[icopyright horizontal toolbar]
The JCPenny fiasco only pointed out what has been apparent for some time, Google and the other search engines are less sophisticated at weeding out link spam than they are given credit for. I believe part of the myth lies with AdWords, which makes a fine point about matching the relevancy of keywords to landing pages. In fact, if you attempt to use wildly inappropriate keywords, Google will hike up you CPC to the point where it’s unprofitable to proceed.
Why Can’t Google Recognize LinkSpam?
If they can do this, why can’t they recognize link spam, forum spam, blog spam, etc? Two answers: (a) they aren’t as good as they say, and (b) for various reasons, they aren’t motivated to apply what they know. In truth, the reason really doesn’t matter, the fact remains that as long as the current system of using backlinks to rank websites is in place, there will be plenty of abuse.
Well, what should the search providers do? The most draconian method is simply to ignore backlinks. Talk about the baby and the bathwater! What would be the replacement proxy for popularity? I propose that search engines disentangle popularity from authority. They are two different concepts and should be handled differently. Bernie Madoff’s Ponzi scheme was wildly popular – enough said.
The bond industry has an interesting model. Companies like Moody’s and S&P rate bonds so that purchasers can properly evaluate the bonds’ prices. Yes, I know, ratings scandals in 2008 helped get us into the economic mess we’re in. The fix was to remove the conflict of interest in how the rating companies were paid, not to abolish ratings.
Let’s extend the model to the backlink business. First, can we stipulate that, say, 100,000 backlinks to a single site is fishy. How about 10,000? 1,000? At some point, the sheer volume of links should become a negative. Second, let’s agree that a link spider can determine if a backlink is “real” in the sense that the linked entities deal with related content. So, our model is a company with a realistic number of backlinks that wants to prove those links to be legitimate, and wants the search engines to reward them for having “good” links.
In this model, 10 good links to Sears would have overcome 10,000 bad links to JCPenny.
Enter the backlink certifiers. Imagine a dedicated army of webcrawlers whose only jobs is to evaluate the relevancy of links; it certifies the good ones and assigns them serial numbers. The serial number database would be available to all subscribing search engines. The search engines would now have a new metric for ranking websites rather than just relying on popularity. From now on, only quality links would be used to rank websites — a practice that would have foiled JCPenny. To keep the certifier’s honest, also set up an industry watchdog to investigate abuses and punish bad actors.
One could ask: why not have each search engine do these link evaluations directly? Well, first, I don’t trust them. They are in a conflicted position as witnessed by the JCPenny case, which is a large paid advertiser. Second, uniformity of results will only arise from employing a common mechanism. Third, cover – the certifier would take the heat from parties unhappy about with its judgment, thereby protecting somewhat the search engines from direct pressure.
I can hear the naysayers now: too costly, too time-consuming, won’t work, will be gamed, etc. To them I say: pilot project. Let’s put it to the test. Let the search engine industry fund a startup consortium for link certification, and apply to it all the lessons learned from the bond-rating industry. Perhaps the industry can be persuaded to license its various forms of content recognition algorithms to assist in determining good links from spam.
Keep it Private
If it works, private industry will soon offer alternative certification sites, and black-hatters will be faced with a new challenge to overcome. If it fails, OK, at least it was tried, and whatever costs involved can easily be absorbed by the likes of Google, Microsoft and Yahoo!.
A few caveats: leave the government out of it, keep the price tag for a link certification low, and agree beforehand on all the rules for judging the appropriateness of a link. For those who think it’s too expensive, ask just how much JCPenny’s competitors would have paid to expose the nefarious plot before they lost online-driven sales to JCP.
The demand for a new link evaluation paradigm is there. Anyone want to fund it?Click here for reuse options!
Copyright 2011 Eric Bank, Freelance Writer