What Is a Podcast RSS Feed and Why Does It Matter?
11 March 2026
No Comments
Your RSS feed is what makes your podcast portable, distributable, and genuinely yours. Here’s how it works — and why it matters more than most beginners realise.
Here is a question nobody in the podcasting industry wants to answer honestly: who decides whether your podcast gets heard?
Not you. Not your listeners. Not even the quality of what you make.
An algorithm does.
That word gets thrown around constantly in podcasting circles — usually in a tone somewhere between reverence and helplessness. But in my PhD research at the University of Pretoria, where I am investigating the political economy of indie podcasting in South Africa, the algorithm keeps appearing not as a background detail but as one of the central forces shaping whether podcasting in this country is viable at all. And the more I read the academic literature, and the more I talk to South African indie podcasters, the more convinced I become that most creators are working with a dangerously incomplete picture of what these systems actually do.
This article is about filling in that picture — in plain language, with practical implications you can act on.
Podcast algorithms are automated systems that decide which shows to recommend, rank, and surface to potential listeners. They do not care about quality. They respond to measurable signals: follows, listen-through rates, metadata relevance, and engagement velocity. Understanding those signals — and the power structures behind them — is now an essential part of running a podcast.
Strip away the mystique and an algorithm is just a set of instructions that produces an output from a given input. A recipe is an algorithm. A scoring rubric is an algorithm. The formula your bank uses to calculate your credit risk is an algorithm.
What makes platform algorithms different — and genuinely difficult to navigate — is that modern platforms no longer use simple rule-based instructions. Spotify, Apple, and YouTube run on machine learning systems that identify patterns in listener behaviour and adjust their recommendations accordingly. The system is not following a rulebook a human wrote. It is drawing conclusions from billions of data points, most of which you will never see.
This matters for a specific reason. With a rule-based system, you can learn the rules and play by them. With a machine learning system, you are chasing patterns that shift as the data shifts — and the platform is under no obligation to tell you what changed or why.
That information asymmetry is not an accident. It is a design choice.
When a podcaster says “I want to chart on Spotify,” they usually mean they want more listeners. That is reasonable. But the way those charts actually work is worth understanding, because it changes what you should prioritise.
Apple Podcasts bases its chart rankings heavily on the rate of new subscriber growth over very short windows — 24, 48, and 72-hour periods matter more than long-term listener numbers. A surge in new follows this week carries more weight than the fact that you have been consistently releasing episodes for three years. This is why new show launches that mobilise an existing community can chart quickly, while slower-burning, high-quality shows often remain invisible despite excellent content.
Spotify’s approach layers in more signals: total follows, unique listeners, completion rates (how much of your episode people actually finish), and consistency of returning listeners. Spotify’s “Top Podcasts” ranking combines total followers with recent unique listeners, meaning a show needs to be both historically popular and currently trending. What this tells you is that Spotify rewards shows that retain listeners week after week, not just shows that spike.
One in two podcast listeners discover new shows directly inside podcast apps, and 70% of them use the search bar. Editorial features and chart positions? They account for a fraction of actual discovery. This is a significant finding that most podcasting advice ignores entirely.
And the review myth? Despite being repeated constantly in podcasting communities, there is very little evidence that five-star reviews directly boost your algorithmic ranking. Reviews matter for social proof and may influence human editorial curation — those “New and Noteworthy” placements — but the algorithm responds to behavioural signals, not star ratings. The persistence of the “beg for reviews” advice is itself a symptom of how little platform transparency exists. Creators fill the information vacuum with industry folklore, and the folklore spreads.
There is a concept in media studies called the “infrastructure of discovery” — the systems, interfaces, and pathways that make content findable. Podcasters spend a lot of time thinking about content. They spend far less time thinking about infrastructure. That is a strategic error.
Consider how localization works. Platforms prioritise content that originates from the user’s own country. For South African podcasters, this should theoretically be an advantage — we are local content, and a South African listener’s algorithm should surface us first. In practice, however, the volume of international content (predominantly North American) is so dominant that local signals often get drowned out unless the creator is deliberate about optimisation.
Personalisation creates a different kind of problem. Platforms build listener profiles based on past behaviour and recommend content within those profiles. For a podcaster trying to reach a new audience — someone who has never listened to South African podcasting before, or someone outside your usual niche — the algorithm creates a wall. It does not know to recommend you to people who have not already demonstrated an interest in you. This is the filter bubble problem, and it is particularly acute for educational podcasters and niche creators.
Spotify tracks how much of your episode people actually listen to. If 1,000 people start your episode but 900 quit after two minutes, Spotify interprets this as a signal that your content is not engaging. This single metric has more influence over your discoverability than almost anything else, yet most podcasters do not structure their episodes with this in mind.
In academic literature, there is a term for what podcasters do when they shape their creative decisions around algorithmic requirements: “algorithmic labour.” It refers to the strategic, often invisible effort required to make content legible to platform discovery systems — to produce content that the algorithm can recognise as worth promoting.
This labour is real and it is substantial. Every time you agonise over an episode title because you know it needs a searchable keyword. Every time you cut your intro short because you know listeners bail in the first 60 seconds. Every time you post a short-form clip to Instagram or TikTok — not because that is where your audience lives, but because that is where the discovery engine might find new listeners to send your way. That is algorithmic labour. You are doing it whether or not you have a name for it.
