If someone on your team is still telling you to nofollow your login link or your utility pages so the "saved" PageRank flows to your money pages, you are probably asking does nofollow sculpting still work, and the short answer is no, that advice is out of date, and has been since 2009. PageRank sculpting with nofollow did work once, in a narrow mechanical sense: Google's own engineer confirmed that the old calculation let you exclude a link from the sharing formula and give the rest of your links a bigger piece. But Google changed that calculation more than a year before it explained the change in public, and nothing since has brought the old benefit back. Here is the grade on the evidence: confirmed that the mechanism once existed, strongly supported that it does not work today, and weak to nonexistent on any claim about how much ranking benefit anyone actually got from using it.
Knowing the nofollow PageRank sculpting history matters because a lot of old SEO advice still repeats the pre-2009 version of the rule without ever mentioning the change.
What did PageRank sculpting with nofollow promise?
PageRank sculpting is the general idea of deliberately changing which links receive or pass along PageRank, so that certain pages get favored. The specific tactic this article is about, nofollow PageRank sculpting, is narrower: you add rel="nofollow" to some internal links while leaving others alone, hoping the links you leave alone pick up a bigger share of the page's PageRank.
The idea traces back to January 2005, when Google introduced nofollow to fight comment spam on blogs. It was meant to stop spammers from dropping links in comment sections to farm PageRank, not to give site owners a lever for redistributing their own internal links.
That changed by 2007, when word got around that Matt Cutts, who was Google's most visible spokesperson on search quality, had talked about YouTube using nofollow on some of its internal links. People in the SEO industry picked up on it and started nofollowing internal links on purpose, hoping to concentrate PageRank on their most important pages instead of spreading it evenly across a sign-in link, a terms-of-service page, or something else low-value.
The core idea: removing a link from the calculation that divides up a page's PageRank meant the other links got a bigger slice. That is a different thing from simply withholding credit from one page. One is subtraction. The other is redistribution. The tactic only works if both happen.
Did it ever work? What the evidence supports
Yes, in the specific, limited way Google itself described. In June 2009, Matt Cutts laid out exactly how the old math worked. Picture a page with ten points of PageRank to pass along and ten outgoing links. If you nofollow five of those links, the other five links split the share that would have gone to all ten, so each one gets two points instead of one. That is a real mechanical advantage, and Google confirmed it existed.
What the evidence does not support is any claim about how much that mechanical advantage actually moved a site's rankings, traffic, or revenue. Rand Fishkin, writing for Moz in June 2009, said his own consulting clients had seen strong results from this kind of sculpting. That is a real account worth taking seriously, but it is one consultant's experience, not a controlled study that isolates nofollow as the cause and rules out everything else that could have changed a site's rankings at the same time.
So hold two things at once here. The redistribution mechanism was confirmed by the company that ran the algorithm. A reliable, measurable ranking payoff for the average site that tried it was never established by anyone, Google or otherwise. Treat "it had a real mechanism" and "it reliably worked for rankings" as two separate claims, because the evidence only backs the first one.
What Google changed before its 2009 disclosure
Here is where the history gets specific enough to matter. Cutts did not say Google turned off sculpting in June 2009. He said Google had changed the underlying calculation more than a year before that, and the June 2009 post was Google finally explaining a change it had already made quietly.
The simplified math looks like this. Same page, ten PageRank points, ten outgoing links, five of them nofollowed.
| Setup | How many links count toward the split | Points per link that still counts |
|---|---|---|
| No links nofollowed | 10 | 1 |
| Old treatment (before the change) | 5 | 2 |
| New treatment (after the change, explained in 2009) | 10 | 1 |
Under the old treatment, the five nofollowed links simply didn't count. The other five split everything, two points each. That's the arithmetic that made sculpting look worth doing.
Under the new treatment, all ten links still count toward the split, nofollowed or not. The five ordinary links go back to getting one point each, same as if nothing had been nofollowed at all. The share that would have gone to the nofollowed links doesn't get handed to the other links. It just doesn't flow anywhere through that page's links.
That one change is the whole story. Nofollow stopped being a way to pull a link out of the sharing calculation while keeping its share for someone else. It became a way to withhold credit from one destination, full stop, with no bonus for the destinations you left alone.
One more date distinction matters here, because it's easy to get backwards. Google explained the change in June 2009. Google made the change more than a year earlier than that, sometime in 2008. If you see a source say "Google killed PageRank sculpting in June 2009," that's describing when the news came out, not when the underlying math actually shifted.
