WhatsApp Open Rates Are 98%. Most Brands Still Waste Them

Before you pour budget into WhatsApp campaigns, it's worth knowing where buying is actually heading. Read: How E-commerce Brands Can Sell Directly Through ChatGPT, Claude, and AI Assistants.

Every WhatsApp pitch starts with the same hook. Open rates near 98 percent, email stuck somewhere in the low twenties, so move your budget over and watch the orders come in. The open rate part is broadly true. The conclusion people draw from it is where most ecommerce brands quietly lose money.

A near-universal open rate tells you almost nothing about whether someone actually read your message or did anything after. All it confirms is that the notification fired and your text showed up on a screen. That is a visibility number wearing a conversion costume. Anyone who has run a few WhatsApp campaigns knows the gap already. The broadcast went out, the dashboard lit up green, and the revenue line sat there.

This piece is about that gap. The real difference between a plain broadcast and an AI-personalised campaign, why the open rate stopped being the interesting number a long time ago, and how to spend on WhatsApp so opens turn into orders instead of unsubscribes. If you are a founder or growth marketer evaluating this channel, the goal here is to save you a quarter of wasted sends.

What WhatsApp Open Rates Actually Measure

WhatsApp open rates run high for structural reasons, not because your copy is sharp. The app sits on the home screen, notifications are on by default for most people, and the platform still feels like a personal space rather than a promotions folder. A message lands, the badge appears, the eye goes to it. That is the entire mechanism.

Email cannot match that, partly because email opens depend on a tracking pixel loading, and privacy features and corporate mail systems break that pixel constantly. So email open rates are both lower and shakier as a measurement. WhatsApp does not have that problem, which is why the comparison looks so lopsided on a slide.

Now the part the decks skip. An open on WhatsApp can mean someone swiped the notification away just to clear the badge. It can mean they glanced, thought "promo," and closed it. None of that is engagement. When a brand tells me their open rate is sky high, I ask the same thing every time. What is your reply rate? How many people are tapping the link? How many are blocking you? Those answers tell you whether the channel is working. The open rate only tells you the pipes are connected.

The number that actually deserves your attention is how many people block or report you. WhatsApp watches that closely, and it feeds straight into how the platform treats your account. A campaign that gets opened by nearly everyone and reported by a noticeable slice of them is a campaign slowly strangling your own reach.

If your WhatsApp reporting begins and ends with open rate, you are flying on the one instrument that cannot tell you whether you are about to crash.

If WhatsApp Open Rates Are 98%, Why Are My Campaigns Stuck at 70-80%?

Open your campaign dashboard and the 98 percent never shows up. Instead you see something in the 70s, maybe low 80s, and it does not match anything you have read about WhatsApp. The gap is not a reporting bug. The 98 percent and your 70-80 percent measure two different things, and the one bothering you is not even the open rate.

The famous 98 percent is a read rate, and it only counts messages that were actually delivered. It says: of the messages that landed on a phone, almost all got opened. Your 70-80 percent is most likely your delivery rate, an earlier step entirely. It says: of the messages you tried to send, only this many actually reached a phone. A message that never gets delivered can never be opened, so it never enters the 98 percent calculation. The high number quietly excludes your biggest problem, and the real question is not why your open rate is low, it is why your messages are not getting delivered. For most ecommerce brands this past year, the answer comes down to one Meta change.

Frequency capping is the main culprit

Meta now limits how many marketing messages a single person can receive in a day, counted across every business, not just yours. Industry reporting and BSP documentation put this at roughly two marketing template messages per user per 24 hours before Meta starts refusing to deliver more. The cap is dynamic and Meta does not publish the exact rules, but the effect is consistent: if your customer already got promotional messages from two other shops this morning, yours may simply not be delivered, no matter how clean your template or list is.

