Blogger Hacks, Categories, Tips & Tricks

Wednesday, February 08, 2006
Retro-Fitting Tags to Your Blog
So you've been convinced of the benefits of tagging your blog posts and have your eye on increased visibility, navigation and integration. But now you've got to backfill dozens or even hundreds of posts with meaningful, traffic-driving tags. How do you craft a set of sensible tags that afford stability and growth? How do you stay sane while spending long hours weaving these tags into your posts? Read on for tips and tricks for retro-fitting tags to your blog.

Getting the right tags for your blog is important. For starters, it will affect how your blog appears to search engines and other services. Next, it will impact upon the way your readers navigate your blog to find other posts. The tags you choose - seen as a whole, say as a tagcloud - also tell the reader a lot about how the blog is positioned. With so much riding on it, getting a good set of tags is worth a moment's reflection. To tackle this problem, I've drawn upon my own experience with retro-fitting tags to two blogs, plus conversations with blog tagging luminaries like Johan, John and Singpolyma. I would love to hear from you (below) with your own stories and advice for getting on top of this.

Tags vs Categories

One of the common distinctions that comes up is whether you're using tags or categories ("tags-as-categories") for your posts. The distinction is that categories are fewer in number, generic, chosen beforehand, possibly hierarchical (sub-categories) and persist for a long time. Sometimes people will file a post under only one category. Tags are much more specific, made up on the spot, are "flat", may be single-use and each post may have half a dozen or more of them. Categories work well for navigation within your blog and integration between blogs (better chance of a match-up with a related blogger). But they can be a bit too generic for readers to get a sense of your blog. Also, if you want to pick up traffic from tag-searchers you may need to go beyond your generic "art" and "culture" and delve a little deeper.

In practice, the category/tag distinction is a question of degree and many authors use a hybrid strategy: tag their posts wildly but reserve ten or so to use as categories. This seems to be a best-of-both-worlds scenario, as your top tags (when sorted by frequency) function like categories while the long-tail of one-shot tags is useful for tag-searchers.

You can also create a traditional hierarchy by including the higher-level tags as well. For example, you could indicate that a post belongs in the "science/physics/theoretical" sub-category by using all of "science", "physics" and "theoretical" as the tags. That way, you can see all posts under science/physics using tag intersections.

Setting Up Your Tagspace

The first step is to recognise that you will want to use your tagging service account (at delicious, simpy or elsewhere) for more than just tagging your blog posts. For instance, you may wish to use it to store your bookmarks! Or for tagging posts in a second blog or guest blog. The easiest way to accommodate this is with an anchor tag for each blog. This means that you have to think up a simple one-word label for your blog and ensure that each and every post is tagged with this label. This will ensure that any subsequent blog hacks you employ will work on a cleanly separated list of tags/posts. It is very frustrating seeing random links or posts from other blogs appearing in your sidebar!

Next, you need to nail your list of "recurrent tags" (or categories, if you prefer). Doing this for a blog that doesn't yet exist is very difficult and amounts to little more than a blog's mission statement. (Sadly, all too often it has the same energising effect - and correspondence to reality - as a corporate mission statement.) This is where retro-fitting tags is a blessing: you already know what the blog is about, what topics are covered and where it's headed.

At this point, you might consider having "themed" tags. For example, You're It! blog uses "who, what, how, why". My letters-to-the-editor blog started out following traditional newspaper categories: politics, foreign affairs, education, media, IT, lifestyle etc. My footballer-scrutiny blog uses crimes: bashing, rape, drugs and so on. While this may be an okay starting point for brand new blogs, I'm a bit wary of using this approach for established ones. The temptation is to shoehorn your tags to fit the pre-conceived theme rather than putting the content first.

[A tip for those with very popular blogs: find out what tags your readers use to describe your blog. Not only will this give you insight into how the world sees your content, but will also provide ready-made "hooks" into the wider tagosphere.]

I recommend skimming your blog's titles if it's more than a few months old, then writing down a list of, say, six tags (as categories) plus a couple of "policies" for dealing with tags targeting the long-tail . For example, you may decide to tag all company names (or product names, or place names, or other bloggers etc) mentioned in each post. Beforehand, decide on a "format" for these tags. For company names, will you use the parent company name or the relevant division? Acronyms? Stock ticker names? You get the idea; consistency may be "the hobgoblin of small minds", but it also makes for a usable tagspace.

OK, once you've got your category-candidates and long-tail policies, trial it out on a few post. Just pick out a nice selection of posts - early/late, short/long, typical/unusual - and see how well it fits. While you won't get each tag being used an equal number of times, I'd suggest that you don't want more than a factor of ten difference between your most-common and least-common recurrent tag (excluding the anchor tag of course). You'll want to fine-tune your tags at this point, but keep in mind that fewer is better. This is because it is relatively easier to split tags than to merge them. Eg. All posts tagged "culture" become tagged with "art" + "media", and you can manually remove "art" (or "media") from relevant posts.

