I just wrote a PageRank vs. TrustRank blog entry making the following main points: - Link value increases with time - Reciprocals are weak - Highly Concentrated Anchor Tags are Bad - Class C IP’s Should be Diverse I still see a lot of people talking about the value of links based on PR and many people give the trite response that PR is meaningless. If the above points are news to you, then the full read might be worth your time. Excuse my lack of blog -> forum -> blog etiquette ... Mods, feel free to give me a smackdown or let me know if I've made a faux paus by posting in this manner...
Interesting article, Nitin. However, a few clarifications: I don't think he was complaining about reciprocal linking per se. I think what he was pointing out is that paid links are fairly easy for Google to spot and discount and that consequently buying links won't help you unless you have extra money hanging around you want to get rid of. I don't see any evidence anywhere, nor official or semi-official statements from any SE, that reciprocal links are "bad". The problem with this theory, apart from the fact that again I see little evidence that it's true, is that organic linking also tends to use similar and sometimes the same anchor text - for example, how many people linking to DigitalPoint or even one of those upstart forums like WMW would use some variation on the term "SEO forum"? Similarly, how many of the organic backlinks to my psychogy forum would you guess contain the phrase "psychology forum"? I don't think "penalization" or "backlink devaluation" is the problem at all with links with similar or identical anchor text. I think the problem is that if all your backlinks target a single search term they can't help you much with your ranking for other search terms - THAT is why it is important to vary the anchor text when you have a choice. I'm not sure the meaning of TrustRank is yet known, let alone whether it is in current use. See thread by Mcfox elsewhere at DP for another view of what TrustRank might mean (having nothing at all to do with page quality or popularity or authority but having to do with security) - it makes at least as much sense as other explanations. This, on the other hand, is dead on. I suggest you might want to download a copy of fryman's "Oh geeze not this shit again" graphic for future use
There's a reason I titled them all "logical theories" instead of "Proven Rules" Based on lots of tests and lots of discussions these are just theories that I believe are logical. Reciprocals - I didn's say they were "bad", I said they were "weak". In other words, I don't think they are going to do much for you. Not sure if there's any official statements anywhere but when Matt was talking, esp, in his unofficial open session (which in many ways was more enlightening than the "official" talk he gave) he gave subtle indications that reciprocals weren't doing much for people. I can't remember the exact words he used but it was something like this ... "Yahoo directory links ... good. DMOZ links .. good. Looks like some other authority links ... good. Lots of reciprocals .. hmph. Paid links ... those aren't doing anything for you." ... that type of almost talking to himself instantaneous review of backlinks was common by Matt and the impression I got from it was that 1) he could easily identify them, and 2) he never told anyone they were helping them... so, my point is ... I doubt they help much and they are such a pain in the ass and time drain compared to other efforts that I just don't think they're worth investing in. Anchor Tags - The exception to the highly distributed rule would be words that are in the domain itself. I should probably clarify my blog entry for this. "Digital Point", "DigitalPoint", "DigitalPoint.com", "www.digitalpoint.com" all would have an expected high distribution. Other than high clustering around the domain or parts of the domain, I don't agree that natural linking will have low distribution. In the natural world, I think it is highly unlikely that the majority of backlinks to DP will include "SEO forum" in the anchor (and I just checked ... and they don't). But, when you say "or some variation" << that's exactly my point... in the natural world they will be variations .. they won't be exactly the same. In the natural world, far fewer than would use the domain name or the words in the domain name. And, my point isn't that natural linking wouldn't give you many with "psychology forum" in the anchor, but that there would be a high distribution of many different anchors. If you have 100 backlinks to your forums and 90 of them are "psychology forum" then you don't have natural backlinks. TrustRank - I should have probably not used "TrustRank" but I can tell you that several knowledgeable SEOs at Pubcon were discussing the theory that there is a "rank" that is different from the PageRank in the toolbar but serves the same purpose. The main point I was trying to make is that all the SE algorithms MUST include backlinks as a significant component in their weighting. Google's toolbar PageRank is no longer something that can be fully trusted when determining the value of a backlink. Often the toolbar will display a PR value for a page but that page won't be passing PageRank. Matt said many times things like "oh, that site doesn't pass PageRank anyway". So, in a small roundtable, we were using the term "TrustRank" .. but the name itself isn't important (to me anyway).
I think you are misinterpreting Cutts comments about reciprocal links, Nitin. There is nothing shady or even "unnatural" about reciprocal links per se, as I said. Natural or "organic" and relevant linking is often reciprocal (definition: site a contains a link to page B; site B contains a link to site A) and occurs even without arrangement between the respective webmasters. Authority sites link back and forth to other authority sites. What Cutts is talking about and what Google is devaluing are non-relevant artificially arranged (e.g., bought and sold) links - nopt reciprocal linking in itself. Again, I disagree. Using Google alerts, I often see non-arranged links to that forum that (1) use anchor text rather than just the URL and (2) use "psychology forum" or "psychology forums" because... well... how else would most people describe it? There are also many that use "Psychlinks forum" (cf. your Digitalpoint example). I doubt that either version is devalued. By the way, I'm not trying to attack your article. I'm just trying to discuss (and debate) the issues raised in it.
And I appreciate it! Quite possibly. It's just the impression I had and many others had as well. And based on my experience with building reciprocals, I just see the effort being way too high vs. other means of getting backlinks so I have a high aversion to them. On backlinks I thing we're saying the same thing... When I talk about high distribution of anchors in backlinks, what I mean is that given 100 naturally occurring backlinks, there would be a lot more than just 2 or 3 anchors. The distribution might loook something like this: 50% A=Domain, 25% Anchor2, 15% Anchor3, 5% Anchor4, 1% Anchor5-9 Often I see people that build new sites and they want to rank for "Purple Widgets" and every backlink they get, whether reciprocal, directory submission, coop, linkvault, whatever ... they use "Purple Widgets" as their anchor. This is extremely unnatural, easily identified, and I theorize that this won't work as well as mixing up the anchors.
Agreed. I'm just not convinced it's because they'll be seen as "unnatural" by Google but more because all of the weight of that anchor text is on a single search term/phrase. Thus, the problem isn't that it will be punished, but that's it's bad SEO.
I don't doubt it. That is exactly the kind of activity that Google is trying to target for devaluation.