When that happens you are not blocked and your account is not broken. The recipient is what the ecosystem calls saturated, and the platform returns a specific error rather than delivering. This is why so many brands watched delivery quietly fall from the 80-90 percent range into the 50-70 percent range over the past year without changing anything on their end.

The other reasons messages go undelivered

Frequency capping is the big one, but a few other things drag delivery down too. A low quality rating means Meta filters more of your marketing messages before they reach people. Recipients who blocked your number receive nothing, and a high block rate worsens the quality rating that controls everything else. Dead numbers on an old list fail silently. And messages Meta judges "less likely to be read" get dropped first, which means stale, generic broadcasts are exactly the ones that disappear. Stack those together and a 70-80 percent delivery rate is not a mystery. It is the predictable result of sending marketing-category messages to a broad list in a system built to throttle exactly that.

What this actually tells you to do

The fix is not to send more to make up the shortfall, because that pushes you straight into the frequency cap and the quality penalty that make the problem worse. Send messages people want and reply to, because a reply opens a window where you are no longer competing against the marketing cap. Utility messages people expect, like order and shipping updates, sail through where promotional blasts get filtered. Personalised, behaviour-triggered messages avoid the cap better than mass broadcasts because they reach fewer people at moments those people care about. Clean your list so dead numbers stop dragging the rate down. It is just working with how the platform decides what to deliver instead of against it.

Broadcast vs AI-Personalised Campaigns: The Real Split

That last point, personalised messages getting through where broadcasts get filtered, is worth slowing down on, because most teams think the choice is between sending one message to everybody and sending a slightly different message to a couple of segments. That undersells it badly. The real difference between broadcast and AI-personalised campaigns is what sets the message off and what the message knows about the person on the other end.

A broadcast is a push. You write one template, choose a list, and fire. Everyone gets the same thing at the same moment, regardless of where they sit in their relationship with you. It is the WhatsApp version of a billboard. Like billboards, broadcast advertising campaigns earn their keep for genuine mass-relevance moments and waste money on everything else.

An AI-personalised campaign behaves more like a conversation the system runs on your behalf. The trigger is behavioural. A cart has sat untouched for a couple of hours. A repeat buyer has gone quiet for two months. A product someone kept eyeing just dropped in price. The content shifts based on what that person actually did or bought, and the timing leans toward when they tend to open things. Nobody hand-writes each message. The personalisation runs off your customer data and a layer of logic that decides what to send and when.

Where broadcast still earns its place

I am not anti-broadcast. A real sitewide sale, a limited drop with honest scarcity, a festival-rush notice about store hours, these are mass-relevance events, and a broadcast is the clean tool for them. The mistake is reaching for broadcast as the default for everything because it is the quickest thing to set up in a dashboard.

A bootstrapped D2C skincare brand I spoke with last year was sending one broadcast every week to their whole opted-in list. Same offer, same time. Opens stayed high the entire time. Revenue per send drifted down week after week, and blocks crept up, because most of the list had no reason to buy that week. They were paying to slowly train their best customers to mute them.

Where AI personalisation pulls ahead

That same brand cut their weekly blast drastically and moved the effort into a small set of triggered flows. Cart recovery, a gentle reorder nudge timed to roughly when each customer runs low, and a win-back for people who had gone dark. Fewer messages went out overall. Revenue per message climbed, and the block rate settled, because people were getting messages that matched their actual moment.

That is the whole argument in one example. Personalisation is not about being clever. It is about not sending people things they have no reason to care about, which protects your conversion rate and your sending reputation at the same time.

WhatsApp Messaging Rates Change the Math

The platform has shifted toward charging per message in a way that makes each unnecessary send a real, traceable cost rather than a rounding error hidden inside a bundle. Marketing, utility, and authentication messages are priced differently, and the marketing category, the one you most want to scale, is typically the least forgiving on cost. So the broadcasts you are tempted to send more of are exactly the ones that stay expensive per message no matter how many you fire. Cost also depends on where your customers are, so a blast that feels almost free on a domestic list can get uncomfortable on an international one.

