Quick answer
Track five numbers: positive reply rate, meetings booked, show rate, pipeline per channel and revenue per channel. Ignore open rate and total sends. Read everything split by segment, watch leading indicators weekly and lagging ones monthly, and treat deliverability as the layer underneath all of it. The first set tells you whether outbound works. The second set tells you a comforting story that is often wrong.
The north star
I'm Hlib Storchak. I build and run outbound systems for B2B founders and sales teams, and most measurement arguments I walk into resolve themselves once the hierarchy is written down.
The only outcome that pays the bills is booked meetings that turn into pipeline. Every other metric matters exactly as far as it predicts that. Once a team agrees on the north star, the debate about which dashboard to trust becomes a much shorter conversation.
This is also the test for any new metric someone wants to add. If you cannot say how a number moves booked pipeline, it is not a metric, it is a fact. Facts are fine in a monthly review. They should not be on the weekly scoreboard.
Positive reply rate
Reply rate includes people telling you to go away. Positive reply rate counts only genuine interest: a question, a request for detail, a referral, a "not now but keep me posted". It is the single best early signal that your offer and targeting are right.
Define it once, in writing, and hold the definition. The most common reporting failure I see when I take over an account is a positive reply rate that drifted because whoever was tagging replies got generous during a slow week. If the definition moves, the trend is meaningless, and the trend is the entire point.
Meetings booked and show rate
Meetings booked is the outcome. Show rate is the quality check on it. A high booking rate with a low show rate usually means the meetings were not really qualified, or the reminder process is weak, or the call was booked three weeks out and the interest went cold.
Watch the gap between the two rather than either alone. If bookings climb and shows do not, something in how you are asking is generating polite agreement rather than intent. That is a copy and qualification problem, not a volume problem, and adding more sends makes it worse.
Qualified, not just booked
A booked meeting with someone who cannot buy is an expense, not an outcome. Track qualified meetings as a separate number, with a definition agreed with whoever runs the call, usually fit plus some evidence of a real problem.
This is the number that keeps outbound honest when it is measured against a target. If the only thing counted is meetings booked, the system will produce meetings booked, and some of them will be worthless. Counting qualified meetings aligns the effort with what sales actually wants.
Pipeline and revenue per channel
Trace booked meetings through to pipeline value and closed revenue, split by channel. This tells you where to put the next hour and the next euro, and it ends arguments about email versus LinkedIn versus calls faster than any opinion will.
Attribution does not have to be perfect to be useful. Tag the first-touch channel on every meeting and follow it through the CRM. Rough attribution that everyone understands beats a sophisticated model nobody trusts, and it beats no attribution by a very wide margin.
Leading vs lagging
Positive reply rate is a leading indicator: it moves first and predicts meetings. Revenue is lagging: it confirms weeks or months later. You steer with the leading ones and you confirm with the lagging ones, and confusing the two is how teams end up either reacting to noise or finding out too late.
Open rate has become unreliable. Mail providers pre-fetch images, which fires the tracking pixel without a human ever reading the message. Treat it as noise, not signal, and never make a subject line decision on it alone.
The vanity metrics
Total emails sent, open rate and raw reply rate all flatter you. High sends feel productive. High opens feel like interest. Neither reliably predicts a booked meeting. The working rule: if a number goes up without meetings going up, it is vanity.
Activity metrics are the worst offenders, because they are easy to game and they feel like management. I wrote separately about why SDR activity metrics mislead and about why I stopped chasing open rates. Both come down to the same thing: measuring effort rather than outcome produces more effort.
The metrics under the metrics
Before you conclude your copy is failing, check whether your mail is arriving. Bounce rate, spam placement and domain reputation sit underneath every number above, and when they degrade, positive reply rate falls with no change to the copy or the list.
Keep bounce rate visible on the weekly view. A sudden rise usually means the list went stale or the verification step was skipped. If replies drop and nothing else changed, deliverability is the first place to look, and my deliverability guide covers the checks in order.
Read everything by segment
A blended number hides the truth. One segment might reply at several times the rate of another while the average sits in the middle, telling you nothing useful about either.
Split every metric by segment, by campaign and by sending channel. The pocket worth scaling is almost always visible in the split and almost never visible in the average. This is the single highest-return change most teams can make to their reporting, and it costs nothing but a group-by.
Sample size, or how not to fool yourself
Small samples swing wildly. Two extra positive replies on a hundred sends looks like a doubling and is usually chance. Teams kill good campaigns and scale bad ones on exactly this kind of evidence, every week.
Set a minimum volume per variant before you agree to read the result, write it down before the test starts, and refuse to draw conclusions until it is met. Also change one thing at a time. A test where the list, the subject line and the offer all moved tells you something changed, not what.
How often to check
Look at leading indicators weekly, so you can adjust while a campaign is live. Look at pipeline and revenue monthly, because they need time to mature. Checking revenue daily adds noise and anxiety and changes no decision you would otherwise make.
