More volume vs better conversion
identical arithmetic, very different ceilings
By Jānis Plūme, Founder, Outbound Pros · 10 min read · 2026-08-06
Quick answer
Algebraically the two levers are identical. Required sends equal positive replies needed divided by your positive reply rate, so doubling the numerator input and halving the divisor produce the same output. Practically they are opposites. Volume is bounded by your market size and your sending infrastructure, arrives immediately, costs roughly linearly and stops when you stop paying. Rate work is bounded by how wrong your current targeting and offer are, arrives after a sample large enough to read, costs mostly fixed effort and carries across every sequence after it. The decision rule we act on is the gate: below 0.5% positive on sends after warm up, fund rate work or kill the sequence. At 1% or above, fund volume. Between them, iterate and do not scale.
This is the most common argument inside a go to market meeting and it is almost always conducted without the arithmetic on the table. Somebody wants more mailboxes and somebody wants better lists, both are describing a real constraint, and the disagreement is unresolvable until the fraction is written down with what it divides by.
So write it down. Positive reply rate on sends is positive replies divided by emails sent. It is not positive replies divided by replies received, which is a ratio that runs an order of magnitude or two higher and belongs in a completely different sentence. Mixing those two is how a planning meeting approves a budget that is wrong by a factor of a hundred, and it is the specific error this property exists to correct, including in our own earlier marketing.
Why the two levers look equivalent and are not
Take a plan needing 200 positive replies. At 0.5% positive on sends that is 40,000 emails. Halve the rate to 0.25% and you need 80,000. Double it to 1% and you need 20,000. The output moves proportionally either way, which is why the levers look interchangeable on a slide.
They come apart on three things the slide does not show. The first is that the levers are not independent. Volume degrades rate, reliably and in two ways at once. You go deeper into the market and reach companies that fit the criteria less well, and you add sending infrastructure faster than reputation accumulates behind it. So the second 40,000 emails almost never convert at the rate of the first 40,000, which means doubling volume produces less than double the positives and the model quietly overstates the plan.
The second is that they hit different ceilings. Volume runs into market size and physical sending capacity, both of which are hard walls the arithmetic can check in advance. Rate runs into a softer wall: there is a maximum plausible rate for a channel, and planning above it is planning on an outlier. As an anchor, the highest weekly reply rate we hold on our largest account is 0.84%, which is 377 replies on 44,649 emails and counts every reply including rejections and out of office notices. Positive replies are a subset of that. Any plan whose rate assumption implies more positives than that total would allow has an input error rather than an ambitious target.
The third is what each lever leaves behind. Volume stops producing the week it stops being funded. A rate improvement that came from a better segment definition or a sharper offer applies to every sequence you run afterwards, including the ones you have not written yet. That is not a reason to always choose it, and it is a reason to stop describing them as the same decision made twice.
The comparison that matters
| Dimension | More volume | Better conversion |
|---|---|---|
| What you are changing | How many companies you reach and how often | What share of the ones you reach respond well |
| Where the ceiling is | Market size and sending capacity. Both are checkable in advance | A plausible maximum rate for the channel, and how wrong your current targeting is |
| Cost behaviour | Roughly linear. Every increment costs about what the last one did | Mostly fixed effort, then it applies to everything you send afterwards |
| Time to effect | Immediate once infrastructure is warm | One full sequence cycle plus a sample large enough to read |
| Effect on the other lever | Usually negative. Deeper lists and newer infrastructure convert worse | Usually positive. A better segment makes added volume worth more |
| Reversibility | High. Turn it down next month | High, but the learning does not reverse, which is the point |
| What it does to measurement | Improves it. Volume buys the sample size a rate reading needs | Degrades it briefly. Changing several things at once makes the next reading unattributable |
| Characteristic failure | List exhaustion and deliverability collapse, both of which look like a copy problem | A quarter spent iterating on a sequence that a defined gate would have killed in week three |
| The reading that says it is exhausted | Required companies exceed your qualified market, or required sends exceed what your fleet can carry | Three iteration cycles on a list you believe in with the rate still under the kill line |
Where more volume wins
Volume wins when the rate is already good. At 1% positive on sends and above, a sequence has proved it works on a segment that exists, and the fastest route to more pipeline is more of the same. At 2% and above we would put everything available behind it. Spending that period optimising a working sequence is the most common way a genuine winner gets under exploited, because optimisation feels like diligence and it is capacity you did not use.
Volume wins again when you do not yet have a readable sample. At a 0.5% positive rate, single positive replies move the measured number noticeably until the sends run into the thousands, which means a team sending hundreds and debating whether the rate improved is debating noise. Volume is what buys the statistical power that any rate decision requires, and this is the case where the two levers stop competing entirely: you need the volume in order to be allowed an opinion about the rate.
And volume wins when the constraint is a deadline. Rate work has a cycle time that cannot be compressed by wanting it more. If the pipeline is needed inside the window and the ceiling checks say the volume exists, buying it is the honest answer even though it is the less clever one.
