Matt Loney's Blog
What it says on the box
Thursday 1 February 2007, 6:27 PM
Why Google is (mostly) right about tags and labels
US organisation Pew Internet and American Life Project released a report this week saying that 28 percent of internet users have tagged or categorized ontent oline such as photos, news storis of blog posts. The report notes that some sites ask people to apply labels. And in the past weeks Google has leant its weight behind the use of 'labels', with the launch of its latest Toolbar which, as you can see in the image below, asks you to add labels, not tags, to describe your bookmarks.
Here's why Google is right.
People are very good at recognising patterns, but when we try to create patterns to define what we see, we're not so good. The reason is that while pattern matching is an automatic, cognitive process, pattern creation requires conscious thought; it gets mixed up with our feelings, our experiences and our knowledge, all of which are particular to the person creating the pattern. HSBC's ads in airports around the world are a great example; one person's holiday is another person's hell, even though it may actually just be a beach.
Computers on the other hand can be very good at looking at content and creating patterns; if you want a clean, precise and untainted definition of something, a computer is your best bet. If you've read my post below you may already see where I'm going with this.
All the Web 2.0 sites that ask people to tag content are falling behind the times. Tagging has been an acceptable term for this human process, but there seems to be a shift in thinking, and not before time. After all, we're used to labelling things. Labelling is a human process; when we label someting we are applying a word to something that describes it in a way that is useful to us.
Tagging is, or should be, a computer process. At ZDNet UK we post at least 40 articles a day - often many more - and we apply labels as we do so, to categorise them. This helps define where an article sits in the navigation, it helps define what treatment the article should get, and it helps define what ads should be served alongisde it. We also apply tags to the article, and this is done automatically by our search engine using the process described in the post below; it is those tags, not our labels, that are used to populate the tag clouds on ZDNet.
For us, tags create a more consistent tag cloud than labels would.
Oh, and why is Google only mostly right? Because Google Docs is still in tag land.
Here's why Google is right.
People are very good at recognising patterns, but when we try to create patterns to define what we see, we're not so good. The reason is that while pattern matching is an automatic, cognitive process, pattern creation requires conscious thought; it gets mixed up with our feelings, our experiences and our knowledge, all of which are particular to the person creating the pattern. HSBC's ads in airports around the world are a great example; one person's holiday is another person's hell, even though it may actually just be a beach.
Computers on the other hand can be very good at looking at content and creating patterns; if you want a clean, precise and untainted definition of something, a computer is your best bet. If you've read my post below you may already see where I'm going with this.
All the Web 2.0 sites that ask people to tag content are falling behind the times. Tagging has been an acceptable term for this human process, but there seems to be a shift in thinking, and not before time. After all, we're used to labelling things. Labelling is a human process; when we label someting we are applying a word to something that describes it in a way that is useful to us.
Tagging is, or should be, a computer process. At ZDNet UK we post at least 40 articles a day - often many more - and we apply labels as we do so, to categorise them. This helps define where an article sits in the navigation, it helps define what treatment the article should get, and it helps define what ads should be served alongisde it. We also apply tags to the article, and this is done automatically by our search engine using the process described in the post below; it is those tags, not our labels, that are used to populate the tag clouds on ZDNet.
For us, tags create a more consistent tag cloud than labels would.
Oh, and why is Google only mostly right? Because Google Docs is still in tag land.
Previous