The other half of the math is how much you are even allowed to send. New accounts do not message the whole list on day one. The platform starts you conservatively and raises your ceiling as you prove you send well, keeping blocks low and engagement healthy, not by pushing harder. A poorly targeted blast that triggers blocks drags down your quality rating, which slows your climb to higher limits and gets more of your messages filtered before delivery. Your sending limit is not a fixed pipe, it is a score you move with every campaign. WhatsApp rate limits behave like a quality system, not a fixed cap.

Put the cost picture next to the sending-limit picture and you get the same answer twice. When every message is billed, promotional sends never get a real discount, and account quality decides how far you can scale, the brand sending fewer, sharper messages wins on conversion and pays less.

The AI Rule Most Brands Missed

If your personalisation plan is to bolt a general-purpose chatbot onto your WhatsApp number, reconsider. The platform tightened its stance here, and open-ended AI assistants without a defined job are no longer welcome on business numbers. Only task-specific agents, built for clear purposes like support or product help, are allowed.

For ecommerce this is barely a constraint, because the AI that actually drives revenue is task-specific anyway. A flow that recommends products from your catalogue, answers an order question, or recovers a cart has a defined purpose. A do-anything assistant parked on your business line does not, and now it is a policy risk on top of being a weak revenue tool.

The useful "AI" here means smart triggering, dynamic content, and timing that adapts to behaviour, all running off your own customer data, not handing the conversation to an open-ended model and hoping it behaves. We have watched teams try to automate every edge case and end up with something that handles nothing well. The point of automation is to remove the repetitive, predictable work, not to replace judgement on the messy stuff.

What Good Looks Like in Practice for Ecommerce

A real WhatsApp programme for an ecommerce brand looks less like a broadcast calendar and more like a small set of always-on flows with the occasional genuine broadcast on top.

The flows that consistently earn their place are cart recovery soon after abandonment, order and shipping updates because they are cheap and genuinely wanted, a reorder nudge for anything people use up, and a win-back sequence for customers who have gone quiet. These are behavioural, personalised, and they run without anyone touching them after setup. They do the boring, repetitive work that drives most of the revenue.

Broadcasts sit on top of that, reserved for the handful of moments a year when the same message really is relevant to your whole list. A 40-property hotel group I spoke with runs only a couple of true broadcasts each quarter, festival packages and an end-of-season clearance, and routes everything else, booking confirmations, pre-arrival nudges, review requests, through triggered flows. Their cost per message dropped and their block rate stayed near zero, because the only mass messages they send are ones people are already expecting.

The honest caveat is that none of this runs fully on autopilot. Personalisation is only as good as your customer data, and messy data produces confidently wrong messages, which land worse than generic ones. Triggered flows still need a human checking the numbers every couple of weeks, watching the block rate, and quietly retiring templates the platform is starting to dislike. And there are always moments, a delayed order, an angry reply, a complicated return, where the right move is to put a real person into the conversation quickly rather than let automation grind on.

If your inbox is already live with us, the next workflow to set up is cart recovery, because it tends to pay for the entire setup faster than anything else you could build.

The Realistic Bottom Line for 2026

The famous WhatsApp open rate is real, and it is also the least useful number on your dashboard. It confirms the pipe is connected, not that anyone cared. The brands wasting WhatsApp are the ones treating that number as proof their broadcasts work, then sending more of them until their best customers start blocking and the platform starts throttling.

The split that matters is broadcast versus AI-personalised campaigns. Broadcast is the right tool for the rare mass-relevance moment and an expensive habit everywhere else, especially now that messages are billed individually and the promotional category stays stubbornly costly. Personalisation, meaning behavioural triggers and dynamic content running off clean data, sends fewer and sharper messages, which protects both your conversion rate and the account quality that decides whether you can scale at all.

