2am, or a day late. Cohort timing you designed is not cohort timing you get. There is no business-hours guard on the send path to protect you from a badly-timed due date either. The offline gateway is the fix, and per the Blueticks pricing page it sits on the Pro plan at $50 per user per month, or $45 billed annually. Free, Basic, and Standard are browser-only. If your cohort launches at 9am on a Monday when you are definitely at your desk, browser-only is genuinely fine. If your steps land while you are asleep, budget for the gateway rather than discovering it the hard way.
FAQ
Can I start a WhatsApp onboarding sequence automatically when someone buys? Not on the drip surface. Enrollment is by audience membership, and steps re-resolve that audience when each step launches. There is no per-customer purchase trigger and no CRM or webhook enrollment into a drip. The workable pattern is a scheduled intake window: add the period's new customers to the audience and launch a fresh drip for that cohort.
How long should a WhatsApp onboarding sequence be? Short. Four to five touches over the first week or two covers almost every case. For the actual day-by-day shapes, use the automation sequence setup guide for the compact version and the drip sequence guide for the longer post-purchase spine.
What's the difference between an onboarding sequence and a nurture sequence? Onboarding drives one specific action for someone who already bought. Nurture keeps someone warm who has not. Same machinery, different success condition. The full comparison is in the setup guide; I am not duplicating it here.
Do I need the WhatsApp Business API for this? No. Sending from your own number means no template categories, no approval queue, and no per-message fee. You take on the Cloud API's category and pricing rules only if you choose that path.
What if a customer replies halfway through? With stop-on-reply enabled on the drip, they exit immediately and are excluded from every later step. The exit records the step they had reached and a best-effort snapshot of what they sent. That switch is per drip, so it applies to every step or none.
Can I add someone to a sequence that already started? You can add them to the audience, and they will be picked up by the next step. They will not receive any step already sent, including the welcome. For anyone who arrives mid-cohort, either hand-send the opening message or hold them for next week's cohort.