Why the loophole closed, and why the industry disputed it
Cutts said the change came out of concerns raised by Google's own crawling, indexing, and search quality teams, who were watching sites try to channel PageRank around in ways that worked against the open structure the system was built to measure. He was also fairly blunt that he had never been a big fan of nofollow sculpting as a recommendation in the first place, and that it had always been a minor, secondary idea rather than something Google pushed hard.
Fishkin, in that same June 2009 period, offered his own read on why Google might have made the change: he suspected it had to do with abuse, sites gaming their internal link structure in ways that hurt the quality of what Google was trying to measure. That's worth including as context, but it is Fishkin's informed guess, not a confirmed statement from Google about its internal reasoning. Keep those two separate when you're deciding what to believe.
What made this genuinely frustrating for marketers at the time was the gap between when the change happened and when anyone outside Google explained it clearly. Cutts himself admitted the delay: he said Google figured site owners running their own tests would notice the shift and said something, but they didn't. So advice telling people to nofollow sculpt kept circulating at conferences and in blog posts for over a year after the underlying math had already changed out from under it. That gap is a big part of why outdated advice about this tactic is still floating around today. It got baked into a lot of SEO training material before anyone corrected the record, and old material doesn't disappear just because the facts underneath it moved.
Did the nofollow hint change bring sculpting back?
No, and this is the point where a lot of confusion creeps back in. In September 2019, Google announced that nofollow, along with two new attributes, ugc for user-generated content and sponsored for paid links, would be treated as hints rather than hard directives. That ranking-related change took effect right away. A related change, where nofollow also became a hint for crawling and indexing rather than a strict block, took effect the following March.
Here's what that announcement did not say: it did not say Google brought back the old denominator-exclusion math. It did not describe any new way for a nofollowed link's unused share to flow to other links on the page. It changed how Google interprets a nofollow signal when deciding whether to follow a link or count it as an endorsement. That's a different question entirely from whether nofollowing some links gives the others a bigger slice of PageRank.
So if someone tells you the 2019 hint update quietly reopened the sculpting loophole, that's reading something into the announcement that Google never put there. A hint about whether to trust a link is not the same thing as a redistribution rule for how much credit flows through the links you left alone.
Current Google documentation on qualifying outbound links, last updated in December 2025, backs this up. It says ordinary links don't need any special markup at all, and that nofollow remains useful for links you genuinely don't want to vouch for, things like paid placements or user-generated content where sponsored or ugc fits better. Nowhere in that guidance is there a mechanism for recovering PageRank you've withheld from a link. Nofollow still has real, legitimate uses. It just isn't, and hasn't been since 2008, a dial you can turn to redirect credit toward your favorite pages.
The verdict, and how to spot outdated advice
If you came here asking does nofollow sculpting still work in 2026, here is the short version: nofollow sculpting had a real mechanism, Google shut that mechanism off more than fifteen years ago, and nothing since has brought it back. If your current advice tells you to nofollow internal links specifically to concentrate PageRank somewhere else, it's working from a version of Google that stopped existing around 2008.
A few ways to spot this kind of outdated guidance when you run into it:
Any claim that nofollowing a link saves its PageRank share for your other links is describing the old math, not the current one. Under the math Google explained in 2009, all your links still count toward the split whether you nofollow them or not. Withholding credit from one page does not hand that credit to another.
Any claim that the 2019 hint change reopened sculpting is stretching an announcement about how Google weighs a signal into a claim about redistribution that Google never made. Those are different mechanisms, and only one of them was ever confirmed to exist.
And any claim that using internal nofollow automatically triggers a penalty is overcorrecting in the other direction. Losing a redistribution advantage you never really had a reliable fix on is not the same thing as Google punishing you for using the attribute. Nofollow still has legitimate jobs to do, marking sponsored links, marking user content you can't vouch for, and you don't need to strip every instance of it from your site out of fear.
What would actually change this verdict is a new, explicit statement from Google describing a restored way to concentrate PageRank through selective nofollowing, or solid, independently reproducible evidence that such an effect exists today. Neither one currently exists. Until it does, treat nofollow as a tool for marking the links you don't want to vouch for, not as a lever for steering authority around your site.
If your backlog of SEO advice includes anything written before 2009 that mentions sculpting PageRank with nofollow, it's worth a second look. That is exactly how the nofollow PageRank sculpting history keeps resurfacing: a lot of old SEO wisdom gets quietly carried forward into briefs and style guides years after the thing it describes stopped being true.
