Ask any business owner how many conversations reached the company’s WhatsApp last month and how many turned into sales. The silence that follows is the answer — and it’s costly. The channel where the sale actually happens is usually the only part of the operation without any numbers, and therefore the only one where problems go unnoticed for months.
In this article
Why no one measures WhatsApp
Companies measure revenue, inventory, costs. But the channel where the sale actually happens is usually the only place without any numbers. The reason is simple: in the regular app, there is nothing to measure. It shows conversations, not metrics. No one can say how many conversations arrived last month, how many went unanswered, or how many turned into sales, because that information was never recorded anywhere.
The consequence is a specific and costly blindness: the company knows that "traffic dropped," but doesn’t know if the number of contacts dropped, if the response rate dropped, or if the closing rate dropped. These are three problems with completely different solutions — invest in ads, hire more agents, or train salespeople — and without numbers, the choice becomes a guess.
The classic symptom: the meeting where someone says "this month was weak" and the response is "I think the team is taking too long to respond." Two opinions, no data, and the usual decision — increase ad budget to compensate for a problem that might actually be in customer service.
The seven key numbers
1. New conversations per period
How many new contacts arrived in the week or month. This is the top of everything: if this number drops, none of the others can save the result. Separate by source — ads, website, referral, Google profile — because this is what allows you to know where to invest.
2. Time to first response
How long, on average, the customer waits for the first human response. This is the metric with the greatest effect on conversion and the easiest to improve. Also measure the worst case, not just the average: an average of eight minutes can hide conversations that waited five hours.
3. Conversations without response
How many conversations went completely unanswered during the day. This is the most uncomfortable number on the list and the one that yields the most money when it decreases — each unit here is a customer who reached out to your company and was not served.
4. Conversation conversion rate
Out of every 100 new conversations, how many turned into sales. This is the quality indicator of customer service and qualification of incoming contacts. If volume rises and this rate falls, the problem is usually with the contact source, not the team.
5. Revenue per conversation
Revenue for the period divided by the number of new conversations. This translates WhatsApp into money and answers the key question: how much is each new conversation worth to your company? This number justifies investing in ads — or stopping investment.
6. Time to closing
How many days, on average, between first contact and sale. It serves two practical uses: to size follow-up cadence and to predict how much of this month’s incoming contacts will turn into revenue next month.
7. Conversations stalled in the funnel
How many opportunities have had no movement for more than X days, by stage. This is the only metric on this list that points to immediate action: it’s not diagnosis, it’s a task list. Proposals sent more than five days ago without response are money on the table waiting for someone to follow up.
A real month, number by number
Service operation with three agents. The closed month looked like this:
Month funnel
Conversion on conversations: 11%. Conversion on quotes sent: 31.8%. Revenue per conversation: R$ 41.68.
Now the reading, where the value lies. 124 conversations went unanswered — 20% of all that arrived. Applying to this group the same conversion rate as the others, there would be about 14 more sales, or R$ 5,320 per month. More than R$ 63 thousand per year lost not due to price, not due to competition, but because messages were never opened.
Notice that no new campaign would enter this calculation. The money was already there, and only appeared because someone counted the unanswered conversations. While this number doesn’t exist, the company’s natural reaction would be to hire more ads to bring more 124 conversations — which would probably also go unanswered.
When the month is weak: diagnosis in four questions
The value of having the numbers appears precisely in the bad month. Instead of discussing impressions, go through the funnel from top to bottom and stop at the first step that dropped — that’s where the problem is, and each step calls for a different action.
| What dropped | What this indicates | Where to act |
|---|---|---|
| New conversations | Source problem, not team problem | Ads, Google presence, referrals |
| Conversations answered | Lack of capacity or organization | Distribution, schedule scale, first automatic response |
| Quotes sent | Conversation that doesn’t advance | Qualification and service script |
| Sales closed | Closing or follow-up | Follow-up cadence and objection handling |
| Average ticket | Discount or product mix | Discount policy and value presentation |
Note that only the first line is solved with more marketing investment — and it’s precisely the solution almost every company applies to any of the five. Bringing more conversations when the bottleneck is response or closing increases waste instead of revenue.
Continue reading for free
Fill out once and unlock this and all other blog articles. We don’t charge anything — we just want to know who we’re writing for.
Takes 20 seconds. Prefer to talk directly on WhatsApp?
