December 19th, 2011

Twitter: Big Sites Fail Big

No, I am not talking about the infamous Fail Whale. And moreover, the big news is that it looks like the biggest fail of the last 5 years has just been fixed – but not due to Twitter’s efforts.

Twitter is huge, this is not news to anyone. A site:twitter.com search in Google returns 1,750,000,000 results. Yup, that’s close to 2 BILLION. Yet, most of those results are actually non-existent pages.

Yes, you heard me right. Google keeps in its index close to 2 billion non-existent pages from one domain alone. How come? Let’s look at the typical URL structure of a Twitter user profile:

http://twitter.com/#!/username

Now, what kind of URLs do we see in the aforementioned SERPs for the site:twitter.com query in Google? Something like:

http://twitter.com/username

Notice the difference? the “/#!” part is missing. Infact, it is not even possible to figure out if Google has at least a single URL with the “/#!” bit indexed as these symbols would get ignored by Google so searching for site:twitter.com inurl:/#!/ just won’t produce any results different from site:twitter.com.

Where did the whole issue come from? Some of you may remember that the new URL structure for user profiles came to exist over a year ago – for some time afterwards, it was still possible to switch back to the old (less-Ajaxy) interface preserved under the old URL. Then, the old interface was killed and all old URLs were redirected to the new ones.

Only, Twitter has never got the redirects right. They use 302 instead of 301! Here is a 2007 blog post by Google’s John Mueller detailing what Twitter’s got wrong and how it should be fixed. Did they ever fix their redirects? – No! Do they think they are too good for SEO? Heck, even CNN has an SEO, and did you ever think CNN should care much about search engines?

Until recently, this profile URL redirect issue used to cause some serious troubles with cached versions of the corresponding pages in Google – all of them appeared as “that page does not exist”:Twitter 302 Redirects Cache Screenshot

Lately, however, Google got better at indexing and caching their 302 redirects so the cache screenshots look better. But it was due to Google’s own action only, not Twitter’s. Are the folks at Twitter THAT blind and deaf?

November 2nd, 2011

Google Products Updating – Usability Fails

I must confess: I am pretty conservative. Not to the point of totally not accepting any changes – but if there is no real pressing need to change anything, I won’t be looking to change it. Moreover, you can guess I am not typically particularly happy when a change affects something that was well usable in a negative manner.

With a rush of Google’s product updates recently, I keep wondering what’s going on – have they fired and completely replaced their internal staff dealing with interface designs and who are those people working on these things now? Are they clueless newbies, do they even test what they create? Has Google decided to cut down usability testing expenses? Because a lot of what I see looks like a complete disaster.

Take just two products, for example. Those I use quite a lot in my daily work.

First, there is Google Analytics. OK, you sell me on getting access to real time data – but considering the new interface it is just not worth the move for me. I wonder who the typical Google Analytics user is, how many sites/accounts do they have under their login? I of course might be more of an exception with my dozens and dozens, but I’d suspect typically people would have at least 3-4 sites there.

Now, when I check my Analytics I like to have a quick glance first at what’s going on with the traffic and its dynamics and then dig deeper to identify any possible issues, and with the old interface it is easily doable:

But compare that to this sad shit of an interface in the new Analytics:

Takes me 3 clicks just to get to that same data for each single account, nevermind getting the data for all my sites! And they don’t make it easy switching back to the first screen, either – took me quite a while to figure out how to do it (for those still wondering, click on the dropdown in the upper left corner with the list of all accounts and click the button appearing there that says “Accounts list”).

OK with Analytics I am not a typical user but what about Google Docs? I only really use them when I need to share something with other people and have it updated in real time, otherwise I don’t think Google Docs will ever be able to replace the real Office applications for most of my document creation and editing needs – there’s just too much stuff missing. Anyways, I found out that the search shortcut (Command+F or Control+F, depending on the OS) doesn’t work in some browsers. Anotehr issue, again, is a less-than-straightforward way of switching back to the old interface. The only place where the link to do it appears prominently is the Google Docs home page, but when you are viewing a document you just don’t see how to do it. Here’s how: click Help and select “Use the classic look”.

I could go on and dissect the failures of the newly updated Reader but I will stop for now and maybe get back to it some other time. For now, let me just say: I’m sticking to the old GA and Docs for as long as they stay alive.

October 20th, 2011

Google’s Move to Hide Search Queries and SEO Feedback

Google has announced its switch to SSL by default for search for logged in users and it immediately became a big topic of discussion and even outrage for the SEO community. Here’s a detailed explanation of what this move means by Danny Sullivan. Here are some SEO reactions across the web to these news. Main outrage is caused by the fact that Adwords advertisers will still have access to the referring queries data for their paid campaigns, which is seen as obviously a double standards implementation by Google and an attempt to switch focus from SEO to PPC.

What surprises me most in this whole discussions, however, is that people somehow fail to see the other side of this whole issue. I was looking if anyone would mention this but nobody did so far – don’t you think what Google is doing is actually an attempt to push more users into (1) creating Google accounts and (2) using Google’s services via those accounts? In other words, encouraging users to stay logged in (of course as a result Google gets to collect more and deeper data, regardless of SSL). Here’s something they say in their official announcement that can act as proof of the above intention (bolded by me):

we’re enhancing our default search experience for signed-in users

In other words, you want security and privacy – you can have it, you just have to log in. Seems like it’s their marketing move to improve the sign in rate which seems to be not so high, to quote Barry Schwartz:

Google says this will likely only impact less than 10% of the searchers. So in a sense, Google is saying that less than 10% of their users are signed in when performing Google searches?

In any case, I wonder if this feature will be as half assed in its implementation as the new Google Analytics and Google Docs interfaces (will post about it separately).