The scoreboard
| Metric | Type | Cadence | Worth tracking? |
|---|---|---|---|
| Positive reply rate | Leading | Weekly | Yes, primary |
| Meetings booked | Outcome | Weekly | Yes |
| Qualified meetings | Outcome, quality | Weekly | Yes |
| Show rate | Quality | Weekly | Yes |
| Bounce rate | Health | Weekly | Yes, as a guard |
| Pipeline per channel | Lagging | Monthly | Yes |
| Revenue per channel | Lagging | Monthly | Yes |
| Open rate | Vanity | Never | No |
| Total sends | Vanity | Never | No |
Keep one simple weekly view: positive replies, meetings, qualified meetings, show rate and bounce rate, all by segment. That is enough to run the week. Pipeline and revenue belong in the monthly review.
Cost per meeting and payback
Once volume is steady, cost per meeting is the number that tells you whether to keep going. Here is the model. Every input below is an assumption I am showing you deliberately, not a researched figure, so replace all of them with your own before you act on the result.
Assumptions. Monthly run cost is tooling plus data plus sending infrastructure plus human time: call it 1500 to 4000 a month for a small operation, and check current pricing for whatever tools you actually use. Meetings booked per month, 8 to 25. Show rate 70 to 85 percent. Qualified share of shows 50 to 75 percent. Qualified meeting to closed deal, 10 to 25 percent. Average deal value: use your own, because this is the input that decides everything.
| Step (assumptions, swap in your own) | Conservative | Optimistic |
|---|---|---|
| Monthly run cost | 4000 | 1500 |
| Meetings booked | 8 | 25 |
| Cost per meeting booked | 500 | 60 |
| Show rate | 70% | 85% |
| Qualified share of shows | 50% | 75% |
| Qualified meetings per month | 2.8 | ~15.9 |
| Cost per qualified meeting | ~1430 | ~94 |
| Close rate on qualified | 10% | 25% |
| Cost per closed deal | ~14300 | ~380 |
Put your own average deal value next to that last row and you have your answer. The point of the model is not the numbers, it is the shape: cost per closed deal is far more sensitive to show rate and qualification than to how much you spend on tools. Two teams with identical budgets can sit at opposite ends of that table.
What to change when a number is low
Each weak number maps to a specific lever. Work down the list in order, because fixing a late-stage problem while an early-stage one is broken wastes the effort.
| Symptom | Most likely cause | What to change first |
|---|---|---|
| High bounce rate | Stale or unverified list | Re-verify before the next send |
| Replies fell with no other change | Deliverability | Check placement, reputation, sending volume |
| Low positive reply rate | Offer or targeting | Tighten the segment, then rewrite the offer |
| Good replies, few meetings | Call to action or reply handling | Simplify the ask, answer replies faster |
| Meetings booked, poor show rate | Qualification or scheduling | Book closer in, confirm, qualify harder |
| Meetings show, no pipeline | Fit | Revisit the ICP, not the copy |
That table is most of the diagnostic work. After 2000+ meetings booked for B2B clients, almost every campaign I have had to rescue was failing at one identifiable step, and the team was working on a different one. Find the broken step before you change anything else.
Key takeaways
- Booked meetings that become pipeline is the north star. Any metric that cannot be tied to it stays off the weekly view.
- Positive reply rate is the best leading indicator, but only if the definition is written down and never quietly loosened.
- Count qualified meetings separately, or the system will produce meetings that nobody wanted.
- Split every metric by segment. The pocket worth scaling is invisible in a blended average.
- Keep deliverability numbers on the scoreboard. When replies fall with no other change, that is usually why.
- Set a minimum sample before you read a test, and change one thing at a time.
FAQ
Is open rate really useless now?
For cold email it is close to it. Mail providers pre-fetch images, which fires the tracking pixel with no human involved, so the number is inflated by an unknown and unstable amount. Do not make decisions on it, and do not report it as interest.
What positive reply rate is good?
It varies enough by market, offer and segment that a universal benchmark is not useful. Compare against your own baseline and your own trend, split by segment. A number that is improving on a consistent definition tells you more than any published figure.
Should I track cost per meeting?
Yes, once volume is steady enough that the number is not swinging month to month. Early on, positive reply rate tells you more, because cost per meeting on a handful of meetings is mostly noise. Track cost per qualified meeting as well, since it is the one that maps to revenue.
How do I attribute revenue to a channel?
Tag the first-touch channel on every meeting and follow it through the CRM. It is imperfect, especially in multichannel sequences, but rough attribution everyone understands beats a complex model nobody trusts, and it beats no attribution comfortably.
What if my sample is too small to trust?
Wait for more volume before reacting. Set the minimum before the test starts so the decision is not made under pressure later. Small samples swing wildly, and most campaigns killed for underperformance were killed on a handful of replies.