Where better conversion wins
Rate work wins whenever a ceiling is close. If the model already requires more companies than exist in your qualified market, or more sends than your infrastructure can carry inside the window, volume is not a lever you have. It is a number you will fail to reach, and the only remaining variable is the rate. This is the case people discover in month four instead of in the twenty minutes it takes to check.
It wins decisively below the kill line. Under 0.5% positive on sends after warm up, adding volume multiplies a sequence that is not working, which spends market and reputation to buy proportionally more of a bad result. Our fleet baseline sits near 0.05% positive on sends as a derived figure, an order of magnitude below the kill threshold, precisely because the portfolio includes everything in warm up and everything about to be stopped. A portfolio mean and a per sequence gate are different instruments and neither excuses scaling something under the gate.
It wins on order of operations too. Prioritisation, in the sense of changing which companies are on the list at all, is a bigger lever than copy and it is frequently mistaken for a rate improvement when it is really a different denominator. If your market is very large, deciding who to send to is worth more than deciding what to say, and both are worth more than sending more of the current thing.
The gate that settles it
We run four thresholds, applied per sequence and per segment, after warm up, over a window agreed in writing before the first send. All four read the same fraction, positive replies divided by emails sent. Below 0.5%, kill the sequence. Between 0.5% and 1%, iterate and do not add volume. At 1% or above, fund more volume. At 2% or above, put everything you have behind it.
Those are decision rules we act on with our own clients, including when killing a sequence costs us revenue, which is the only reason to take them seriously rather than as content. The gate resolves this comparison automatically in most cases, and the hard part was never choosing the threshold. It is committing to the window before you see the numbers, because a threshold negotiated after a bad week is whatever the loudest person in the room wanted and everyone will remember it as a decision.
One reading means fund neither. If you have run three genuine iteration cycles on a list you believe in and the rate is still below the kill line, the problem is not volume and it is not copy. It is that this segment does not want this offer at this price, or that outbound is the wrong instrument for the deal size. Continuing to fund either lever at that point is the most expensive decision available, and it is the one that gets made most often, because both levers feel like action.
Who should pick which
Buy volume if your rate clears 1% on a segment you can describe in a sentence, your market ceiling and sending capacity both have room, and your calendar can hold the meetings the volume will produce. That last condition is not decoration. Where calendar discipline is broken, roughly half of booked meetings die, which means a volume increase converts into half the pipeline you funded it for.
Buy rate work if the rate is under the gate, if a ceiling check is tight or broken, or if your market is large enough that who you contact matters more than how many. Change one thing per cycle, because changing the list and the offer and the copy together produces a reading you cannot attribute and therefore cannot repeat.
Do both, deliberately sequenced, if you are running more than one segment. Fund volume on the segments above the scale gate and run rate work on the ones between the gates, with separate reporting for each. A blended rate across both hides that one segment is carrying the other, and a team reading a blended number will scale the wrong one. This is the single most common reporting mistake we correct.
Test the trade off with your own inputs in the pipeline model. Run it once at your current rate and once at the rate you believe you could reach, then compare both against your market and capacity ceilings. If only the optimistic version fits, you do not have a plan, you have a hope with arithmetic attached. AllboundPros is part of the Outbound Pros group and the group sells managed outbound, which means we are paid more when the answer is volume. The gates above are the ones we apply anyway, and if your reading says fund neither, the useful next step is an audit of the inputs rather than another quarter of sending.
The model runs a conservative and a target rate in parallel and shows which ceiling breaks first. The spread between the two columns is usually larger than the argument being had about them.
Scale or optimise, asked properly
If the two levers are algebraically the same, why not always pick the cheaper one?
Because they are not independent and they do not hit the same wall. Adding volume tends to lower the rate, since you reach less well fitting companies and add infrastructure faster than reputation accumulates, so the second half of a volume increase converts worse than the first. And volume has hard ceilings the arithmetic can check in advance, market size and sending capacity, while rate has a soft one you can only discover by running into it. Cost alone is the wrong comparison.
How many sends before a rate reading is stable enough to act on?
Enough that one positive reply landing or not landing cannot move the measured rate more than you can tolerate. At a 0.5% positive rate that means thousands of sends rather than hundreds. A team comparing 0.4% to 0.6% on a few hundred emails is comparing noise, and the correct response is not a better analysis, it is more volume until the sample is readable or a decision to stop on other grounds.
Our rate dropped after we scaled. Did we break something?
Probably not, and that is the point worth internalising. A rate drop on scaling is the expected shape rather than a fault, because you extended into a less well qualified part of the market and onto newer infrastructure at the same time. What matters is whether the blended rate still clears the gate. If it does, the scale was worth it. If it does not, you have found the volume ceiling for the current segment definition and the next move is rate work, not more volume.
When is the answer to fund neither?
When three genuine iteration cycles on a list you believe in leave the rate below the kill line, when the required company count exceeds your qualified market, or when cost per meeting held cannot be carried by your deal size. Those are three different disqualifications and only the first is about execution. All three are cheaper to discover in twenty minutes of division than in a quarter of spend, and all three are reasons we turn work down.
Last updated: 2026-08-06