The Sandbox and Delicious
January 4th, 2006 by Michael Gray in Case Study, SEOIf you're new here, you may want to subscribe to my RSS feed. Read my top posts or learn more about Michael Gray. Want more frequent updates follow me on Twitter. Thanks for visiting!
Over on the Link Building Blog Andy’s talking about how the blog is still sandboxed. Not that Andy needs my help, but I’d like to take this opportunity to prove the value social bookmarking can play in getting your website out of the sandbox. I’ll be dissecting my own site and Rand of SEOMoz has been kind enough to allow me to pick apart his traffic a little too. Get out the pencils and paper here we go …
First lets take a look at my traffic for most of 2005
I’m not sure why the data from January through March is missing but basically my page views (cyan) and visits (yellow) have tripled from April to December. I’d say that graph is pretty typical of a standard growth rate, some ups and downs but overall and upward trend.1
Now let’s take a look at my feedburner subscription stats
Again beats me what was going on in 2004 and why there is no data before hand, but looking at March through December the slow steady growth I talked about is reinforced.
Lastly we are going to look at the Alexa traffic graph for the website. Now I know Alexa is not the most accurate piece of data out there, and can be manipulated if you are motivated enough. However using the Alexa data in combination with the above graphs a few key points stick out.
Ok so no activity to speak of before April 2005 on all three graphs, I know I was blogging so I guess no one was reading (sob). The Alexa graph is much more spiky than the two previous graphs. In July and August we see two big “pimples”, two smaller “pimples” in October and then a big spike up in December.
In late July I published my Adsense Series (Google Adsense: Tips, Tricks, and Secrets) followed by a whole series of articles on building better blogs and content based websites. Now my Adsense series made the Delicious Popular Page, brought in huge amounts of traffic and links. In fact people are still adding it to Delicious. In October I got residual traffic from the SEO ranking paper I contributed to (more about that later), and made a few snarky Google posts. Now other than things like Google Bowling, Bacon Polenta, Eco-Friendly Flip Flops, and the occasional odd key phrase here and there I don’t do a lot of targeting here (not to say you won’t see some camouflaged tests now and again). But I can tell you Google referrals picked up in November, and I believe this site came out of the Sandbox at that point. Now someone is going to point out OK so why did your traffic drop in November if you just came out of the sandbox? First I went to pubcon in Las Vegas and then was blogging light when I went to Disney the last week of November. So while I wasn’t posting as much and people weren’t visiting traffic dropped, but Google referrals made up for most of the lost traffic. Once I got back to posting regularly the impact of coming out of the sandbox was completley evident shown by the large traffic jump in December.
SEOMoz
As I said in the begining of this post Rand was nice enought to let me poke at his traffic numbers at bit so here goes. Back in July Rand mentioned how his website was in the sandbox. From his sponsors page lets take a look at his traffic:
February 2005 - 7,190 unique visitors 97,386 page views
March 2005 - 7,357 unique visitors 139,020 page views
April 2005 - 8,274 unique visitors | 228,993 page views
May 2005 -11,050 unique visitors | 214,207 page views
June 2005 - 9,276 unique visitors | 104,978 page views
July 2005 - 6,792 unique visitors | 108,082 page views
August 2005 - 8,950 unique visitors | 219,334 page views
September 2005 - 15,062 unique visitors | 229,935 page views
October 2005 - 37,493 unique visitors | 292,779 page views
November 2005 - 25,758 unique visitors | 260,376 page views
December 1-15 2005 - 57,803 unique visitors | 252,119 page views
While the numbers are completely different there is still a general upward trend. Lets pull in the Alexa Graph
Two really big spikes stick out, the first in October came from the search ranking research paper he compiled. The second came from his write up and being featured in Newsweek Magazine. Now stepping back in time I know that the Search Engine Ranking Report made the Delicious Popular page and got lots and lots of links. Then come November his site came out of the sandbox.
Correction
Rand corrected me, in December the spike was caused by the The SEO Beginners Guide which was tagged by delicious users 350 times so far.
My math teacher Sister Miriam told us we need three points to draw a trend line, and I only have two, so I’d love to hear from someone else who has a similar experience. Two websites is also a very small data set to draw sweeping conclusions from, so again I welcome people sharing some research.
