Vetting Sources for a Curated Feed With a 30 Day Trial
Last spring I counted the feeds behind my weekly link roundup: 143. Then I checked which of them had produced a link I actually used in the previous three months. The answer was 19.

Last spring I counted the feeds behind my weekly link roundup: 143. Then I checked which of them had produced a link I actually used in the previous three months. The answer was 19. The other 124 were noise I scrolled past every morning, and a few of them had quietly sent me stories that turned out to be rewrites of rewrites.
Vetting sources for a curated feed is the unglamorous half of curation. Everyone talks about picking great links; far fewer people talk about deciding which sources earn the right to be read at all. This is the 30 day trial I now run on every new source, with the signals I check, the numbers I track and the rules for cutting.
Why Vetting Sources for a Curated Feed Comes First
Your readers never see your sources. They only see the 8 or 12 links you pick each week, and they trust that you saw the original before you passed it on. One wrong link from a careless aggregator costs more trust than ten good links earn.
There is also a plain time cost. At 30 seconds of skimming per unread headline, 120 weak feeds posting twice a day eat two hours a week. That is two hours you could spend reading the strong sources properly and writing better notes on what you share.
Where New Candidates Come From
Good candidates rarely arrive from lists titled "best blogs to follow". They come from the trail behind something you already liked.
- Citations inside strong articles. When a piece you saved links to a study, a dataset or a small blog, that cited source is the candidate, not the article that cited it.
- Bylines that keep reappearing. If the same researcher or writer turns up in three of your saved links, find where they publish directly.
- Replies, not reposts. On social feeds, the person who corrects a viral claim with a primary document is often worth more than the account that posted it.
Many of these sites have no visible subscribe button. The tricks in our guide on finding the hidden RSS feed on almost any website will turn up a feed for most of them.
The 30 Day Trial in Practice
Every new source goes into a single folder in my reader called Probation. It never goes straight into a main topic folder. Once a month, on the first Sunday, I review every feed in that folder against three numbers.
| Number | How I count it | Keep if |
|---|---|---|
| Posts published | Shown by most readers per feed | Fewer than about 40 a month |
| Posts opened | Read past the headline | At least 1 in 4 |
| Posts saved or shared | Starred items in the reader | At least 2 in the month |
A source that passes moves into its topic folder. A source that fails gets one more month only if I can name a specific reason, such as a holiday lull. After two failed months it goes, with no guilt. The point of the trial is that the decision is made by what I actually did, not by how impressive the source felt when I subscribed.
High volume sources get one extra chance. A news site posting 300 items a month will fail the first number, but if one section of it is excellent, look for a category or tag feed covering only that section. Subscribing to the narrow feed often turns a failure into one of the best sources in the reader.
Quality Signals Worth Checking in the First Week
The numbers tell you whether a source is useful to you. They do not tell you whether it is reliable. For that, I spend ten minutes on each candidate in its first week.
Does it link out to the original?
Open three recent posts. If the claims trace to a named report, court filing, changelog or dataset, good. If every link points to another article from the same site, you are reading a closed loop.
Is there a named person and a corrections habit?
Look for author names and an edit note on any post that changed. A source that never admits a mistake has either never made one, which is unlikely, or never fixes them.
How far is it from the event?
Count the steps between the source and the thing it describes. A developer posting their own release notes is zero steps. A blog summarising a news site summarising that post is two. For most topics, keep sources at zero or one step away.
The Advice I Stopped Following
The common advice is to follow the biggest names in your area, on the theory that the most cited sources must be the most valuable. In my experience the big names are the least useful feeds to add, for a simple reason: you will see their best work anyway. It arrives through ten other people quoting it.
The value in curation is the source your readers would not have found on their own: a small agency blog, a hobbyist who benchmarks carefully, a researcher's notebook updated twice a month. I now favour feeds that post less than once a week and have fewer followers than they deserve. Of the 19 sources that survived my cull, 13 publish fewer than six posts a month.
Keeping the List Small Once It Is Clean
A clean source list decays without a rule. Mine is one in, one out: when a source graduates from Probation, I have to remove one from the main folders. That keeps the total near 30 feeds, which I can read properly in about 25 minutes a day.
I also keep a short note against each feed saying why it is there, one line like "first to post official API changes". When I review the list in January, a feed whose note no longer holds is the first to go. If you organise feeds by topic, the folder layout described in organizing RSS folders so you stop missing posts makes this review much faster.
Source Vetting Questions
How many sources should a curated feed have?
Fewer than most people think. Somewhere between 20 and 40 is manageable for one person reading daily. Past 60, you start skimming headlines and missing the posts that matter.
Should I keep a source that is often wrong but sometimes first?
Keep it out of your main folders. Put it in a separate tip line folder and treat anything from it as a lead to confirm elsewhere, never as a link to share directly.
Does this work for newsletters and social accounts too?
Yes. Route newsletters into your reader where possible and run the same three numbers. For social accounts, a private list works as the Probation folder.
Start Your First Trial This Week
Create a Probation folder today and move every feed you added in the last three months into it. On the first Sunday of next month, check the three numbers and cut anything that fails. Then, when you share what survives, write a sentence on why each link matters; the method in writing link annotations readers actually click pairs well with a source list you trust.
More in Curation
Curation
A Tagging System for Saved Articles That Survives a Year
I exported my read-later library last spring to check something unrelated and found 1,847 saved articles carrying 240 distinct...
Curation
How to Handle Paywalled Links in a Curated Reading List
The complaint that finally changed how I do this arrived in March, from a reader who had followed my weekly links for about two...
Curation
How to Share an OPML Bundle People Will Actually Import
Someone asked me in September for my feeds on soil science and small presses.