30 Recorded Conversations in 90 Days: The Math

Most agencies set their guarantee at the number their funnel model produces on a good day. We run outbound for 50 plus B2B companies and have shipped over 8 million cold emails this year, and the 30 recorded conversations we back in 90 days is roughly a tenth of what that same send volume models out to. Below, where the 30 actually comes from, why the window is 90 days, what one recorded conversation costs against a booked meeting, and the 4 things that break the number.

What Does 30 Recorded Conversations in 90 Days Mean?

The wording matters more than it looks. A guarantee is only as strong as the unit it is denominated in, and most outbound guarantees are written in units the buyer cannot audit. Leads, opportunities, and qualified meetings all mean whatever the contract says they mean, and the definition usually lives in a paragraph nobody reads until month 3.

A recorded conversation is different because it either exists as a video file or it does not. You can count them from your own calendar. You do not need our dashboard to tell you whether it happened.

Recorded Conversation

A decision maker inside the client's defined ICP who shows up and completes a recorded interview on the client's own show, recorded on Zoom or Google Meet. The unit is the completed recording. A no show does not count. A junior employee sent in place of the buyer does not count. The later sales conversation is a separate event and is never counted toward the 30.

Guaranteed Unit vs Reported Metric

A reported metric is anything the vendor measures and shows you, such as sends, opens, replies, or positive replies. A guaranteed unit is the one thing the vendor is financially exposed on. The distinction decides who carries the risk when a quarter goes sideways. Vendors who report 6 metrics and guarantee none of them have transferred all of it to you.

There is a second reason to denominate the promise in recordings rather than meetings. A booked meeting is an intention. Show rates on cold booked appointments commonly sit in the 70 to 80 percent range, so 30 booked meetings is often 21 to 24 held ones, and the 6 to 9 that evaporated still counted against the contract. Counting only completed recordings removes that argument entirely. We go deeper on the unit problem in what a qualified meeting actually is and what a sales qualified meeting means.

If the invitation model itself is new to you, start with what reverse outbound is and B2B podcast lead generation. The short version: instead of pitching a cold buyer, you invite them onto your own show as the expert, and the recorded conversation does the trust work a first sales conversation normally has to do cold.

Where Does the Number 30 Come From?

Here is the part most vendors keep private. The engagement sends 10,000 personalized email invites a month, so 30,000 across a 90 day window. Run that volume through our own book averages and the arithmetic is not close to 30.

Our reply rate across 50 plus campaigns sits at 4.6 percent against a templated market median of 3.43 percent, per the Instantly 2026 benchmark report. Roughly 40 percent of those replies are positive, and roughly 57 percent of positives complete a recording. That chain, at 30,000 invites, models to 315 recorded conversations in the window.

We back 30. Here is the sensitivity table that explains why.

Read the fourth row again. A campaign running at 1 percent reply, with only a quarter of those replies positive and only 40 percent of the positives making it to a finished recording, is a bad campaign by every internal standard we hold. Our own kill rule fires below 1 percent reply or below a 30 percent positive share after 5,000 sends, and when it fires we rework the list before we touch a word of copy. That bad quarter still produces exactly 30 recordings.

That is the whole design. The guarantee is not set at what a strong quarter delivers. It is set at what a failing campaign delivers before we intervene, which means the number holds even when the first list is wrong and we have to rebuild it mid engagement. Benchmarks for the underlying rates are in cold email reply rate benchmarks, and the stage by stage version of this funnel lives in how many invites it takes to book one recording.

Why Cap It at 30 When the Model Says 315?

Because the top of the funnel is not the binding constraint. The host's calendar is.

30 recordings over 90 days is 10 a month, or about 2 to 3 a week. Each guest costs the host a 15 minute alignment call plus a 45 minute recording, so roughly one hour per conversation. Across the full 30 that is about 30 hours in a quarter, which is under 20 minutes a day. A founder who also runs the business can hold that. The same founder cannot hold 105 recordings a month, and pretending otherwise would make the guarantee a trap rather than a floor.

