Midroll ads in audio articles: where to cut, how many
The best-paid audio ad format, applied to a news article: where the break has to fall, how many to run, and the length below which it is not worth it.
A midroll is an ad played partway through a listen, sitting between the pre-roll that opens it and the post-roll that closes it. In podcasting it is the most expensive slot on the market, for a straightforward reason: a listener who has reached the middle of an episode has demonstrated attention, and demonstrated attention is exactly what the advertiser is buying. Move the format onto a narrated news article and it keeps that advantage while inheriting a constraint podcasting does not face to the same degree. An article is constructed prose. You do not cut an argument in half to fit a spot.
Why midroll sells for more than pre-roll
Three gaps explain the price difference, and they compound.
The first is proof of attention. A pre-roll plays to someone who just pressed play and has proved nothing yet. A midroll plays to someone who stayed. The second is skip rate: on a player embedded in an article, the listener who started playback is usually occupied elsewhere (commuting, doing something with their hands) and is not holding the screen ready to tap past the message. The third is environment. A midroll inside an identified news article is a context few programmatic inventories can offer, and context gets paid for.
In practice, across markets where digital audio is mature, the rate gap between pre-roll and midroll runs from roughly a third to double, depending on inventory quality and the type of advertiser. Those figures move, and they move most according to whether you sell direct or through a third-party network, but the direction of the gap is constant.
Where the break goes
This is the decision everything else follows from, and there is only one right answer: on a boundary in the text, never on a clock.
Fixed-duration insertion, "a spot at three minutes", is what most tooling does, and it is what produces a cut in the middle of a sentence, inside a quotation, or worse, between a figure and its unit. The listener does not think "badly placed ad", they think the player is broken. The credibility cost dwarfs the revenue from the spot.
The break therefore has to land on a natural boundary: the end of a paragraph, ideally the end of a section, where the writing itself takes a breath. That is what becomes possible when the audio is produced from the article's structured text rather than from a flat file, the boundaries are known before the voice is even generated, so the slot can be positioned on them instead of being retrofitted onto them.
Two further rules, learned in radio long before digital audio existed. Never cut immediately after a hook, at the moment the listener is waiting for the payoff; you have put them in suspense and charged them for it. And never cut inside the final thirty seconds, because that is not a midroll, it is a mispositioned post-roll that makes people abandon just short of the end.
How many, and above what length
Ad density is the setting publishers most often get wrong in their own disfavour.
Below six or seven minutes of listening, a midroll has no place: the pre-roll is enough and the extra interruption is paid for in drop-off. Between seven and fifteen minutes, a single midroll placed after the first substantial section is the setting that holds. Above fifteen minutes (a long read, an investigation, a feature) two breaks are defensible, with at least five minutes between them.
The temptation to add more is understandable and turns against you quickly. The number to watch is not impressions but listen-through rate: if adding a second break costs you a tenth of the listeners who would have reached the end, the extra inventory has been bought with audience, which is a poor trade for a title that lives on audience. Listening analytics are what settle that question, rather than instinct.
What the listener needs to be told
An audio ad has to be identifiable as one. That is an editorial-standards requirement in news, and it becomes a regulatory one wherever the format can be confused with editorial content, a spot read in the same tone, in a voice close to the article's, inside a continuous playback, is precisely the confusing case.
The treatment that works has three parts: a short sonic transition marking the message in and out, a voice or a treatment distinct from the editorial voice, and an explicit mention when the content is sponsored. None of this is a commercial handicap. A listener who knows where the advertising starts trusts what surrounds it more, and that trust is what gives the slot its value.
Midroll inside a wider strategy
Midroll is a lever, not a strategy. It comes after the pre-roll, which is simpler to sell and has no placement constraint, and it combines with the other revenue sources audio opens up: the registration wall that converts a listen into an account, and podcast distribution that reaches a different audience entirely. The article on the three monetisation levers puts the whole picture together, and the midroll glossary entry gives the short definition.
A sensible order: get listening established and measured first, sell pre-roll to confirm the inventory finds buyers, and only introduce midroll on long formats once you know how many people actually reach the end. Midroll inventory sold against an audience that drops out at minute four is inventory you will not sell twice.
At WeDispatch the ad slots are positioned on the article's structure rather than on a counter, which makes a cut mid-sentence structurally impossible; the audio advertising page covers the setup. The quickest way to judge the result is to hear it on one of your own long reads.
Give your articles a voice with WeDispatch
This blog is itself voiced by WeDispatch. Curious how it sounds on your content?
Book a demo