Is getting links and getting tagged in del.icio.us a sure fire way to get out of the sandbox? Nope don’t think so. However getting links from lots of other websites, on different IP’s, with lots of varying surrounding and anchor text, is going to look a lot more natural than links from SEO driven websites ever could ever dream of, something you may want to think about.
Footnotes
1. The dark blue and green bars represent files and hits. Hits are an almost useless metric I wish would disappear. The file number is wonky to because I was using includes. So the cyan bars which represent page views and yellow which represents visitors all show an upward trend over time.
Sphere It
January 6th, 2006 at 9:16 am
If you’re happy to take my word (not my stats) then I can add quite a few sites that will allow you to draw your trend graph
January 6th, 2006 at 1:53 pm
Michael - quick correction, in December, the traffic spike came from the Beginner’s Guide article (getting Slashdotted, etc.). I didn’t see a noticable traffic rise from Newsweek, and it didn’t provide many links (most folks linked only to the MSNC site and not to SEOmoz.org)
January 6th, 2006 at 3:08 pm
I don’t think these stats show much.
IMO You need to isolate all the traffic you got from google into a seperate log and then run a stats package on it. I know for me that I am still getting more traffic from direct links at the moment but you would never know that just by looing at my stats.
January 6th, 2006 at 3:19 pm
Also I’d be more than happy to do what I suggested and send you the graphs of the analyzed data. My blog hasn’t been around long though so it may not be useful for you yet.
January 6th, 2006 at 4:09 pm
Hmmm.. I’d be interested in the referrers also, especially the non-google ones. You can’t track the backlinks reliably otherwise, but referrers might provide some insite into the monitored traffic (Alexa, toolbar) giving you rep.
Of course there is also time frame. Many, many sites that reported “sandboxing” last spring came out around the same time. It’s fun to see causality, but perhaps it is circumstance?
January 7th, 2006 at 5:07 am
It’s interesting, but I’m not seeing the overall spike that happens when sandboxing suddenly finishes and a site starts ranking for bigger keywords. IMO, that’s what is really indicative of sandboxing. Really, your stats would actually look more like sandboxing if reversed - you’ve actually lost out on unique visitors over time, which doesn’t suggest to myself a sudden release from sandboxing.
In terms of stats - I once posted stats I thought showed sandboxing in action, demonstrating how a website could not rank for it’s targeting keywords via links for 3 months, then suddenly impacted.
That was from the early days of sandboxing in 2004, though, and the process is obivously more complex now.
I think rather than being a specific gatekeeper on anchor text, as sandboxing was originally reported to be, “sandboxing” seems to have become a generate euphemism for a range of “anti-spam” features.
It’s hard to make clear statements on what constitutes sandboxing now, but last night I found it interesting to note that - although http://www.linkcondom.com/ has PR5 and is indexed by Google, it doesn’t rank in Google for “link condom”. Sandboxing in action, or a penalty?
Either way, sandboxing seems to become an overall concept that becomes more confusing to determine specifics for.
2c.
January 10th, 2006 at 1:10 am
I agree it’s a fuzzy concept, and the link condom example is strange. Could be an SEO Community link penalty.
January 10th, 2006 at 7:16 am
[...] It’s getting easier to accomplish the things that were considered SEO in the past like getting dynamic URL’s spidered, but the demand areas are just shifting. Rather than needing LOTS of links, people now need the RIGHT links. They also need viral marketing which I think will be a large part of SEO in the future. Gray Wolf commented recently on the correlations between del.icio.us and the sandbox, and I highly doubt that this is just a coincidence. The engines will definitely be implementing user data to validate the integrity of the algorithmic results. If you have all the right links, but no click through and very few users, you are not going to have rankings. The emphasis on what is priority to an SEO campaign will be shifting as nearly quickly as the engines change how sites are ranked. This is one of the things that make scaling the growth of an SEO company difficult. Employees must be prepared for a dynamic shift in job responsibility to continuously be effective for clients. [...]
June 3rd, 2007 at 10:31 am
No Doubt link baiting can provide us alot of links in no time but shouldnt we have to do it daily.
As on digg.com every second home page stories are shifted down.
Whats future of this social bookmarking thing ?
Cant we see 1000 of sites like digg.com - like i just saw one of indian guys site http://www.Tagza.com
Where this storm will end by the way ?