One of my earliest SEO wins involves a tale of two sites.
Site #1 dominated the rankings. It ranked #1 for "linux hosting" and dozens of other nationally competitive results for years. It ranked for tens of thousands of nice variations, too.
Site #2. That was a mystery. About as old. Also commerical. More authority from inbound links than the other. No spammy practices. I was running both from the start and ran both the same way. Yet the second site just wouldn't rank for anything.
So I struggled. I posted on the Search Engine Watch forums. Back then, Danny Sullivan was in charge, and that was basically the only place you could find those mythical SEO gurus.
There, I gathered lots of terrible-to-passable advice. But for the most part, it stumped everybody.
Then I did a crawl. I realized that all internal links were redirects. It was nearly invisible to the naked eye because it only involved a trailing slash on the end of every web address. Menus, direct links to other pages, and so forth: all of them lacked a slash at the end of the URL, and all of them redirected.
So I fixed it.
Then, after 2+ years of struggle, I watched site #2 blow site #1 out of the water, ranking on every expensive telecom/managed services query from "dedicated servers" to "[CITYNAME] colocation" and beyond.
From that point on, fixing redirects seemed to be one of my most reliable SEO magic tricks ever.
Why Did That Work?
Almost nothing in SEO happens in a vacuum, so we're never certain. Then, my uneducated guess was just that I was signaling that I was a more attentive website-keeper.
As I've leveled up my SEO skills, I learned that Google had told all of us, as direct as can be, that PageRank decays.
In 2005, people were popping up "link wheels" all over. Linking a "circles" of different sites, attempting to throw Googlebot into an infinite loop for infinite link value.
When you think about it like that, decaying PageRank just makes sense. Even before considering that Google's going out of their way to shut down those shenanigans directly. You shouldn't be able to pass an infinite value to your own stuff by interlinking.
Might Google "decay" redirects, so that each redirect has incrementally less PageRank value?
That's been directly stated by Matt Cutts more than once, and even in Larry Page's original PageRank patents.
The simple message we were given? Redirects decay PageRank just like links.
Then, it all changed (maybe).
Last week, Google's Gary Illyes tweeted:
30x redirects don't lose PageRank anymore.
— Gary "鯨理" Illyes (@methode) July 26, 2016
Well strap a pig to the mud and feed it a biscuit Google, what are we supposed to think now?
My take: I have no reason to doubt Gary. Or Matt. Or Larry. Google's staff have said plenty of things in the past that don't pan out in testing. Is it because their agenda is to get everybody acting a certain way? Do ends justify means?
Maybe. But probably not.
Just like a lot of other contentious SEO topics, Google and actual testing can disagree and both be right.
Ranking factors are independent functions and Gary is only talking about one of them. Here's an example.
Redirects, especially chains of them, slow everything down. And we know this:
And I'm sure some of you can find plenty more conflicting scenarios if you thumb through the ranking factors.
I'm mostly taking Gary at his word until I see conflicting evidence, but for now, we're downgrading Redirected Internal Links from Probable to Iffy on the list of direct, negative factors.
Still, if you want competitive SEO, take nothing for granted. The best practice is still to avoid unnecessary redirects. All that's up for debate is how much impact this has on your rankings.
And you can settle that for yourself by just doing things right in the first place and paying attention to what happens next.