Archive for September 27th, 2011

27
Sep
11

What Page Deserves #1 Ranking?

One of these things is not like the others

One of these things is not like the others

This has been one of our hobby horses for a long, long time. Does a page deserve to rank number 1 in a search engine for a particular keyword search because it has more links pointing to it? Or because it uses the exact keyword phrase in the title and <h1> tags? Or because the site it belongs to is big and old? Or because a horde of pixies has conspired together to deluge the page with +1s? (We’ll avoid the obvious conclusion that the most deserving page is the one we did the SEO on.)

Or does a page deserve to rank #1 for any given keyword phrase because it is a better answer to the query?

That’s the answer we desperately want to believe. That’s the answer that promotes quality content over all else. And makes the internet (and thus, the world) a better, happier, more useful, and more interesting place.

So, then, what happens when someone searches for some very specific topic—let’s say a product called WOMENS HARLEY DAVIDSON JERRI STILETTO SANDALS—and every one of 50 pages in the search result is the exact same catalog page: product photo, manufacturer’s description, price & shipping info, and a big fat glowing “buy now” button? Which one of those do you rank #1? What if every one of them has the same number of ++++++ (zero) and the same number of links (zero) and the same number of Facebook friends (zero)?

To the searcher, it probably doesn’t matter who’s on first. If they’re looking to buy WOMENS HARLEY DAVIDSON JERRI STILETTO SANDALS and they don’t look at a whole bunch of those pages, they’re very poor internet shoppers. And if they’re just doing research, well, it doesn’t matter which page they hit if the content is all the same.

To the merchant, though, it does matter—matters a lot. Because whoever’s on top gets more visits, first impressions, and better opportunity to close a sale.

What’s a search engine to do?

If you’re Google, you’ll rely heavily on domain size, domain age, and number of inbound links, +1s, likes, and whatever that  point to the domain as a whole, even if none point to the page in question. If you’re Yahoo, you look to domain links and maybe Alexis traffic data. If you’re Bing, you pull out your 20-sided fuzzy dice.

If you’re us, you’d treat merchant sites differently from all the other kinds of sites. You’d use a less-focused algorithm that says “On merchant sites, if the content is virtually the same, the page ranks are equal, despite any other of the usual ranking criteria.” And then we’d let all the identical product-description pages rotate through the ranks, randomly, evenly, equally.

And let the consumers sort ’em out.