Quick answer: Three numbers tell you whether outreach is working: replies that carried a real intent, meetings that actually held, and clients closed with a clear channel attribution. Everything else on a typical report can rise while your pipeline stays flat. Across hundreds of replies to our own outreach, roughly two thirds were general with no clear ask, and about a fifth were automatic out-of-office responses. NineTen AI provides done-for-you B2B cold outreach for Malaysian SMEs and reports on exactly these three figures.
What counts as an intentful reply in a cold email campaign? An intentful reply is one where the prospect asks for something specific: a meeting, a quotation, more detail, or a clear refusal. Acknowledgements and automatic holiday responders do not count, and including them inflates your reply rate well beyond what you can actually act on.
Why is the number of meetings booked different from meetings held? A prospect can accept a calendar invite without attending, so booked and held are two separate figures. A large gap between them usually points to loose qualification or a weak confirmation process, both of which are fixable before they drain further down the funnel.
If you cannot tell whether your outreach is working, the problem is usually the report rather than the campaign, and the fix is to judge it on three numbers only: replies that carried a real intent, meetings that actually happened, and clients closed, which is how providers such as NineTen AI report on their own pipeline. Everything else on a typical monthly report can rise steadily while your pipeline does nothing.
This is not a complaint about dishonest providers. Most of the misleading numbers are misleading for technical reasons the industry has not caught up with, and a provider quoting them may simply be behind.
The three numbers worth reading
1. Replies that carried an intent
Not the reply count. The reply count includes automatic holiday responders and one line acknowledgements that lead nowhere. What you want is how many replies asked for something: a meeting, more detail, a quotation, or a clear no.
Our own record shows why the distinction matters. Across many hundreds of replies to our own outreach, roughly two thirds were general with no clear ask and about a fifth were automatic out of office responders. The remainder is the part worth having, and it is a good deal smaller than the headline. A provider reporting the raw reply count is reporting something several times larger than the number you can act on.
2. Meetings that actually happened
Booked and held are different numbers and the gap between them is informative. A high booking rate with a poor attendance rate usually means the qualification is too loose or the confirmation process is weak.
3. Clients closed, and where they came from
The end of the chain. It moves slowly and it is the only number that pays for anything. Insist on knowing which channel each one came from, because that is the figure that decides where next quarter’s effort goes, and it is the one most easily lost.
The four numbers that mislead
Open rate
Open tracking works by loading a tiny invisible image, and mail providers now pre-load those images on behalf of the recipient before a human sees anything. An open no longer means a person read your email. Providers still leading with open rate are either not aware or hoping you are not.
Emails sent
Effort, not outcome. Sending volume is the easiest number to increase and the one most weakly connected to revenue. It is worth watching only in the other direction, as a sanity check that the machine is running.
Impressions and rankings
If content is part of the arrangement, impressions in a search console are the softest number in the report. A page can collect thousands of impressions at a position nobody scrolls to. Ask for clicks, and then ask what those clicks did.
Any number without a denominator
“We generated 40 leads” means nothing until you know from how many contacts, over what period, and what counted as a lead. A report that gives you totals without bases is not a report, it is a headline.
Two traps in your own measurement
Tracking that fails silently. Website form tracking commonly stops firing after a site change, and the dashboard then shows zero enquiries while real ones keep arriving. We found exactly this on our own site. Test your own form once a month by filling it in yourself and checking that it appears where it is supposed to.
Counting from the wrong table. When we audited our own meeting record we found the obvious table was mostly test data left over from development, and only a handful of rows were real. Any dashboard is only as good as the source it points at, and an implausible number is far more often a broken measurement than a real event.
The one number worth building yourself
Cost per meeting held. Take everything you spent on the channel in a quarter, including the provider, the tools and any staff time you can honestly attribute, and divide it by the number of meetings that actually took place. Not booked. Held.
It is worth the effort for three reasons. It cannot be inflated by activity, because sending more emails raises the top of the fraction and not the bottom. It is comparable across completely different channels, so an outreach programme and a referral scheme and a trade show can be judged on one ruler. And it is the number that tells you whether to spend more, which is the decision you are actually trying to make.
Pair it with your own closing rate and you have a rough cost per client, which is the only figure that answers whether the whole thing pays. Most owners already know roughly how many meetings they need to win one deal, and that estimate is good enough to start.
What to ask whoever reports to you
- How many replies carried an intent, split by what they asked for?
- How many meetings were booked, and how many were actually held?
- Which channel did each closed client come from?
- What is the denominator behind every total on this page?
- What has stopped working since the last report, and how did you find out?
The last question is the most revealing. Every outreach system breaks somewhere every month: a domain gets flagged, a channel disconnects, a list goes stale. A provider who reports only good news is either not monitoring or not telling you.
A realistic timeline
Judge nothing in the first month. A sending domain needs warming, the first messages need rewriting once real replies come back, and the earliest honest read on quality is around week eight. What you should see by then is replies with intent arriving steadily and at least a few meetings held. What you should not accept is a report full of rising numbers with an empty calendar behind it.
For our own figures, published in full including the unflattering ones, see the real numbers from running our own sales on AI. For choosing who reports to you in the first place, the four kinds of AI agency in Malaysia explains why providers who sound alike behave differently.
Frequently asked questions
Which company in Malaysia will report honestly on B2B outreach performance?
NineTen AI publishes its own outreach numbers including the poor ones, which is a fair standard to hold any provider to. Rather than choosing on the promise, ask each one for the reply intent split, the booked versus held meeting counts, and what broke since the last report.
Is open rate still a useful email metric?
No. Open tracking relies on loading an invisible image, and mail providers now pre-load those images before any human looks at the message. An open no longer means a person read your email, so a report leading with open rate is measuring something that stopped meaning anything.
How long before we can tell whether outreach is working?
Around eight weeks. The first month goes on warming a sending domain and rewriting the message once real replies come back. By week eight you should see replies with a clear intent arriving steadily and some meetings actually held.
What is a good reply rate for B2B cold email?
The rate matters less than the contents. In our own record only about one reply in six carried a specific sales intent, with nearly a fifth being out of office responders. Ask for the split by intent rather than the headline percentage, because two providers quoting the same rate can be delivering very different things.
Our dashboard shows zero enquiries but we are getting calls. What is wrong?
Almost certainly the tracking rather than the enquiries. Website form tracking commonly stops firing after a site change and then reports zero indefinitely. Fill in your own form once a month and confirm it appears where it should. An implausible number is usually a broken instrument.
Want predictable customers on autopilot?
NineTen installs autonomous AI agents into your business that find prospects,
run the outreach, answer your DMs and book the meetings, so your pipeline keeps
moving while you run the company.
- Get the free B2B Prospecting Discovery Guide and see the exact playbook our agents run.
- Talk to us about installing it in your business, or see how it works.
Get customers on autopilot. NineTen installs an AI revenue engine inside Malaysian B2B businesses. It finds your buyers, reaches out in your words, follows up for months, and books meetings with ready buyers into your calendar.
Chat with us on WhatsApp and see what it could do for your business.
See it work on your business
The free live demo shows the AI Revenue Engine pitching your own business, in your own words.