How to Take Control of Your News Algorithm (Instead of Being Controlled by It)

Every news feed you scroll is ranked by an algorithm — and almost none of them work for you. Engagement-optimized feeds surface whatever keeps you scrolling. Chronological feeds bury the important under the recent. Either way, the ranking logic is a black box you can't see, question, or change.

Feedpik takes the opposite position: the ranking algorithm belongs to the reader. Here's what that means in practice, and how to use it.

Start with sources you choose

Feedpik speaks RSS, Atom, and OPML, so your feed is built from sources you picked — tech blogs, Substack newsletters, Google News queries for a ticker, YouTube channels, podcasts, even Bluesky accounts. Paste any URL or keyword into Discover Sources and Feedpik figures out the feed. If you're coming from another reader, one OPML import brings everything over.

That's the foundation: no algorithm can be yours if the inputs aren't.

See why every article is ranked where it is

Open the Recommended view and every article carries a score. Click one and Feedpik shows the reasons: how much came from topic match, source preference, freshness, language fit. No "the algorithm decided" hand-waving — an explanation you can read and disagree with.

Disagreeing is the point. Which brings us to tuning.

Tune it by talking to it

Click ✨ Assistant and tell it what you want in plain language:

• "Less crypto."
• "Boost fresh articles — I write a daily newsletter."
• "Why am I seeing so much Apple coverage?"

The assistant translates your intent into a concrete change — blocked topics, interest boosts, ranking weights — and shows you a diff card of exactly what would change. Nothing applies until you click Apply. Every applied change becomes a new numbered version of your policy.

Prefer sliders and forms? Everything the assistant can do also lives in Algorithm & Agents → Recommendation Settings, the ground truth for your ranking policy.

Roll back anything

Because every change is a version, there's no fear in experimenting. Cranked freshness too high and your feed turned into a firehose? Open your policy history and roll back to the version that worked. The audit trail records what changed, when, and who (or what) changed it — including changes made by an AI agent you've connected.

Save whole configurations as personas

A ranking policy tuned for deep AI research makes a lousy market open scan. Feedpik lets you save each configuration as a persona and switch between them in one click. Import a community Pack and it lands as a new persona instead of overwriting your setup.

The takeaway

"Controlling your algorithm" isn't about becoming a machine-learning engineer. It's three capabilities no mainstream feed gives you:

1. Transparency — see why every item ranks where it does.
2. Authorship — change the ranking with a sentence, see the diff, apply it.
3. Reversibility — every change is a version; rollback is one click.

Start free with the full reader, import your OPML, and ask your algorithm the question no other feed will answer: why am I seeing this?