The weekly dashboard: 10 minutes, every Monday
A metric that doesn’t become routine doesn’t change behavior. The format that works is short and always the same:
- The four numbers of the week side by side with the previous week: new conversations, time to first response, conversations without response, and sales closed.
- One question per number that worsened. Not an explanation — a question. "Why did the response time rise from 6 to 19 minutes?" usually reveals simple things: someone on vacation, message peak at a certain time, campaign that brought volume without warning.
- The list of stalled conversations, distributed immediately among those responsible. It’s the only part of the meeting that generates tasks.
- One goal for the following week, and only one. A company that tries to improve seven indicators at once improves none.
Measuring by person without turning into surveillance
Numbers by agent are useful and can quickly sour the environment if used as a public ranking for judgment. Three precautions make a difference.
Compare similar contexts. Those who handle cold ad leads and those who handle referrals are not playing the same game. Comparing the two conversion rates without this distinction produces unfairness and distrust.
Use the number to train, not to punish. A salesperson with good conversion and few conversations might be slow; a salesperson with many conversations and low conversion might be rushing qualification. These are different development conversations, and none require embarrassment.
Show the number to the person before showing it to the group. Most correct themselves when they see their own data — and start trusting the metric instead of trying to game it.
Mistakes that make measurement useless
- Measure everything. Twenty indicators on a dashboard lead to zero decisions. Seven is already generous; start with four.
- Confuse message with conversation. Counting messages inflates volume and says nothing about opportunities. The unit that matters is the conversation, or better, the customer.
- Look only at average response time. The average hides the bad case, and it’s the bad case that loses the sale.
- Compare months with different numbers of business days. February always looks worse than March. Normalize by business day when the difference is relevant.
- Record manually. A spreadsheet filled out by hand lasts three weeks. The data needs to come from the service system itself, without extra work — it’s a basic function of a WhatsApp CRM.
- Measure and don’t act. A nice dashboard that generates no tasks on Monday is expensive entertainment.
Start with one number
If your company measures nothing today, don’t try to build the full dashboard. Choose one: conversations without response. It’s the easiest to gather, the most uncomfortable to see, and almost always the one that returns the most money in the first month. Count on Monday how many conversations from the previous week went completely unanswered, multiply by your conversion rate and average ticket. The result usually ends the discussion about whether it’s worth organizing customer service.
After that, add time to first response, and only then the rest. The company that consistently measures four numbers decides better than the company that has fifteen indicators on a dashboard that no one opens — because measuring is not about having data, it’s about having, every Monday, a conversation different from the one that starts with "I think the month was weak."
Frequently Asked Questions
Which metrics should be tracked in WhatsApp customer service?
Seven cover a commercial operation well: new conversations per period, time to first response, conversations that went unanswered, conversation-to-sale conversion rate, revenue per conversation, average time to closing, and number of stalled conversations by funnel stage. If the company measures nothing today, start with conversations without response and time to first response — these two usually return money fastest.
How to measure response time on WhatsApp?
Time to first response is the interval between the customer’s initial message and the first human response. Track the average and also the worst case for the period, because the average hides conversations that waited hours — and those are precisely the ones that lose sales. The regular app cannot track this; the data needs to come from a system that records the time of each message.
What is a good conversation-to-sale conversion rate?
It varies too much by sector, ticket, and contact source to have a universal number. The correct use of the metric is comparative: track your own rate over months and by contact source. If conversation volume rises and the rate falls, the problem is usually the quality of incoming contacts; if volume is stable and the rate falls, first look at response time and follow-up.
Is it worth measuring performance by agent?
It is, as long as it’s used for training and not embarrassment. Compare only people who handle the same type of contact — those who receive referrals and those who receive cold leads are not in the same game — and show the number first to the person themselves. Low conversion with many conversations and high conversion with few conversations indicate different problems, each requiring a specific development conversation.
Can these metrics be tracked in WhatsApp Business?
Not usefully. The app shows conversations, not metrics: there is no time to first response, conversion rate, funnel, or history by salesperson. To have these numbers without manual work, service must happen inside a system that records each event — otherwise, any measurement becomes a spreadsheet filled out by hand, and spreadsheets filled out by hand usually last three weeks.
Your customer service numbers, without spreadsheets
VeyloCRM records every conversation, calculates response time, conversion, and funnel by salesperson — and shows every week where sales are leaking.