The finding
The people who write about Apple mostly build light websites. 6 of the 9 contacted two advertising domains or fewer. The median is 1 ad domain and 11 third parties, against 2 and 14 for the other 868 sites in our catalogue. 4 of them run exactly one ad domain, and it is the same one in every case.
Then the range breaks. The lightest source contacts no advertising domain at all, and the median across the other eight is 1. The heaviest contacts 58, and 110 third parties in total. There is no sensible multiple to quote against a source that loads zero, which is rather the point: these publications compete for the same readers on the same stories and are not on the same scale.
The second pattern is that ad weight and reading comfort are not the same axis. The two sources with the lowest design scores load three advertising domains each, not fifty. What costs them is typography: 14px body text at 100 characters a line on one, and a surface sitting over the article on the other. The heaviest site by ad count has perfectly reasonable body text.
| Source | Ad domains | Third parties | Body text | Design | Reading | Shift |
|---|---|---|---|---|---|---|
| The Eclectic Light Company eclecticlight.co | 0 | 3 | 18px / 67cpl | 6/10 | 10/10 | 0.003 |
| MacStories macstories.net | 1 | 9 | 18px / 70cpl | 6/10 | 10/10 | 0.005 |
| TidBITS tidbits.com | 1 | 7 | 18px / 68cpl | 8/10 | 10/10 | 0.006 |
| Six Colors sixcolors.com | 1 | 11 | 15px / 92cpl | 6/10 | 8/10 | 0.009 |
| 512 Pixels 512pixels.net | 1 | 13 | 17px / 70cpl | 6/10 | 10/10 | 0.057 |
| Michael Tsai mjtsai.com | 2 | 2 | 16px / 87cpl | 5/10 | 8/10 | 0 |
| AppleInsider appleinsider.com | 3 | 11 | 18px / 93cpl | 3/10 | 8/10 | 0.014 |
| MacRumors macrumors.com | 3 | 15 | 14px / 100cpl | 2/10 | 7/10 | 0 |
| 9to5Mac 9to5mac.com | 58 | 110 | 18px / 87cpl | 5/10 | 9/10 | 0.004 |
Source by source
The Eclectic Light Company
0 ad domains · 3 third partiesThe only source in the beat that contacted no advertising domain at all on the load we measured.
feedhttps://eclecticlight.co/feed/
MacStories
1 ad domain · 9 third partiesMember-funded, and it shows in the measurement: one ad domain and nine third parties.
recognised: doubleclick.net
feedhttps://www.macstories.net/feed/
TidBITS
1 ad domain · 7 third partiesBest design score in the beat, and running since 1990. The only one whose consent banner we had to clear by hand.
recognised: doubleclick.net
feedhttps://tidbits.com/feed/
Six Colors
1 ad domain · 11 third partiesLight on ad tech. The reading score is held back by a 15px body size and long lines, not by advertising.
recognised: doubleclick.net
feedhttps://sixcolors.com/feed/
512 Pixels
1 ad domain · 13 third partiesThe most layout shift in the beat, though 0.057 is still well inside what Google calls good.
recognised: doubleclick.net
feedhttps://512pixels.net/feed/
Michael Tsai
2 ad domains · 2 third partiesFewest third parties of all nine, and zero measured layout shift. Carbon rather than Google.
recognised: carbonads.com, carbonads.net
feedhttps://mjtsai.com/blog/feed/
AppleInsider
3 ad domains · 11 third partiesThree ad domains, but a low design score: our vision pass flagged a surface covering part of the page.
recognised: doubleclick.net, googlesyndication.com, jwplayer.com
feedhttps://appleinsider.com/rss/news/
MacRumors
3 ad domains · 15 third partiesThree advertising domains, and joint highest reach in the beat. The lowest scores here are typographic: 14px body text at 100 characters a line. It is also the only source in the beat that meets the blocking-surface test in our wider report, on an open navigation menu rather than an advert. Our classifier separately flagged apple.com, which we have excluded: it is an iTunes and affiliate link, not advertising.
recognised: doubleclick.net, googlesyndication.com, skimresources.com
feedhttps://feeds.macrumors.com/MacRumors-All
9to5Mac
58 ad domains · 110 third parties58 advertising domains and 110 third parties, against a median of 1 and 10 across the rest of the beat. Free to read at scale, funded by programmatic advertising, and the only source here running a full consent management platform.
feedhttps://9to5mac.com/feed/
Method, and what these numbers do not say
Each homepage was loaded once in a real browser on September 26, 2026, from Ireland, at desktop size, with the cookie banner accepted so the page ran the way a reader actually sees it. We counted third-party registrable domains and classified them with Mozilla's public third-party list, measured layout shift over a six second dwell, then opened an article and read its real computed body size and line length. Design is scored by a local vision model on a screenshot. Same instruments, same thresholds, as every other source we publish.
Five limits worth more than the numbers. This is one page load on one day from one country, and ad auctions vary between loads, so read a per-source count as a sample rather than a verdict. It is a homepage, and readers arrive on article pages. Our classifier left some domains unrecognised on every site, so each ad count is a floor rather than a total. Nine sources is a small sample, and a median of nine is a fragile thing. And an ad domain is not a sin: several of these publications are free to read because of it, and the heaviest is also one of the most prolific.
We should also say where we stand. FeedsBar is a Mac app, so these nine are our press, and we carry all of them in the app by choice. None of them asked to be measured, none paid to be here, and nothing about these numbers changes which we carry. Full method and weights: feeds.bar/signal-500/method. The wider study of 890 news sites is here.
Publishers: if a number here looks wrong, it may well be. Tell us what you see and we will re-measure and correct the page, the same way we corrected our own interstitial figure. Cite or reproduce any of this freely with a link back.
All nine ship in FeedsBar as the Mac set, one tap to switch on. Every feed URL is listed above, so you can equally paste them into whatever reader you already use, and hundreds more feeds are scored and copyable the same way.