Kevin Rose has a good rant on why Google needs Buzz…
In a nutshell, his premise is that google knows they missed the boat on real-time search and is going to do what it takes to get this one right…
Kevin Rose has a good rant on why Google needs Buzz…
In a nutshell, his premise is that google knows they missed the boat on real-time search and is going to do what it takes to get this one right…
A few weeks ago I moved the Rain City Guide servers over to new servers. Starting a few days before I moved the site over, I noticed a drop in the amount of traffic that coming to RCG as measured by Google Analytics:
The fact that it started a few days beforehand didn’t really surprise me since I had been having trouble with the host (hence the move!).
However, the huge drop that Google Analytics shows is almost entirely from a drop in search engine (think: Google) traffic: 
It’s not pretty, but I figured it was a reality and just part of the consequences of moving the site from one server to another and I figured Google would hopefully start sending lots search traffic again over time. But here’s where it gets weird.
The ONLY indication I have that our search traffic has decreased is Google Analytics… All other indications are that the traffic is just about the same.
For example, Quantcast total daily visits from the same time frame (which, just like Google Analytics, uses a bit of javascript I installed in the footer of the site):
Diving deeper into the search terms I can see that for terms we used to rank really well for, like [moving to Seattle], Google Analytics said we dropped to only one person visiting us from that search term yesterday:
But, I simply can’t believe that for two reasons:
For those wondering, I didn’t touch the Google Analytics code when I moved servers. Nor did I install any caching or other plugins that would be related to this in any way that I can think of…
And finally, just to show you that the issue really is only related to search engine traffic, both the direct traffic and referring traffic have remained pretty consistent throughout this time period:
So finally…
My questions for the 4realz community:
UPDATE:
After only 14 hours or so of a fix (thanks to an idea I had inspired by a comment below), the search engine traffic on RCG as measured by GA has had a nice bump: 
(Notice the jump at the right of the chart… That’s today’s traffic from search engines)
At least for one of the WordPress blogs that I run, all evidence points to the fact that the existing tracking code went bust at some point in mid-May! Obviously, if you’re also noticing that GA is measuring a drop in search engine traffic, as some others have mentioned in the comments, then try upgarding to the latest javascript code.
Weird. Something clearly changed about May 24. But if SERPs didn’t change, it’s hard to believe people’s surfing patterns fundamentally shifted (all on the same day).
My GAs seem fine. Why it would affect just RCG is beyond me. Sounds as if something somewhere is corrupted. What and where is beyond my capability!
I poked through a few accounts to see if I found similar results and I didn’t. So off the top of my head I don’t have a clear answer. But here are some factors, perhaps:
1) The Google updated its algorithm around the time your issue begins.
2) Potential issues with your old server may have caused Google to re-rank your site at various data centers that you don’t see from your location. (Unlikely since you seem to rank well still).
3) Search interest in the keywords and phrases that worked for you in the past suddenly dropped off (again, unlikely but it’s a possibility… maybe no one cares about Seattle real estate blog).
4) GA servers were all screwed up at the end of May making a hash of a lot of data for a week or so, perhaps your issue is related.
Have you tried installing another tool just to double-check the GA results?
I also put a call out on the #WA twitter hashtag, I’ll let you know if I hear anything.
Jay and g.dewald,
Playing around on GA a bit more, I noticed that Google apparently updated their tracking code at some point from “urchan.js” to “ga.js”. I don’t think it *should* make a difference, but I just updated off of the legacy tracking code to find out.
urchin.js and ga.js are two different code sets. The new one has lots of features (event tracking being the most exciting) but it also could use a little more baking.
I currently don’t advise people to switch to the “new ga.js” code yet. But it probably won’t hurt too much. I don’t think it’s related to your issue.
thanks..
I noticed something around the same dates, too, also a traffic drop, though I couldn’t point it to search engine traffic for sure (we’re just starting and we don’t get enough traffic to detect that kind of pattern yet). We do pop up as a first result with relevant terms on Yahoo, though, but not Google. I’ve noticed some traffic coming from google recently, but the formerly usual Yahoo referrals have all but evaporated.
We’ve always been using the newer script and we haven’t made any other changes. I thought it was just a drop due to actual content, but hearing a similar story makes me wonder about the accuracy of the tool itself. I’m only using that and a server-side Webalizer account, but maybe I should install additional trackers in light of this issue…
You got me curious, so I poked around in my Google Analytics stats for a while and was unable to see anything similar.
