blog / Automation

How to measure a WhatsApp campaign: the metrics that actually matter

heysteff is the AI platform for customer service and sales over WhatsApp, Instagram, Messenger, Gmail and Shopify; Steff is the AI agent that runs it. Knowing how to measure a WhatsApp campaign is what separates a business that repeats what works from one that blasts its list out of habit. Most teams stop at read counts, which are the easiest number to find and the least useful; this article covers which metrics to track and in what order, how to attribute revenue honestly, and which costs usually get left out.

What does it mean to measure a WhatsApp campaign properly?

Measuring a WhatsApp campaign properly means answering one question: did it create more value than it cost? Deliveries, reads and clicks are intermediate steps that explain why the answer is yes or no, but they are not the answer.

A WhatsApp campaign (often called a broadcast) is a proactive message sent to a list of contacts, usually with a Meta-approved template, to announce an offer, a launch or a reminder. Because the customer did not message you first, every send has a direct cost and an attention cost: you are interrupting someone in their personal chat. Good measurement accounts for both what you gain and what you put at risk.

Think of the campaign as a five-step funnel: delivered → read → replied or clicked → converted → revenue. Each step has its own rate, and finding the step where most people drop off tells you what to fix. Low delivery points to the list. Reads without replies point to the message. Replies without purchases point to the conversation or the offer. Meta lays out the same rates in its guide to measuring WhatsApp campaign effectiveness.

WhatsApp campaign metrics: which ones matter and which ones mislead

The metrics that matter are reply rate, conversion rate and revenue per message sent; read rate is a useful diagnostic but a misleading goal. A message can be read by nearly everyone and still sell nothing if the hook is catchy but the offer is wrong for that person.

Reply rate (replies divided by delivered messages) is the metric most specific to WhatsApp. Unlike email, people answer on WhatsApp: they ask about sizes, shipping, price. Every reply is an open sales conversation, and a campaign with plenty of replies but few sales usually means the team could not respond fast enough, not that the campaign failed.

Conversion rate (purchases divided by delivered messages) ties the send to the business. Calculate it on delivered, not sent, so invalid numbers do not drag the campaign down.

Read rate has a technical ceiling: not every user shares read receipts, so the real number differs from what you see. Use it to compare two sends against each other, not as an absolute figure. And be wary of clicks on their own: a click without a purchase might be curiosity, or a shopper who checked the price, saw the shipping cost and left.

How to attribute sales to a WhatsApp campaign without fooling yourself

To attribute sales to a campaign you need three things: a tracked link, an attribution window you set before sending and, when volume allows, a holdout group. Without them, every sale that week looks like a campaign win.

A tracked link uses UTM parameters, tags added to the end of a URL so your analytics knows where the visit came from. Google explains how to build them in its help article on custom campaign URLs. Use a different tag per campaign and per message variant so you can compare later.

The attribution window is the period during which a purchase counts as an effect of the campaign. A short window fits a deadline-driven offer; a longer one fits a pricier product launch. What matters is fixing it in advance rather than picking whichever window flatters the result.

A holdout group is a random slice of your list that does not receive the campaign. If the group that got the message clearly bought more than the holdout, that gap is what the campaign actually generated. It is the only honest way to discount customers who would have bought anyway, and on a list of repeat buyers there can be a lot of them.

Watch for sales that never touch the link. A customer who replies, chats with your team and pays with a payment link sent in the conversation is still a campaign sale. Abandoned cart flows have the same blind spot, which is why cart recovery on WhatsApp should track both clicks and attributed sales. If you only measure the original link, you undercount exactly the sales that needed a conversation.

The real cost of a campaign: what usually gets left out

The real cost of a WhatsApp campaign is what Meta charges for the messages, plus the team time spent handling replies, plus the contacts you lose to opt-outs and blocks. Most ROI calculations only include the first.

On the official API, Meta charges per delivered template based on its category and the recipient's country; current rates are in its platform pricing documentation. Picking the wrong template category changes that cost, so check whether your message is marketing or utility before sending; our guide on WhatsApp template categories walks through it.

Team time is the invisible cost. A campaign that triggers a wave of replies at 8 p.m. on a Friday, with no one on shift, turns interest into frustration. Count the hours the campaign consumed in replies and add them in.

Opt-outs carry a delayed cost: every person who asks you to stop, or who blocks your number, is a contact you cannot reach next time. Track opt-outs per campaign alongside revenue. If a campaign sold well but drove a lot of opt-outs, you probably reached the wrong people. We go deeper on that in WhatsApp campaigns without getting blocked.

Example: a skincare brand measuring its first campaign

Picture a direct-to-consumer skincare brand launching a new serum. It messages customers who bought facial care in recent months. Before sending, it sets aside a random slice of that list as a holdout, builds a UTM link and decides to count purchases over the next seven days.

Three days in, the dashboard shows lots of reads and few clicks, and the first instinct is to call it a flop. But the inbox tells a different story: many customers replied with questions about sensitive skin and whether the serum works with products they already use. Several bought after that conversation, through a payment link in the chat, without ever touching the UTM link.

When the window closes, the team compares: the group that received the campaign bought the serum noticeably more than the holdout, opt-outs were low and most sales came from conversations rather than clicks. The practical takeaway is not "WhatsApp works" but something more useful: for products that raise questions, the campaign is the start of a conversation, so next time it should go out when someone (or an agent) is available to reply quickly.

Common mistakes when measuring WhatsApp campaigns

The most common mistake is measuring only what the send dashboard shows: delivered and read. Those numbers tell you the message arrived, not that it worked.

The second is comparing campaigns that are not comparable, such as a clearance sale to the whole list versus a back-in-stock alert to people who asked for that item. Compare campaigns with similar goals and audiences, and benchmark against your own history more than industry averages.

The third is not separating service replies from sales replies. A campaign can surface open complaints, like late orders or unresolved returns, in the replies. That is not a campaign failure, but it is an operations signal worth reading on its own.

The fourth is changing several things at once. If the next campaign has new copy, a new send time and a new audience, you will not know what moved the result. Test one variable at a time, or split the list in two with a single difference.

If you want measurement built into the process instead of living in a spreadsheet, heysteff's automation flows and campaigns keep replies in the same inbox your team already works in, mark people who ask to stop as opted out, and keep the same contact from getting several campaigns on the same day.

Frequently asked questions

What is a good reply rate for a WhatsApp campaign? There is no universal number: it depends on your industry, your existing relationship with the contact and whether the message invites a reply. Your own history is the most useful benchmark; measure your first campaigns and improve against them rather than against vendor-published averages.

How do I know if a customer bought because of the WhatsApp campaign? Use a UTM-tagged link, set an attribution window before sending and also count sales closed inside the conversation. To know how much the campaign really added, compare against a holdout group that did not receive it.

Can I see who read a WhatsApp campaign message? The API reports each message's status (sent, delivered, read), but it only marks a message as read when the user shares read receipts. That makes read rate an approximation, so give more weight to replies and purchases.

◆ how heysteff does it

Replies to every campaign land in the same inbox Steff or your team already handles, so the sale that closes in conversation is not lost between the send and the payment.

Related

◆ next step

Send your next WhatsApp campaign with every reply in one inbox.

Start free → Get a demo See pricing