How I Turn RSS Stars Into a Weekly Link Roundup

For most of last year, my Friday link roundup took longer than anything else I wrote that week. I keep around four hundred feeds, I star whatever looks worth sharing, and by Friday afternoon I would open my starred list to find 60 or 70 items with no memory of why I had saved half of them. Re-reading enough of each to decide took three hours on a bad week. Twice I skipped the roundup entirely and told the team it was a "quiet week," which it was not.

The problem was never a shortage of good links. It was that starring had become a way of postponing a decision, and Friday was when all those postponed decisions came due at once. What fixed it was not a new tool. It was changing what a star means and doing about thirty seconds of work at the moment I save something.

Why a starred list turns into a second inbox

A star in most feed readers carries exactly one bit of information: you liked this, at some point. It does not record whether the piece is worth sharing, who it is for, or what you thought when you read it. So the list grows with items of very different value, and every one of them looks the same on Friday.

I counted one typical week. Of 64 starred items, 9 were things I wanted to read later for myself, 11 were reference material for an article I was drafting, 17 were mildly interesting, and only 27 were real candidates for sharing. I eventually used seven of those. The roundup was taking three hours because I was sorting four different kinds of saving in one pass, and doing it with no notes.

If your reader already feels like a backlog, I wrote about the broader version of this in why an RSS reader turns into another inbox. The starred list is the same failure in a smaller container.

Give a star one job

The first change was to stop using the star for anything except roundup candidates. Read-later material now goes into a separate read-later app. Research for my own drafts gets a tag named after the draft. The star means only one thing: "I might share this on Friday."

Most readers can do this without anything extra. Feedbin, Inoreader, NetNewsWire, and Miniflux all let you separate starred items from tagged or saved ones, though the names differ. If your reader only has stars, you can still get most of the benefit by sending read-later items somewhere else, like Instapaper or a plain notes file, before you star them. Winifred has a useful piece on building a reading list you actually finish that covers that side well.

The second change mattered more. When I star something, I write a single sentence about why, right then, while the article is still fresh. In Feedbin and Inoreader I use the notes or annotation feature. When a reader has no notes field, I copy the link into a running text file with the sentence underneath it. It takes perhaps half a minute. That sentence becomes the blurb in the roundup almost unchanged on most weeks.

A star without a reason attached is a decision you have only postponed until Friday.

"Share everything good" makes a worse roundup

The usual advice for link roundups is to be generous. Include everything interesting, let readers pick what they want, more is more. I followed that for about a year and my click data from our team newsletter said otherwise.

When the roundup had 15 or more links, the median link got clicked by 4 of our 38 readers. When I capped it at seven, the median rose to 11. The total number of clicks across the whole roundup was a little higher with seven links than with fifteen. People were not reading more of a longer list. They were skimming it and picking fewer.

I now hold to seven, with at most two from the same source. If I have nine good candidates, the two that lose either get held for next week, if they will still be relevant, or get dropped. Dropping links I liked was hard at first. It got easier once I noticed that nobody ever asked about the ones I cut.

The Friday routine, now about forty minutes

This is the workflow I use every week. Times are from my own timer over the past two months.

1. Open the starred list and sort oldest first (2 minutes). Anything older than ten days gets unstarred unless it is clearly evergreen. News-style posts go stale fast.
2. Read only the notes, not the articles (8 minutes). The one-sentence reason decides it. If the sentence is weak or missing, the link probably does not make the cut.
3. Pick seven (5 minutes). Check for variety of source and topic. Two pieces from the same outlet are fine, three almost never.
4. Write the blurbs (15 minutes). Start from the saved sentence. Add what the reader gets from clicking: a number, a method, a counterargument. Keep each to two sentences at most.
5. Order and title (5 minutes). Put the strongest link first, not the newest. The roundup subject line names that first link.
6. Clear the stars (2 minutes). Unstar everything, including what did not make it. Held items go into a tag called "next week," and that tag gets checked first next Friday.

The last step feels wasteful, but it is the one that keeps the list from rebuilding itself into seventy items. A starred list that never empties stops being a curation tool.

What a single entry looks like

For anyone who has not written one, here is the shape I use. A bare link with the original headline tells the reader nothing about why you picked it. Compare the headline alone, which might be something like "Notes on feed polling intervals," with what actually appears in my roundup:

Notes on feed polling intervals (a developer's blog). Measured how often 1,200 feeds really change and found most update less than once a day, which is why polling every 15 minutes mostly wastes server time. Useful if you run a self-hosted reader.

The source in parentheses, one finding with a number, and who should care. That last part saves readers a click when the piece is not for them, and they trust the next recommendation more for it.

Folder structure helps the starring side too. Feeds that reliably produce shareable material sit in one folder I check daily, so candidates surface early in the week instead of all on Thursday. If your folders are a mess, organizing RSS folders so you stop missing posts is a good place to start.

Try it on next Friday's list

From today until next Friday, every time you star something, write one sentence about why before you move on. Do not change anything else yet. On Friday, build your roundup using only those sentences and see how many of your stars you actually need to open again. For me, that one habit cut the job by more than half before I touched any other part of the routine. More on keeping feeds manageable is in the curation section.