This is why capacity, not volume, sets the ceiling on any host led channel. Sending more invites into a calendar that is already full does not produce more conversations, it produces more reschedules and more people who agreed to something that never happened. Once the calendar is the constraint, the operating question changes from how do we get more replies to how do we get better guests into the same 10 slots. That is a targeting problem, and it starts at the list. How to define an ICP for cold email is where that work happens.

There is a quality argument here too. A channel that fills a calendar with 10 senior buyers a month beats one that fills it with 40 marginal ones, because the later sales conversation only exists if the guest was worth talking to in the first place. We gate guests on firmographics before invites go out for exactly this reason, which is also what keeps the guaranteed unit honest. Related reading: how to get decision makers to reply and how to get high profile podcast guests.

Why Is the Window 90 Days and Not 30?

Three reasons, and the first one is purely mechanical.

1. Infrastructure eats the first month. Sending domains need 3 to 4 weeks of warmup before the first real invite goes out. A 30 day guarantee would spend most of its clock on DNS records and warmup traffic, which means it would be a guarantee about setup, not about conversations. The full ramp is in how to warm up a new email domain and podcast invite deliverability.
2. One send is not a dataset. 90 days leaves roughly 60 days of live sending, which is enough to run a full pass at the list, read the reply data, and rebuild the targeting if the first cut was wrong. Our own rule is to judge nothing before 5,000 sends, and a 30 day window rarely clears that bar with warmup taken out of it.
3. It matches how buyers actually move. The median B2B sales cycle now sits around 84 days with a mean closer to 134, per Ziellab's 2026 cycle length analysis, and Optifai's benchmark set across 939 companies puts high ACV deals well past that. A 90 day window is roughly one cycle. It is long enough to produce conversations that turn into pipeline inside the same quarter, and short enough that nobody is waiting a year to find out whether the channel works.

The 90 day window has a companion commitment that covers the early part of the clock, because the honest weakness of any quarterly guarantee is that a client can sit for 6 weeks wondering whether anything is happening. First recording within 14 days of the first invite sent, or the next month is free. That collapses the feedback loop from a quarter to 2 weeks, and it puts the pressure back on us during the window where a client has the least information.

The third piece is what happens on a miss. You get your money back, and you keep every recording produced up to that point plus the warmed sending infrastructure. A refund that also strips the assets is a refund that punishes the client for our failure. The recordings are yours either way, which is the point of running the show on your own channel rather than someone else's. More on that asset side in B2B podcast ROI explained and how to repurpose podcast episodes.

What Does One Recorded Conversation Actually Cost?

The engagement is $8,000 flat, one offer, no tiers, with 0 percent financing available. Across the 30 guaranteed conversations that is about $267 each. Every recording past 30 pushes it down, since the price does not scale with volume. Here is how that lands next to the alternatives.

The retainer and per appointment ranges come from published 2026 pricing guides, including Leadium's appointment setting pricing guide and DemandNexus cost benchmarks, which put qualified meetings in a $300 to $800 band for mid market and C suite meetings materially higher. The SDR output row uses median productivity benchmarks from TamToTarget's 2026 SDR meeting benchmarks and Growthspree's SDR quota and productivity data, where the median outbound rep books 8 to 15 meetings a month and quota attainment sits under 70 percent.

Two honest caveats on that table. First, the units are not identical. A recorded conversation is 45 minutes of a senior buyer's attention with no pitch in it, and a booked sales meeting is 30 minutes where they arrived expecting one. Those are different goods, and which one is worth more depends entirely on how long your sales cycle is and how much trust it needs. Second, an in house rep is an asset you own and a vendor is not, which is a real argument on a multi year horizon even when the quarterly math looks worse. We laid both sides out in podcast invites versus appointment setters, podcast invites versus an SDR agency, and done for you outbound versus hiring an SDR.

If you want to run your own version of this arithmetic against your deal size rather than ours, cold email ROI by ACV and how to lower cost per booked meeting both work the math from the revenue side backward.

What Breaks the Number, and Who Pays for It?

4 things move this math, and only 3 of them sit on our side of the line. Naming the fourth is the part most vendors skip.

