When a show buys growth it is usually buying one of three things, and they are not interchangeable.
Downloads are the cheapest and the most commonly sold. A download is a file being fetched, which means it is the easiest of the three to produce and the easiest to produce badly. Downloads raise the number on a sponsorship rate card. On their own they do very little for chart position, because a download with no listen behind it is a pattern platforms recognise.
Listens are somebody playing the episode. They are harder to produce, they cost more, and they are what actually moves a show, because they carry completion data with them. A campaign that delivers listens is judged on whether people stayed, which means the traffic has to be from people with some reason to care about the subject. This is the number to buy if the goal is ranking.
Followers are the compounding one. A follower is next week's download arriving without being asked for. They are the slowest to accumulate and the most valuable per unit, because they change the show's baseline rather than its week. A campaign that produces follows has moved something permanent.
The mistake worth naming: buying downloads to look bigger for sponsors. It works on the rate card and it fails on the campaign, because the sponsor measures response against the download number they paid for, and a download that nobody listened to cannot respond. The result is a flight that underperforms and a sponsor who does not return, which costs more than the downloads were worth.
What we would suggest for most shows: listens as the main line, because they are what the platform reads and what a real audience produces. Downloads as a by-product rather than a target. And follows tracked as the measure of whether any of it stuck, because a month after a campaign ends, follows are the only part still working.
If a show is being prepared for a specific sponsorship conversation with a download-based rate card, that is a different problem with a different answer, and it is worth saying so out loud rather than optimising for both at once.