Debra Mastaler asks Matt Cutts: Does Google pass 2nd link value?

By Halfdeck on July 10, 2008 at 4:45 pm

Matt Cutts: “Your question is short, but the answer is more complex. Typically if the anchortext on the two links is identical, we would probably drop one of those links.”

Good job pushing this issue. Again the majority of SEOs turned out to be off the mark.

By seo_scientist on July 10, 2008 at 10:46 pm

hey guys, don’t give him a break. be as tough on him as you are on all the other seos and testers out there. Cool of him to respond on this issue. If he could only now say what happens in case there are two links to the same page using different anchor texts…

By seo_scientist on July 11, 2008 at 7:22 am

“That’s what tests are for neyne”

but as we saw, every test can (and will) be disputed. At the end of the day, Google is a black box and any test result can be just representative of that day’s algo, the settings of the browser of whoever is checking the results. No test can take all of the factors involved, since we don’t know what all of those factors are. So we shouldn’t test at all ? I know that I try to keep as many variables constant (as much as I can) except for the ones I am testing. Reversing the situation helps a lot and repeating the results on unrelated, separate testing ground helps even more, but you have just increased the probability that your results are teaching you something valid, there is never a definite proof. So I think it is a bit unfair to come with such demand.
In any case, I would be more than happy if Matt came out and said that Google will always pass value from all the links to a single domain. Even if that means my results are completely wrong, I would at least know what not to look at the next time. And we would have a definite answer on that issue, so it would all be worth it.

Unfortunately, his vague answer did not solve this issue. He could have responded “no Debra, Google will not pass value only from the first link, although there are some situations when this could appear to be the case”. Would this reveal something about the algorithm? I don’t think so. So you can look at what he said and you can look at what he didn’t say. That is what happens with vague answers.

By Halfdeck on July 11, 2008 at 4:19 pm

“Unfortunately, his vague answer did not solve this issue.”

Neyne, I don’t expect Matt Cutts to try to dumb down 1000 lines (figuratively speaking) of if-then-else-which-case code into one yes or no answer. Matt Cutts also has no obligation to respond to these kinds of questions. So I appreciate whatever he contributes to the discussion. If someone else wants to squeeze more juice out of Matt, that’s fine with me. But that’s Graywolf and Rand Fishkin’s area of expertise, not mine. If Graywolf published another anti-Google post slamming Matt Cutts for giving out vague answers, Matt might take the bait.

Previously I imagined Googlebot maintaining a List object, to which Googlebot would add one URL on a page at a time after pulling them out of a page using a regexp call (or whatever HTML parsing tool they use). The first URL would automatically make the List; duplicates would not be added. Then Googlebot would check for nofollow, and if a link has nofollow, it would be recorded but wouldn’t be added to the link graph.

That process only involves two IF statements. It also wrongly assumes links are processed immediately during crawl. Michael Martinez’ post about second link at least temporarily sticking suggested to me the process is more complicated than that; Matt Cutts confirmed it. First, Michael’s test (regardless of whether its dirty or not) suggests all links on a page is recorded during crawls; some are then dropped later post-crawl. That delay allows for more complicated algos to be run (since doing them during crawl I imagine would slow down Googlebot). Second, Matt’s answer confirms there are more than two IF statements - probably many more. We now know for certain one of those IFs checks for duplicate anchor text.