What concerns me, from a research perspective, is what this labour costs. The podcasters who can sustain it — who can afford to hire editors for social media cuts, who can post consistently across multiple platforms, who have the time to study platform changes — are not the typical South African indie podcaster. They are usually those who are already resourced, already established, or already backed by institutional money.
The algorithm, in other words, does not just discover the best content. It discovers the content produced by people with enough capacity to speak its language.
Here is where the Global South context becomes critical, and where my PhD research is most directly relevant.
Most of the advice about gaming algorithms is built for creators in the US, UK, and Australia — markets with higher per-capita ad revenue, larger local audiences, and listeners with reliable high-speed internet. South African (and by extension African) podcasters face a different set of constraints. Data costs remain a barrier for many potential listeners. The local advertising market is smaller. English-language content competes directly with international giants. Multilingual content — which reflects the actual reality of South Africa — is not well understood by algorithms trained predominantly on English-language listening data.
This means that simply importing international optimisation advice is not good enough. We need to understand which algorithmic signals are worth prioritising given our constraints, and which require resources most South African indie podcasters do not have.
There is also a subtler issue. Algorithms are trained on historical data. They reproduce existing patterns. If South African podcasting has historically been underrepresented in global listening data — and it has been — then the algorithm’s default is to keep it underrepresented. It is not malicious. It is mechanical. But the effect is the same.
This is one of the reasons the South African Podcasters Guild matters. Collective visibility — cross-promotion, community building, shared audiences — is a way to generate algorithmic signals that individual creators cannot generate alone.
1. Make your first 60 seconds the best 60 seconds. Spotify’s completion rate signal is one of the most influential factors in its recommendation engine. If listeners are bailing early, your discoverability suffers. Cut any preamble — the long intros, the “welcome back,” the sponsor reads at the top. Earn attention before you ask for it.
2. Optimise your metadata like it is your front door. 97% of podcasts that appear in Apple Podcasts search results include the user’s search term at least once in their metadata. Your show title, episode titles, and descriptions need to contain the words your potential listeners are actually searching for — not the words you think describe your content perfectly. Use the language your audience uses, not the language your discipline uses.
3. Ask for follows, not reviews. Reviews matter for credibility. Follows drive your algorithmic ranking. Every time you ask listeners to “leave a five-star review,” you are directing their energy toward a signal that has less algorithmic impact than a simple follow. Ask for the follow first. Every time.
4. Prioritise consistency over volume. The algorithm rewards predictability. A show that releases every Tuesday builds a listener pattern that platforms can detect and reward. Erratic publishing — even with higher-quality episodes — sends weaker signals. Find a cadence you can sustain for six months without burning out. That is your publishing schedule.
5. Stop chasing every platform. Go deep on two. The “breadcrumb strategy” — posting everywhere to drive traffic back to your main show — sounds logical but requires significant ongoing labour. For most South African indie podcasters, the ROI is not there. Pick the two platforms where your audience already listens (likely Spotify and YouTube, based on current data), and master those. Do less, better.
There is a phrase that keeps appearing in my research notes: “the optimization of culture.” It refers to what happens when creators — not just podcasters, but musicians, writers, filmmakers — begin shaping their creative decisions around what platforms will reward rather than what they actually want to make.
This is not abstract. Every podcaster who has shortened an episode because they worried the algorithm would penalise the length, or given a topic a more sensational title than it deserved, or abandoned an experimental format because it felt too risky, has experienced this pressure. It is real. It is widespread. And it has consequences for the kind of podcasting that gets made.
I am not arguing that you should ignore the algorithm. In the current landscape, that is not a viable strategy unless you have a large, loyal audience that exists independently of platform discovery — and very few South African podcasters are in that position.
What I am arguing is that you should understand the algorithm clearly enough to make conscious choices about when to optimise and when to protect what makes your podcast worth making in the first place. Those are not always in conflict. But sometimes they are, and you need to be able to see the difference.
The algorithm is not going away. But neither is the need for podcasting that reflects South African voices, South African realities, and South African stories told on South African terms.
Understanding the machine is the first step to refusing to be completely controlled by it.
Hendrik Baird is completing a PhD at the University of Pretoria researching the political economy of indie podcasting in South Africa. He is co-founder of Baird Media, host of the Become a Podmaster™ podcast, and a winner of the Muse Award for Best Podcast Script for STRIPPED – An Audio Drama Podcast.
Your voice is your brand. Your podcast should sound like it.
We help creators, coaches, and businesses make shows that stand out – for the right reasons.
Book a Free Strategy Session and let’s build something powerful.
Your RSS feed is what makes your podcast portable, distributable, and genuinely yours. Here’s how it works — and why it matters more than most beginners realise.
South African podcasting has a production problem — but it isn’t a talent problem. It’s structural. This Industry Insights analysis examines why the gap exists, what it costs the ecosystem, and who is positioned to close it.
Audio and video podcasting are converging. Here’s how South African brands can choose the right format — or combine both — for maximum impact.
Showmax is closing. The audience it served hasn’t gone anywhere. Here’s what South African brands and commissioners should be thinking about next — and what a professionally produced podcast series actually costs to commission.
The question is deceptively simple. The answer, research suggests, is not what the gear guides want you to believe.
Branded podcasts in South Africa are the most underused content strategy for corporates. Here’s why the opportunity is real — and why now is the time.