Operational guide
Response rates on WhatsApp are much better than email and much worse than the marketing suggests, depending almost entirely on one decision you make at registration. Here is the whole operational picture from seven deployments, including the parts that went wrong.
Three moments from a real deployment, reconstructed. The template that opens the conversation, the discovery that follows, and the helpdesk traffic that ends up dominating the volume.
Sena
online
Hi Marie, you are registered for FOST New York next week.
I am Sena, the event concierge. Tell me what you are hoping to get out of the two days and I will put together your sessions, the sponsors worth your time, and who to meet.
A sentence or two is plenty, and a voice note is fine if you are walking.
09:12
Sena
online
Mostly looking for a payments provider we can actually integrate quickly
09:31
Useful, thank you. Two quick things so I get this right:
Are you evaluating now or later in the year, and is the blocker integration effort or pricing?
09:31
Evaluating now, we want something live before Q4. Integration effort is the blocker, pricing is fine
09:34
Got it. Three exhibitors match that, and two have engineers on the stand rather than sales.
Booth 214: ships a drop-in SDK, 2 week typical integration
09:34
Sena
online
which stage is the AI infra panel on
14:02
Stage 3 (Level 2, past the coffee bar on the right).
Starts 14:20, so you have 18 minutes.
14:02
anything after it worth staying for
14:03
Same stage at 15:10 there is a session on cost control at scale, which lines up with the integration problem you mentioned earlier.
14:03
Conversations are illustrative reconstructions in the shape our deployments produce. They are not transcripts, and no real attendee is depicted.
Yes, though not the way most organizers first try it. Messaging a registration list from a personal account or the free WhatsApp Business app gets the number banned, usually the week of the event. Doing it properly means a verified sending identity, consent captured before you send, and every pre-event message cleared through Meta review. That is a compliance project with queues you do not control, not a software integration.
The channel works. The reason most organizers hear about it and never ship it is that nobody mentions the compliance part until they are three weeks in. This page is what that project actually looks like, so you can decide whether you want to run it yourself.
We are not neutral about this: we sell a product that arrives with the whole setup already done. What we can offer that a neutral guide cannot is the failure list. Everything in the "what breaks" section below happened to us, on a real event, and is now a problem you would be paying us to have already had.
Every number below is a count over its base, from a named deployment. Percentages without denominators are the reason this category has a credibility problem.
Reply rate by how hard the organizer promoted it
113 replies of 123 delivered (CPME31). 65% across the full funnel, including everyone the send never reached.
67 replies of 177 delivered (FOST Singapore) - though 119 of 177, or 67%, were read.
Typical response rate, the baseline most organizers already know.
A deployment where the organizer never mentioned the channel to attendees.
Sources: CPME31 and FOST Singapore deployment reports. Survey baseline is the industry figure organizers already compare against.
Swipe the table to see every column
| Channel | Measured | Where it comes from |
|---|---|---|
| WhatsApp, promoted | 92% replied to the template (113 of 123 delivered); 65% funnel-wide | CPME31 Toulouse, opt-in on the registration form plus on-site promotion |
| WhatsApp, lightly promoted | 38% replied (67 of 177 delivered), though 67% were read (119 of 177) | FOST Singapore, content-led event, no heavy organizer push |
| WhatsApp, not promoted | Around 6% engagement | A deployment where the organizer did not mention Sena at all |
| Written survey | 10 to 20% typical response | Industry baseline, the comparison most organizers already know |
| Event app | 20 to 30% of registered attendees typically use it; ~40% for well-run dedicated networking apps | Buyer-stated benchmarks from organizer conversations |
The spread between rows one and three is the entire argument of this page. The channel is not the variable; the opt-in placement and the promotion are.
Note the second row carefully, because it is the most instructive. Two-thirds of attendees opened the message and only a third replied. Read receipts prove delivery and attention; they do not prove the message earned an answer. At a content-led event the pre-event WhatsApp ask is competing with the programme itself, and a template that does not make its value obvious in the first line loses that competition.
Five things have to be true before a single message reaches an attendee. Four of them are ours. The one in the middle is yours, and it is the one that decides whether the other four are worth anything.
01
Event messaging cannot run from a personal WhatsApp account or the free Business app. The number has to sit on the WhatsApp Business Platform under a verified business identity, and that means a provider relationship plus a Meta review queue with no published turnaround. It is the single most common reason a first deployment slips. Sena arrives on a number that already cleared it.
02
Anything you send to an attendee who has not messaged you first has to be pre-approved by Meta. Anything that reads like generic marketing comes back rejected, and every rewrite goes to the back of the queue. Two weeks is not enough runway on a new number. We bring templates that have already been through this.
03
This one is yours, and it decides whether the other four matter. One checkbox on the form your attendees are already completing, naming who will message them and what they get. We will give you the wording and the placement; we cannot put it on your form for you, and collecting consent later reaches a fraction of the people.
04
Phone numbers are the dirtiest field in any registration export, and the column recording consent is almost never named the same thing twice. Both fail quietly rather than loudly, which means you find out at broadcast time. We reconcile the export against what actually delivers before anything sends.
05
The moment the first message lands you have an inbox open, and attendees use it for everything: which room, what is next, and telling you they cannot come. Reply windows close on their own schedule, and an unanswered reply is worse than never messaging. Sena answers in seconds, from your programme and floor plan.
On the registration form. If you take one thing from this page, take that.
We have now watched this variable move engagement by more than an order of magnitude across deployments, and it dominates everything else: the template copy, the send timing, the agent quality. At CPME31 the checkbox was on the form and the organizer talked about it on site: 159 of 209 Speed Business participants opted in, 76%. At FOST Singapore the opt-in existed but the event was content-led and the push was light: 274 of 1,596, or 17%. Same product, same channel, same continent-agnostic mechanics. The difference was where the ask lived and how loudly the organizer made it.
