Fourteen of 34 blogs never tell Google who wrote the article. Thirteen of them are engineering blogs.

A client asked me last week whether author bylines are worth the argument with their legal team. I said yes. Then I realised I had no idea what anybody else actually does, so I wrote a script that reads one article from each of 35 blogs and looks for a name.

Thirty-four feeds answered. Asana has no feed I could find, so it fell out.

Twenty of them hand a search engine an author. Fourteen hand it nothing.

On its own that is a useless number, because these 34 are not one population. So I split them, and the halves barely resemble each other.

The marketing and SEO blogs name a person almost every time. Thirteen out of fourteen. The one exception is Buffer, which I will come back to, because I cannot explain it.

The engineering blogs, developer docs and design research do it six times out of twenty. Same script. Same half hour.

The people who write about demonstrating expertise mark up their expertise. The people who have it mostly do not.

What to check on your own site

Check whether your CMS emits an author at all. Not whether a byline renders. That is a template question and you already know the answer. Open the page source, find the JSON-LD, look for the field.

If it exists and points at a name with nothing behind it, write the page. Two hundred words on what this person has actually done. A rater can use that. A reader wants it. It also takes real time, which is presumably why it is the piece most often missing.

Then check the tool you are auditing with. Mine lied to me for most of an afternoon.

The blogs I did not expect

I have read web.dev for years. Its writers are Google engineers. Several of them wrote the specifications they are explaining, and their names sit on the page where anyone can see them. In the structured data there is no author field at all. Chrome Developers, the same. Then Cloudflare, Netlify, Stack Overflow, Shopify Engineering, Nielsen Norman Group, A List Apart and Buffer. Nine sites in this set run author pages you can click, and still publish articles that, to a parser, nobody wrote.

Then the opposite fault. Smaller, more irritating. Moz, Stripe, Semrush, Yoast and Baymard name a person, and the name goes nowhere: no profile, no bio, no way to learn whether this person has ever done the thing they are describing. Litmus manages something stranger. Kimberly Huang sits in the markup and never appears on the page a reader sees.

The part where I was wrong

My first run said eleven blogs emitted an author with no type. I read that as sloppiness and nearly wrote it up as such. Ahrefs, Yoast, Kinsta, Hootsuite, VWO, apparently none of them able to say whether their writer is a human being.

The field holds a reference, not an object. Yoast puts a bare @id in there, a pointer at a Person node defined elsewhere in the same graph. Ordinary JSON-LD, valid since 2014. My parser took the pointer at face value, found no type inside it, and shrugged. Following the pointer moved all eleven from unknown to Person. It also moved my headline number from 8 to 19, which is the kind of correction you want to make before publishing rather than after.

Forty minutes. One wrong article avoided. Make sure whatever you audit with follows @id, or it will call perfectly good markup broken.

What I still cannot explain

Buffer is the one marketing blog in the set with no author in its markup, and Buffer has written more about transparency than anyone else on this list. Their byline is on the page. Their author pages exist. The field is simply absent.

I have guesses and I believe none of them. A migration dropped it. Somebody made a performance call and never wrote it down. Or the web.dev thing happened: engineers built the site, engineers do not think about raters, the ticket was never filed.

One article per blog is one article per blog. It tells you what a crawler met on the day it arrived, which is roughly what a crawler gets anyway. It does not tell you why.

Questions people ask about this