On the other hand, I poked around in your Quantcast statistics a little bit as well, and found some charts that seem to indicate a fairly similar trend to what Analytics is showing.
They’re showing a decline 5/15-5/31, followed by something of a leveling off since then. I am inclined to think that the issue has something to do with your Google ranking, perhaps on some search terms other than the obvious ones that come to mind?
My $0.02.
Interestingly, the search engine traffic that GA recognizes is WAY up today over the past few weeks. I only started using the new script code at 10am and despite the fact that it’s not yet 8pm, the total search traffic for the day is almost double the usual search engine traffic.
I’ll give it another day to confirm things, but I think the problem has been solved and that the old tracking code was missing lots of search traffic.
I see a similar drop from May 24. Normally 300 visitors daily sice then reduced to 50. Next to GA I also use GetClicky, here business as usual. So it must be something in GA.
I noticed something similar happening on my site and it’s freaking me out.
Organic traffic from Google was about 70% of my total traffic up until late May. Since then it is more like 30% and it is actually DROPPING.
My Page rank has remained unchanged. Was 3/10 and still is 3/10.
Now, I did go on vacation around that time and stopped posting for about 3 weeks, so that could have something to do with it, but I only ever post about once a week anyway so I wouldn’t think a 3 week hiatus would result in a drop like that.
In my case it seemed to fall off the cliff around the last couple days of May.
Consistent across Google Analytics and Sitemeter.
I guess I’ll try upgrading to the latest GA code first and see what happens.
Erno and Andrew: The upgraded code solved the problem for me! My traffic according to GA has been back to normal levels ever since the change!
Upgraded GA code has not changed anything for me in the 24 hours since. still not getting ANY google traffic and my page rank of 3 is just puzzling me. how can i still be a 3 and not be getting ANY google traffic. is google punishing me for something?
I found your blog when searching “google analytics shows a drop in keyword search” because I’ve been trying to figure out what has happened to my site that totally freake me out this past week. I have been raking pretty steady and in Google’s top 10 for over 2 years for a few select money making keywords. One day I did a check and I had fallen to page 6 for two of them, and even further down for a few more. I had been updating my site with a new logo so for a couple of weeks I had been working steady on it. My first thought was I must had messed something up and didn’t realize it. In looking at my G-analytics account I could see that for certain keywords the search for all of them happened around June 19-21. This was about the time I started uploading changed pages on my site. However I didn’t see the big bottoming out until July7th. All the top keywords in G-analytics showed the same thing a big decline (and I dont’ mean just a swoop down I’m talking about a straight line to the bottom and then flat line…like I had died for two weeks).
I’ve been frantically trying to retrace my steps ..and tonight in looking at my GA account I couldn’t help but think it has something specific that happened during those 3 days.
Tonight I just happen to compare my home page (which has been updated) to an internal page that has not been updated and is ranking very high. After reading the above post in you blog..and checking to see which code I’m using ..the old or new… I realized my updated pages don’t have ANY analytics code in them. But my older pages do! And, my older pages are still ranking ok.
I think I must have accidentally removed the GA tracking code completely…so I’m fixing to go put it back right now and I’ll come back and update here to let you know if it had an impact.
Donna: Accidentally removing the GA tracking code would definitely cause such a problem!
Dustin,
may I have a backlink or a blogroll?
Oh, I already have a blogroll. Thanks so much.
It’s hard not to like Matt Cutts (pertianing to the linked article) he knows how to limit the things to say. I think that’s smart his five steps may sound simple but that alone can be overwhelming for people starting out with SEO.
Reno: No doubt keeping it simple is key. That’s why I love that Matt can be quoted as saying “linking is where it is at”. And if ranking well in Google is your goal, then having an strategy to build quality inbound links is key.
Galen makes his first move since his new found riches… Congrats to the Estately team.
All the talk about Yahoo, Microsoft, and Google reminded me that I hadn’t checked to see if Microsoft has made any improvement to their search results in quite a while…
I know one term doesn’t make a rule, but I’ve found the term [Seattle Real Estate Blog] to be great for testing out the search engines over the years. Over the past three years, Google has always delivered relevant results for the term, while it used to be full of spam for both Yahoo and Microsoft. It’s interesting to note that Yahoo has made great strides over the past few years, while Microsoft is still lost in the dark.
Here’s the top five results by search engine:
Google [Seattle Real Estate Blog]:
All of these are great blogs and worthy of being in the top five.
