Google’s Algorithm Biased Towards New Web Pages?
Well, to some extent we already knew this (read Could Indexation Time Affect SEO?).
The Google Operating System blog, however, is reporting an extrapolation of this pattern. According to the article Google might be artificially promoting recent web pages to the point of out placing trusted and established websites like Wikipedia.
They gave the example of the term “January 1 tcp/ip.” Yesterday Google changed its logo to celebrate the anniversary of the famous Internet protocol, resulting in a sudden increase on the number of related search queries.
If you performed a search for this term, though, on the first page you would find only recently created web pages like Digg submissions and Blogspot blogs that were trying to leverage this traffic.
I checked again today and results are still dominated by recently created material. If nothing else this is an interesting trend to watch. The question that emerges is: “for how long will the results keep this bias?”
It is also a good idea to keep an eye on search queries that become popular overnight. If you move fast using the right URL and keywords you could outrank Wikipedia!
Update: Patrick from Blogstorm has a possible explanation for the whole issue.
21 Responses to “Google’s Algorithm Biased Towards New Web Pages?”
I can say that back in November I wrote a post titled â€œWhatâ€™s wrong here?â€ that almost instantly went to the top of search rankings in Google. Traffic from that post was 5-10x my normal traffic for all posts, if I recall correctly.
I’m trying to cut down the amount of work I have to do to promote my site.
How about a post, which reads “Google doubles pagerank for old sites with few new pages”.
I live in hope.
Yes blogs get an extra boost in search engines for a couple of days. But then drop out untill google starts figuring out what to do with them. The same happened with my blog
So interesting!.. digg is still #1â€¦
I agree with Paul Saunders
Established websites with lots of backlinks do indeed do well in the rankings, especially if they update their content regularly. I believe that Google also take account of how often people click on links to sites that appear in the search results.
It therefore follows that old, dead sites that don’t update and don’t get clicked on, will die out.
If Google algoritm work for established websites with millions backlinks their results will look like archieve of 1995-1997 websites.
So interesting!.. digg is still #1…
I completely disagree with Patrick’s comment that “itâ€™s only biased towards new pages for â€œhotâ€ queries.”
My own personal tests and results have matched those that Paul describes in detail above. As with Paul’s testing, my own testing has been with queries that are certainly not what I’d describe as ‘hot.’
Have you actually done any testing on this with ‘non-hot’ queries Patrick, or is your conclusion based on that single TCP/IP example and the fact that Wikipedia didn’t have a dedicated page for it?
I’m surprised there isn’t more written about this and that more people haven’t picked up on it. I’d be very curious to see if this could be manipulated by blogs to appear highly during the time that a fairly small scale event was taking place (i.e. not the Superbowl or Olympics as that will be covered by the likes of digg, but a smaller local event).
I second your opinion. I’ve a news portal website. Only the news in the front page get listed by google. As soon as the news leaves the front page, it get very bad position in google search result and therefore very few traffics.
Posting again to get emailed
Nothing new here, move along. Has been happening for over a year now.
This is very interesting… I think Partick’s explanation would fit this situation very well. I’ll have to look into this.
I think that new blog posts are floated to the top of the results for a few days and then judged on the basis of relevancy by monitoring whether or not people actually click through.
I have to disagree with Patrick, it’s not just hot queries, it works for common phrases too. I noticed this behaviour when I started a blog 7 months ago, and although I searched for information about this at the time, I couldn’t find any.
To begin with, I started my blog with a post called “Welcome to the Weblog”, hardly an uncommon phrase. Within it month it ranked 3rd in Google out of nearly 15 million results! Sometime later (I’m not sure how long) it disappeared without a trace.
The same thing happened with “My First Mountain”, another common phrase. After 14 days it was nowhere to be seen, but after 17 days it ranked 16th out of 29 million. It also subsequently disappeared.
On 26th May I posted “Worms Head Sunset”. On 19th June it ranked 30th and on 22nd June it ranked 1st (out of 580,000 results). Needless to say it also disappeared after a while, but here’s where it gets interesting. I checked it again today and I’ve discovered that it’s back in 1st place, 7 months later!
This suggests a trend, that blog posts are artifically boosted in importance soon after posting, then they lose that bonus, then later they age and increase in importance once more.
However, most of my other posts seem to have disappeared permanently into the void. So why is “Worms Head Sunset” ranking so highly? As far as I’m aware it doesn’t have a single external link, and it isn’t even very popular. It’s only had 47 hits since it was first posted. It presume it must be due to SEO alone.
Another thing I’ve noticed is that the number of results is significantly less now, only 44,000 instead of 580,000. I’ve noticed the same thing with the others I checked, the total results are down with all of them. Are Google getting stricter and eliminating more results?
Which brings me to normal web pages. I have a number of high ranking pages on my website, and they get a decent number of hits, but my blog pages were a disaster in comparison. This has led me to believe that web and blog pages are ranked differently.
Web pages seem to benefit from ageing, while blog posts seem to benefit from “freshness” and thus lose importance over time, unless they get a lot of links perhaps? This doesn’t explain the high ranking of my Worms Head page though.
I’ve suspected for a while that my blog posts would do better as normal web pages, so now I’m about to test that theory. I discontinued my blog a few days ago and I’m going to convert most of the content to normal pages instead. It’ll be interesting to see if they do become more popular than they were on my blog, and if so, how long it will take for that to happen.
i am the owner for byrev.net (listed up) ,
This result is “overnight story” from google trends.
– 1st, title like in google trends,
– 2nd, i ussed digg for promote (not imediatley)
– 3rd, some backlink (not to much, 2-3)…
this efect is verry rare ~ 1% “g trends inspired” articles.
last trafic spike (nov 28, 2007) was with “zoey zane” … My server has shut down unexpectedly 2h ( myqsl limits, cpu overloaded – flood from google search )
and btw: google is very friendly with blogspot 🙁 !!!
I can say that back in November I wrote a post titled “What’s wrong here?” that almost instantly went to the top of search rankings in Google. Traffic from that post was 5-10x my normal traffic for all posts, if I recall correctly.
The weird part is that the post is still ranked number 5, but traffic has fallen off quite a bit. I know the time of year is a factor, but even disregarding that it seems to have fallen off a lot more than other posts.
There may be something to it. I started my blog not too long ago and was surprised to notice that google started sending me traffic just a few weeks after I started it. Not a lot of traffic but consistent traffic. Thanks google; you are a great search engine! 🙂
Interesting! Thanks for noticing. Definetly worth keeping an eye on this topic.
There were many ad hoc blogspot blogs there also, who probably had been created the day before :).
I think your explanation is feasible, but Google should pay attention to the issue nonetheless.
It’s only biased towards new pages for “hot” queries. In this case Google unleashed millions of searchers on a query that had no previous history meaning the normal spam and domain age filters were turned off to give people topical results, just as you would expect.
In this case Wikipedia didn’t have a page to match so was pushed well down the results. Digg, another trusted domain, did have a page to match so it was at the top.
Comments are closed.