How to Audit a Podcast Lead Generation Agency

Most people audit a podcast agency by reading the monthly report, which is the one document the agency built specifically to survive being read. We run invite led outbound across 50 plus B2B campaigns, and every engagement that went bad was visible in the input data 60 days before it showed up in the summary slide. Below, the 6 layers worth auditing, the 9 numbers to ask for in writing, and what each answer tells you about whether to fix the engagement or end it.

What Does It Mean to Audit a Podcast Lead Generation Agency?

There is a difference between reviewing an agency and auditing one.

A review reads what you were given. An audit asks for what you were not given, and the gap between those 2 lists is usually the whole finding. This is also different from vetting a vendor before you sign, which is a separate exercise with separate questions, and from spotting the warning signs in a pitch. An audit runs on an engagement you are already paying for.

Agency Audit

A structured review of the inputs, outputs and contract of a live engagement, run by the buyer, using raw data rather than the agency's own summary.

Attributable Recording

A completed interview with a decision maker who matches your target profile and could sign your contract. It is the only output in a podcast acquisition engagement that connects directly to revenue.

The reason this matters is timing. Podcast led outbound has a real ramp, covered in how long this channel takes to produce revenue, so judging it on closed deals at day 30 is unfair and judging it on nothing at all until day 180 is negligent. The audit is how you get a verdict in between.

Why Is the Monthly Report Not an Audit?

Because the party being measured picked the metrics.

A report can be completely accurate and still be useless. Downloads went up, 4 episodes shipped, audience grew 10 percent. All true, none of it answers whether a person who could hire you sat in that chair. Content Allies puts average completion on B2B episodes under 30 minutes at 38 percent, which is a fine number to know and a terrible number to run a business on. The audience was never the point. Measuring a B2B show without downloads is the alternative frame.

The other problem is scope. KKBC's B2B lead generation audit checklist organizes the work into 5 layers covering strategy, process, technology and data, performance and commercials, and operational fit. It is a solid frame, and the reason it exists is that almost nobody audits past the performance layer. Strategy and data hygiene are where the problems start, and neither of them ever appears on a monthly slide.

What Are the 6 Layers to Audit?

Six layers carry a podcast acquisition engagement. Audit them in order, because a failure upstream makes everything downstream unreadable.

Notice that only layer 4 shows up on a standard report, and it shows up as booked rather than held. That single substitution hides more underperformance than any other reporting choice in this category.

How Do You Audit the List and the Invite Engine?

Start with the export, because the list is the highest leverage thing in the engagement and the easiest to check.

Sort everyone invited by job title and company size, then count how many could sign your contract without permission. If that count is under half, you do not have a copy problem or a deliverability problem, you have a targeting problem, and the fix lives in how the target profile is defined and in whether anything gates the list before invites go out. Agencies that bill per booking have a quiet reason to let the list drift wide, which is covered in how the pricing models compare.

Then check the plumbing. Ask which domains the invites send from, whether they are separate from your main domain, and what the bounce rate has been. Domains and warmup are not a technicality, they decide whether anything arrives at all, and a bounce rate above 3 percent usually means the list was never verified. If the agency sends from your primary domain, that is worth raising today, because the whole point of secondary domains is keeping your real mail safe. Ask about spam placement too, since open rate looks identical whether a message landed in the inbox or the promotions tab.

Last, read the actual copy. Not the version in the deck, the version sending this week. Then read 10 real reply threads end to end. You are looking for whether the reply handling answers the question a guest asked or just pastes a booking link, because that single behavior moves booked recordings more than subject lines do. Reply rate benchmarks give you a yardstick for what you are seeing.

How Do You Audit the Recordings and the Handoff?

Two numbers, and the second one is the one nobody volunteers.