What makes an opt-in convert
One warning worth the five minutes it costs: the consent column is almost never named the same thing twice. Across FOST events alone we have had it arrive under three different headings, once labelled for profile visibility rather than messaging. Confirm which column actually means "may we message this person" before you send anything. Read the wrong one and you either reach nobody, or you reach people who never agreed.
WhatsApp is a processor question, not a WhatsApp question. Using the channel does not create a compliance problem by itself; messaging people who did not agree, or being unable to say who processes their data, does. This is not legal advice, and an EU-facing event should have someone qualified read its setup. Five things have to hold, and on a Sena deployment they are already holding:
Every item below cost us real time on a real event. None of them are in a provider's documentation, and each one is a week you would spend finding it out.
This is the worst one, because it does not look like what it is. Inbound messages keep arriving; every outbound send fails. The error surfaced to us was a generic authentication failure, which sent us straight into rotating API keys and re-linking the number, none of which helped, because the real cause was Meta freezing the number after a display-name review was rejected. Nothing resumes until the freeze lifts, with the same credentials and no code change. Everything about the symptom points away from the cause, and we lost the time twice before we recognised it.
A template with more than one placeholder needs every parameter supplied for every recipient, in the right order. Miss one and Meta rejects the send with an error that names a parameter count rather than the attendee. The failure is uniform: it does not fail for one person, it fails for everyone. That makes it obvious in testing and catastrophic if you skip it, and testing it properly means real handsets on more than one country code.
A free-form reply is only allowed within 24 hours of the attendee's last message. During the event this is invisible, because attendees message constantly. The morning after, the same code path that worked all day starts failing, because every window has quietly expired. Anything you want to send afterwards has to be a template, which means it needed writing and approving before the event ended, back when nobody was thinking about it.
Missing country codes, landlines, numbers with no WhatsApp account, and typos all fail at delivery rather than at import. At CPME31 the gap between the 159 people who opted in and the 123 who received a template was almost entirely this: numbers that were malformed, unmatched to a record, or missed by a manual send batch. The gap only shows up if someone reconciles opt-ins against what actually delivered, and it is never zero.
Templates go to Meta for review. Rejections are common for anything that reads like generic marketing, and each rewrite goes back into the queue. Two weeks before the event is not enough runway on a number that has never sent before, which is exactly when most organizers start.
If you run several events off one WhatsApp number, moving it to a new event can disable the routing for the old one. Attendees from the previous event keep messaging and nothing answers. Nothing errors, nothing alerts, and the first you hear of it is an attendee from the last event asking why nobody replied.
Once you open WhatsApp, attendees will use it for everything, and mostly not for what you built it for. At FOST Singapore, 78% of conversations were navigation questions (which stage, which room, what is on next) rather than networking ones. At FOST Paris the channel handled 2,634 attendee questions, each answered in under three seconds.
Two implications. First, volume is higher than an email inbox and the expected response time is much shorter, so a human-staffed WhatsApp desk at a large event needs more people than organizers plan for. Second, whatever answers has to be working from your actual programme, venue map and exhibitor list. An assistant guessing at a room number does more damage to your event than silence would.
The unexpected uses are worth planning for too. Attendees message to say they cannot make it, to ask for an organizer, to report problems with the venue. Once the channel exists, it becomes the event's front door.
Yes, but not from a personal account or the free WhatsApp Business app, which gets the number banned. It requires a verified business sending identity, consent captured before you send, and Meta review on every message you initiate. That is a compliance project with review queues you do not control, and it is the part organizers underestimate. Sena runs on a number that already has all of it in place.
Any message you initiate outside 24 hours of the attendee's last message must be a Meta-approved template. Once they reply, a 24-hour service window opens in which you can send free-form messages. The window resets with each attendee message and closes silently, which is why sends that worked during the event start failing the next morning.
Put the checkbox on the registration form, with a specific promise about what they get. Placement is the single biggest determinant we have measured. At CPME31, where the opt-in was on the form and the organizer promoted it on site, 159 of 209 Speed Business participants opted in. At a content-led event without that promotion, opt-in was 274 of 1,596, or 17%.
It can be, and it is your responsibility rather than Meta's. You need a lawful basis, which in practice means explicit opt-in captured at registration; a privacy notice naming the processors involved; a documented retention period; and a working opt-out honoured on the first request. Attendees do ask about this. At CPME31 one asked in detail how the data was sourced and processed.
It depends almost entirely on promotion, and the honest range is wide. At CPME31, 92% of delivered templates got a reply (113 of 123) and 65% across the full funnel. At FOST Singapore, with lighter promotion, 38% of delivered templates got a reply though 67% were read. Where an organizer did not promote it at all, engagement sat near 6%. Treat any vendor quoting one number without a denominator with suspicion.
For reach, usually yes, because there is nothing to install and the app is already open on the attendee's phone. For agenda, maps and session bookmarking, no. An event app is the right tool and WhatsApp does not replace it. The two solve different problems and most events end up running both.
Volume is higher and the expected reply time much shorter than an email inbox, so the staffing most organizers plan for is not enough. FOST Paris ran 2,634 attendee questions through the channel, each answered in under three seconds, and at FOST Singapore 78% were navigation questions rather than networking ones. Whatever answers has to be working from your real programme and floor plan, which is the job Sena does.
The anonymized sponsor report from a live 1,306-attendee deployment, in your inbox.
A real, anonymized sponsor report from a live deployment. No newsletter, no sequence.
Setting this up yourself and stuck on one of the failure modes above? Ask us and we will tell you what we did, whether or not you buy anything.