Most B2B Companies Should Not Start a Podcast

Look at the distribution rather than the average. An episode needs 27 downloads in its first week to sit in the top half of all podcasts, 409 to reach the top 10%, and 4,579 to reach the top 1% (Buzzsprout, 2026).

A B2B show in a narrow category will land somewhere in the low hundreds and stay there. That is not failure. It is what a show aimed at 4,000 qualified people in the world looks like.

The metrics that mean something for a B2B show are different.

Who accepted. Twelve guests who each run a department at a target account is a better quarter than 5,000 downloads from an unqualified audience.

Whether guests share it. A guest sending the episode to their own network is the entire distribution model for most B2B shows.

Clip volume shipped. Episodes recorded per quarter multiplied by usable clips per episode. This is the number that connects the podcast to the rest of your marketing.

Deals where the podcast appears in the story. Not attributed, mentioned. Ask on discovery calls and count it.

Only 27% of B2B marketers used podcasts in the last 12 months, down from 30% the year before, while 76% used video and 58% rated video their best-performing content type (Content Marketing Institute, B2B Content Marketing Benchmarks, 2025, 980 respondents). The teams quitting podcasts are mostly the ones who measured them by downloads.

It buys an hour with someone who would not take a sales meeting

A cold sales email to a VP at a target account gets ignored. An invitation to be interviewed for 45 minutes about their area of expertise gets accepted at a rate that surprises people the first time they try it.

You are not selling on that call. You are spending an hour with a senior person who is talking freely about their problems, on a recording, with their guard down. Six months later that relationship exists and the cold email does not.

Treat the guest list as an account list. Book the people you want to know, not the people with the biggest followings.

It produces raw material at a volume nothing else matches

One 60 minute conversation with a good guest contains more usable statements than a week of writing. It is the cheapest way to generate short-form at a cadence you can hold.

The catch is that a flowing conversation does not cut into clips by itself. You have to record it in a way that produces standalone answers, which is a specific set of interviewing habits covered in our guide to turning one recording into a month of short-form. Skip that and you get an hour of good talk and three usable clips.

People imagine a studio. What you actually need is smaller and less negotiable than they expect.

Two cameras. Any two cameras that shoot the same frame rate and can be color-matched. They do not have to match models.

A wired microphone per person. Not a shared shotgun mic, not laptop audio, not a Bluetooth lavalier.

Controlled light on faces. Two soft sources and no window behind the subject.

A room with soft surfaces. Carpet, curtains, a sofa, bookshelves. An empty glass-walled meeting room is the worst room in most offices.

A backup recorder running the whole time. Separate device, separate card.

What you do not need: a set build, a logo wall, a switcher, moving camera work, or a control room. None of it changes whether the episode gets watched.

Remote is the right answer more often than studio people admit. It is how you get a guest in another country, and a guest you can actually book beats a guest you filmed beautifully. The honest quality gap is not resolution. It is that remote gives you no cutaway, no shared eyeline, and audio that depends on a room you have never seen.

Record remote for reach. Record in person for the episodes you intend to cut hard.

This is the single change that separates an edited show from a recorded meeting.

With one camera, every cut is visible. Remove a rambling 30 seconds and the subject's head jumps. So you either leave the ramble in, which costs you the viewer, or you cover the cut with a graphic, which looks like a patch because it is one.

With two angles, cuts disappear. You take out the ramble, switch to the other camera on the join, and nobody can tell. That means you can shorten a 62 minute recording to 41 minutes of the parts that work.

It also doubles clip yield, because a clip cut from a single static angle looks like a Zoom recording, while the same words with one angle change read as produced. The long-form episode and the clips come out of the same two files.

Put the B camera at a different height and a tighter focal length, not just a few degrees off the A camera. Two similar angles give you nothing.

This is counterintuitive and it holds. 92% of weekly podcast consumers say they "listen" to podcasts rather than watch them, and only 8% say they only watch (Cumulus Media and Signal Hill Insights, Podcast Download Fall 2025, 603 weekly consumers).

Video expanded the format rather than replacing the audio. 58% of Americans aged 12 and over consumed a podcast in the last month, and 57% have both listened to and watched one (Edison Research, The Infinite Dial, 2026). Most of that watching happens in a tab that is not in front of the viewer.

Practically: a viewer will forgive soft focus and mixed white balance. They will not forgive room echo, a mic that clips, or two speakers at different volumes. If you have one afternoon to improve the show, spend it on microphones and the room, not the camera.

Weekly is the default recommendation and it is the reason most B2B shows die around episode six. The failure is never the recording. It is the editing backlog and the guest booking pipeline running dry at the same time.

Fortnightly gives you 26 episodes a year, a booking runway, and an edit window that absorbs one bad week. Batch it further: record four episodes in two days, once a quarter. Publish on a schedule the audience can predict while the production happens in bursts.

Commit to 12 episodes before you launch, and have 6 recorded before the first one goes out. Roughly a quarter of companies currently produce a podcast, with another 10% planning to launch (Wistia, State of Video Report, 2026). Most of the ones launching this year will not reach episode 12.

How long should episodes be? Whatever the conversation is worth, usually 35 to 55 minutes after editing. Do not pad to an hour. A tight 38 minute episode gets finished; a loose 70 minute one gets abandoned at minute nine.

Do we need a host who is good on camera? You need a host who is good at listening. The most common failure is a host who talks 40% of the time because they are nervous about silence.

Can we use the podcast as our whole video strategy? No. It produces conversation and clips. It does not produce a product demo or an explainer, and those do a job the podcast cannot.

Should the episodes go on YouTube? Yes, and treat it as a second show rather than a dumping ground. Chapters, a real thumbnail, and a title written for someone who has never heard of the guest.

What do we do with clips from guest episodes? Clear it in the release, then post them. Guest clips outperform host clips because the guest shares them, and they slot straight into a paid hook test if the statement is strong enough.

How do we know when to stop? Set the review at episode 12, not episode 3. Judge it on guests booked, clips shipped, and conversations it started. If none of those moved, stop, and do it without apology. More detail on how we scope this sits in our FAQ.