The first is show rate. Booked and held are different events, and an engagement reporting only the first is reporting the easier half. Focus Digital puts the blended B2B no show rate near 25 percent and cold outbound booked by an SDR at 35 to 45 percent. A podcast invite should beat both of those, because a guest who agreed to be interviewed is in a different posture than someone who agreed to a sales pitch. If your show rate sits in SDR territory, something in the booking is being oversold, and the fixes for guest no shows are mostly confirmation mechanics rather than copy.

The second is who attended. Pull the list of held recordings and read the titles. This is the number that separates a working engagement from a busy one, and it is why the definition of a recorded conversation has to be written down before anyone signs anything. A marketing coordinator finished the interview and the episode sounds fine. Nothing happened.

Then audit the handoff, which is the layer almost every production shop leaves undefined. What happens in the 7 days after a recording ends? If the answer is "we send the edit", the acquisition part of the engagement stops at the moment it was supposed to start. The sequence after a recording and how guests become clients cover what should be there instead, and the conversion rate from guest to client is the number that tells you whether it is working.

What 9 Numbers Should You Ask For in Writing?

Send this list as a single email and ask for a raw export rather than a slide. How fast it comes back is itself a finding.

1. Invites sent in the period, by week.
2. Bounce rate, with the verification step named.
3. Reply rate against invites delivered, not invites sent.
4. Positive reply rate, with the definition of positive attached. The definition moves the number more than performance does.
5. Recordings booked.
6. Recordings held, and the gap between this and the line above.
7. Show rate, stated as a percentage so nobody has to do the division for you.
8. Titles, company sizes and industries of everyone who recorded.
9. Recorded guests who later took a sales conversation, and what happened on it.

An agency running a real engine returns all 9 in a day, because the system already tracks them. An agency that needs a week is assembling them for the first time, which tells you they have not been watched. An agency that argues with the list has answered the question completely.

Run the same arithmetic on top: total spend divided by held recordings with real buyers. That is cost per recorded conversation, and it is the only figure that makes 2 differently priced engagements comparable. Take it further with podcast ROI and how deal size changes the math.

Do You Fix the Engagement or End It?

Sort the finding into 1 of 3 buckets before deciding anything.

• A list problem. Invites are landing, replies are coming, and the wrong people are saying yes. This is the most fixable version and usually costs 2 to 3 weeks. Rework the target definition, re-gate the list, keep going.
• An engine problem. Invites are not arriving, or replies arrive and die. Deliverability and reply handling are both fixable by a competent shop, and a reply drought has known causes. Give it 1 cycle with a written change, not a verbal one.
• An orientation problem. The agency cannot produce raw numbers, reports downloads when you asked about buyers, or treats guest sourcing as outside scope. This one does not get fixed by a meeting. It is not underperformance, it is a different business being delivered than the one you bought.

Timing matters as much as the verdict. Audit at day 30 on inputs only, day 60 on whether the right people are recording, and day 90 on outcome. What the first 30 days should actually look like sets the early bar, and the common failure modes are worth reading before you draw a conclusion, because several of them look identical from the outside and have opposite fixes.

If you land on ending it, do the paperwork before the conversation. Who owns the sending domains, the recordings, the guest list and the published episodes decides how much of the last 90 days you keep. How to end an outbound engagement cleanly covers the sequence, and the asset question is the one people forget until it is too late.

The Practitioner Read on Auditing a Podcast Agency

The hardest part of this is not the arithmetic. It is that the audit usually finds a problem you helped create.

Most engagements that go sideways were sold with no number in the paper, bought with no definition of who counts as a buyer, and reported on with metrics nobody agreed to in advance. That is a 2 sided failure, and the fix on your side is cheap. Write the definition down, name the number, and ask for the raw export every month instead of once a quarter when you are already unhappy.

On our side the unit is the recorded conversation, which is an ICP decision maker who showed up and completed the interview on your show. The commitment is 30 of them with your ideal buyers in 90 days, or your money back. Editing and publishing are included, invites go out by email only, and you own every recording. Rates get scoped on a call because the number moves with the market being invited.

Run the audit at day 30 whether or not anything feels wrong. The version you run when you are already angry is the expensive one.