The last row is the honest condition on the whole promise. If a host cancels 8 recordings in a quarter, the number does not get made, and no amount of send volume fixes it because the constraint was never the invites. That is the trade in a host led channel. You get conversations at a fraction of the cost of buying meetings, and in exchange the channel needs an hour a week of the most expensive person's calendar.

The first 3 rows are ours, and they are the ones we instrument. Placement gets a weekly seed test per domain rather than a campaign average, because a blended number hides the one failing domain until it takes its neighbors down. Positive share gets watched separately from reply rate, since a list that generates replies but not interest is a targeting failure wearing a copy failure's clothes. And no shows get a structural answer rather than a hopeful one: reducing sales meeting no show rate covers the mechanics, and the short alignment call before every recording is the single highest leverage piece of it.

Worth saying plainly: we have run this badly before. The campaigns that miss are almost always list problems that looked like copy problems for 3 weeks first. Diagnosing that faster is most of the operating work. How to fill your calendar with sales calls and how many qualified meetings per month is realistic cover the pattern from the pipeline side.

What the Guarantee Does Not Promise

It does not promise revenue. It never has, and any version of it that did would be dishonest.

We control the invites, the sending infrastructure, the targeting, the booking, and the recording. We do not control your price, your offer, how you run the sales conversation after the episode, or whether the buyer's budget cycle happens to line up with your quarter. Guaranteeing a revenue number across all of that would mean underwriting variables we cannot see, which in practice means writing conditions into the contract until the guarantee quietly stops meaning anything.

So the promise stops where our control stops. 30 recorded conversations with your ideal buyers, or your money back. What those conversations turn into is a function of your close rate, and that is the number worth improving on your side while we work on the top of the funnel.

The reason the conversations are worth having at all is a shift in how buying happens. Gartner research puts the share of the B2B buying journey spent with any vendor at roughly 17 percent, and a 2026 Gartner survey found 69 percent of buyers now turn to sales reps specifically to validate what AI told them. Buyers are doing more of the work alone and then looking for a human to check it against. A recorded conversation is a low friction way to become that human before anyone is selling anything, which is the argument in invite versus pitch in B2B outbound and cold email versus podcast invites.

The conversion side is a separate skill, and it is the one clients under invest in. 30 conversations with senior buyers is a lot of pipeline if you have a real follow up motion behind it, and it is 30 nice chats if you do not. How to turn podcast guests into clients and how to get clients from a podcast are the 2 posts to read before the first recording, not after the tenth.

How to Pressure Test Any Outbound Guarantee

Whether or not you ever work with us, the questions below are the ones worth asking any vendor who puts a number in a contract. They take 5 minutes and they surface almost everything.

• What exactly is the unit, and can I count it without your dashboard? If the answer requires their reporting to verify, the unit is theirs, not yours.
• Is the number a floor or a target? Ask what happens at their worst historical performance. If the guarantee only holds in an average quarter, it is a forecast wearing a guarantee's clothes.
• What is the constraint that caps the number? Every real channel has one. If they cannot name it, they have not modeled the channel, they have modeled the invoice.
• What do I keep if you miss? A refund that also removes the assets means a failed quarter costs you a quarter of calendar time with nothing to show.
• Which inputs are mine? A vendor who claims every input is theirs is either taking risk they cannot price or planning to argue about it later.
• What is the earliest signal that it is working? A quarterly promise with no 2 week checkpoint means finding out in month 3 that month 1 was broken.

Run those against our number and it holds up on 5 of 6, with the sixth being the honest one: the host's calendar is your input, and if it does not stay open the math does not work. That is the disclosure, not the fine print.

The broader direction here is worth naming. Reply rates on generic outbound have fallen from around 5 percent in 2024 to a 3.43 percent median in 2026 as inboxes saturate, and the gap between average and top performers keeps widening. Volume as a strategy is getting more expensive every quarter, while the cost of one genuinely good conversation with a senior buyer stays roughly flat. Guarantees written in volume units will keep getting harder to hold. Guarantees written in conversations will not, because the constraint on those was never the sending. It was always the calendar, and the calendar has not changed. Where this sits in the wider channel picture is covered in the state of AI outbound in 2026 and whether you need an audience for podcast lead generation.