Spend on WhatsApp like the messages cost money, because they do, and treat your account quality as something you can lose. Get those two instincts right and the open rate stops being something you have to worry about.

FAQ

Is the 98% WhatsApp open rate accurate or just marketing hype?

The figure is broadly accurate, but it is widely misread. WhatsApp open rates run high because notifications are on by default and the app sits front and centre, so messages surface reliably. What an open does not tell you is whether the person read, cared, or acted. A swipe to clear a notification can register as an open. Treat it as a visibility metric, and judge your campaigns on reply rate, link taps, and block rate instead. Those reveal whether the channel is actually doing anything for you.

Why is my WhatsApp delivery rate stuck at 70-80% when open rates are 98%?

Because those numbers measure different stages. The 98 percent read rate only counts messages that were delivered, so it hides the problem entirely. Your 70-80 percent is a delivery rate, and it is mostly being held down by Meta's frequency capping, which limits how many marketing messages a person receives per day across all brands. If a customer already hit that cap from other businesses, your message will not be delivered. Low quality ratings, blocks, and dead numbers pull it down further.

When should we use a broadcast instead of a personalised campaign?

Use a broadcast only when the same message is genuinely relevant to your entire opted-in list at the same moment. A real sitewide sale, a limited drop, a festival or store-hours notice. Those are mass-relevance events where one message to everyone is honest. For anything tied to individual behaviour, browsing, cart status, purchase history, lapsed activity, a triggered personalised flow converts better and protects your block rate. The common mistake is defaulting to broadcast for everything simply because it is the fastest thing to set up.

How do WhatsApp messaging rates affect campaign cost?

Messages are charged in a way that makes each unnecessary send a real cost rather than a freebie, and promotional messages tend to be the least forgiving category to scale. Cost also depends on where your customers are located, so the same campaign can be cheap to one list and pricey to another. The practical effect is that high-volume broadcasting is more expensive than it looks, and tighter, behaviour-based targeting usually wins on pure economics before you even count conversion.

What WhatsApp rate limits should a new ecommerce brand expect?

New accounts start with a conservative sending ceiling and earn higher limits over time by sending well, low blocks and reports, healthy engagement, rather than by sending aggressively. Poor targeting that triggers blocks lowers how the platform rates your account, which can slow or freeze your progress to higher limits and even reduce delivery of messages you have already paid for. Think of your sending limit as a score you affect with every campaign, not a fixed pipe you can hammer.

Can we just add a general chatbot to our WhatsApp number for personalisation?

No, and the platform now restricts it. Open-ended AI assistants without a defined purpose are not allowed on business numbers. Only task-specific agents built for clear jobs, like support or product help, are permitted. For ecommerce this is not a real limitation, because the AI that drives revenue, smart triggers, dynamic product content, behaviour-based timing, is task-specific anyway. The useful AI layer decides what to send and when. It does not host a do-anything conversation, and a do-anything bot is now both a policy risk and a weak revenue tool.

What is the minimum setup needed to see real results from WhatsApp?

Start with one behavioural flow rather than a broadcast calendar. Cart recovery is usually the fastest to pay for itself: trigger a personalised message a short while after abandonment, include the product and a direct path back. Add order and shipping updates, since they are cheap and genuinely wanted. Get those two right, keep your customer data clean, and watch your block rate before layering on anything more. You do not need a big stack to begin. You need clean data and one well-timed flow.

How do small teams run personalised WhatsApp campaigns without burning out?

The point of AI-personalised campaigns is that they run without daily effort once configured. Set up your core triggered flows, cart recovery, shipping updates, reorder nudges, win-back, and let them fire on behaviour pulled from your customer data. The recurring human work is light: check the numbers every couple of weeks, watch block and quality signals, retire weak templates, and stay reachable for the conversations that genuinely need a person. Small teams get into trouble when they try to send constant manual broadcasts instead of letting good flows handle the repetitive work.