Yahoo [Seattle Real Estate Blog]:
Again, all of these are great blogs worthy of being in the top five for the search results.
MSN Live [Seattle Real Estate Blog]:
There results are horrible. Two spam sites, two sites with no content, and one page with a few links. There are some great real estate blogs in Seattle, but Microsoft can’t seem to find them as their search algorithms are still so easily gamed by spammers.
Scoble thinks Microsoft should buy up Facebook and keep it closed to all but their search engine. Personally, I think they’d find a way to screw that up.
UPDATE: NYTs says Microsoft is getting ready to PAY people to use their search. How sad.
What we’re seeing with the state of microsoft search is the result of a 5 year plus project costing billions and the product still doesn’t work, period.l…. hence the microsoft / yahoo deal that has been up and down as of late…. microsoft has yahoo pretty much pinned at this point so it is likely they will be working something out…. we will see soon….
You think that’s bad? Check the search results for “atlanta reale state” —>
http://search.msn.com/results.aspx?q=atlanta+real+estate&FORM=MSNH
The 50 (read fifty!) pages are subdomains of the same website. That is 500 search results from the same domain for the most trafficked term for Atlanta real estate”
Ryan: Wow! There’s almost something comical about that!
Ryan, I see your website at the top of the list and a bunch of other websites all on different domains. Not really sure I understand what you mean?
I think Microsoft finally realized this and doggedly pursued Yahoo.
Though I still have my doubts on how they are going to pull it off and do some real competing against google.
…although if you add quotes around the search term,(“Seattle Real Estate Blogs”) I think you’ll see that you get different results on all three. Hmm…
Brian,
Must be a datacenter issue. I posted about this a few months ago on my blog with a screenshot and had people check all over the country. Most people will see subdomain.atlantarealestate.net for the first 498 results!
But hey! If you see me I must be doing something right
somewhere….
BTW Brian, where do you see me? I’m like 700 here in Atlanta. Thank God for Google…
For real estate blog searches, Live Search is certainly worse than Google currently is. However, for other topics, it’s usually much better. Oddly enough, a few months ago I ran a search on “how do CDNs work” in which MSN & Yahoo gave me relevant results on Content Delivery Networks and Google’s results were so wrong (Canadian Viagra anyone?), I laughed for 15 minutes. No search engine gets it right 100% of the time, even Google. (Frankly, Yahoo has been giving me better results than Microsoft or Google lately).
For real estate searches, I think Live Search weighs the domain name too heavily in rankings and generates poor results. For searches on technical topics (networking / software) and product / review information (bluray players / mattresses), I’ve been seeing Google’s relevancy lead over Yahoo & Microsoft shrink during the past few months.
Still, I don’t see Microsoft beating Google any time soon. However, I don’t think Microsoft needs to beat Google, they just need to be a competitive #2 and
force Google to play defense.
Ryan, it could be that I am located in South America. I know results vary depending on IP addresses. After I posted I thought that could be it. Anyways, I believe u. I was just confused when I checked the results.
I did a search on the term “short sale” and google carried the blog post I did last March on the first page. It doesn’t show up on the first page of either yahoo or MSN, and the yahoo and msn results looked very similar.
I like how blog posts are given some love on google.
Robbie: I was hoping you’d come around these parts!
I’ll just have to believe you that it does better in some areas, because it definitely misses the boat on real estate related content.
And Jillayne: That Short Sale article you wrote is still one of the highest traffic articles on RCG on any given month. It’s huge and so much fun that Google loves it!
Sad but true… Microsoft just doesn’t seem to understand that they need to make the results better (and faster) than Google’s if they want to compete. Giving people money is just a short sighted way to spend their hoards of cash. At this point, they can’t even get their own staff to use their search engine consistently from what I hear.
Long-time Rain City Guide contributor, Galen Ward, announced that Estately just got $450K in angel funding! Congrats Galen!
The Estately team is used to operating on a shoe-string budget, so it will be interesting to see what they accomplishes with this money!
Christian Sterner 12:06 pm on August 22, 2008 Permalink |
Hah…Boer stole the show in the comments. Hilarious!
Drew Izzo 12:26 pm on August 22, 2008 Permalink |
Thanks for the post Dustin. It’s definitely the beginning of an interesting and ongoing conversation.
Dustin 12:40 pm on August 22, 2008 Permalink |
No doubt about it Drew… You guys are definitely doing something interesting by tapping directly into the MLS system.