Another point: how do you form your specific tag? For example, should you use "math" (American English), "maths" (British English) or "mathematics"? What about using underscores or hyphens or CamelCase? Singular or plural? While some tagging guidelines are available, there are no standards. One solution is to use all variations at once. While this will ensure you will get the tag-searchers, it also clutters your tagspace, reducing usability and potentially turning off readers. Another approach is to look at which version is most popular on a range of tag sites. Personally, I recommend going local: look at your corner of the blogosphere (blogs you cite, shared readers, blogs that cite you) and see what conventions are followed. If, later on, you decide you backed the wrong version, you can readily rename it.

Tagging Your Posts

The actual mechanics of adding tags to your blog posts are straightforward. There is a range of bookmarklets, greasemonkey scripts and other add-ons to support this. In term of the process, should you have only a few posts, it's best to just go through them in chronological order. I'd suggest that each post will average about two minutes to process (one minute to skim and choose tags, another minute to add them).

If you're facing several hours of this, it's best to break it up into blocks. To that end, I recommend prioritising your tagging efforts. Through mechanisms like web traffic counters and backlinks, you should have a good idea which are your most popular posts. Put these at the head of the queue. Next, your blog may have changed its content or focus over time. Posts from an earlier, less relevant epoch should go to the end of the queue. In the middle, put your "recommended posts" ie those that are not necessarily seeing a lot of traffic but which are likely to be of interest to readers of the high-traffic posts. (Do these in reverse chronological order. More recent means more impact.)

After this experience, chances are you'll want to modify your tag selection again. On the second pass, allow one minute per post for, say, 25% of posts.

Tag Maintenance

From time to time, you will feel compelled to add new categories (recurrent tags). This is natural as you discover "better" tags for existing content (ie more popular or widely-understood versions of your tags) and your blog evolves its scope. However, there is the danger of a "tag explosion" rendering your tagspace unwieldy. To instill discipline, you could try promising to yourself to only add a new category if you're prepared to go back through your posts and reflect the changes throughout the blog history. To that end, ask yourself (in order of ease and frequency):

1) Is this renaming an existing tag?
2) Is this splitting an existing tag into two?
3) Is this a brand new tag that could apply anywhere?

The prospect of wading through all those posts again to see if the new tag applies should be sufficiently irksome to put the brakes on too much enthusiastic tag creation.

Finally, an important aspect of maintaining your tags is to monitor their usage. Most blog category/tag mechanisms (eg built-in, FreshTags, links to delicious and search engine approaches) will provide you with traffic stats about which tags your readers are using. It's worth considering whether or not your readership agrees with the tags/categories that you think are important. Do your readers use them like generic categories? Or do they prefer the really specific tags? What about tag-searchers and other inbound traffic?

Now, I'd like to hear from you. What experiences have you had with retro-tagging your posts? What tips would you offer newbies? What are some pitfalls to watch out for? What about automatic tools to help with this arduous task?

Filed in: , , , , , , , , ,
Posted at 7:23 AM by Greg.
<    >
Blogger Aditya said...
one of the best tools which one can use to tag with technorati and is performancing for firefox, since it posts directly using the API, rather than push you to a login and post page.

it can also read previous posts from your blog, so you can easily load them up in PFF, edit them by adding your tags, and republish them as an edit (the date/time gets recent to the current date/time, which they should fix).

newbies can start by taking it slow. its easy to tag one post with everything you can think of, which will only result in about a thousand tags with 5 posts between them. we don't want that ... we want to make it easier. think of the 3/4 best topics which the post covers, and put them as the tags. readers will appreciate it, and you'll find it easier to track and edit if the need arises.

for example, taking this post (just an example!), see that its tagged under technorati, tags, tagging, folksonomy, blogs, blogging, blogger, blogtech, delicious, how-to -- something which will confuse the heck out of me. rather, try combine the tags such as blogs and blogging into one, tags and tagging into one.

small considerations like these can make a difference. also, make your tags easily accessible to people by using tag-clouds or some hack like freshtags, but make them easily spottable and usable, so that people know how to get the information they want.

hope this helps! :)

<    >
Blogger Singpolyma said...
Great post! I would tend to disagree about the anchor tag (although I know most people seem to use one). I think that one should use one account for one tagspace. That means one account for bookmarks, and one for each blog. It may be a tad less conveniant, but it keeps you more organised.

<    >
Blogger John said...

One limitation of tagging w/ PFF at this point. You can't customise the URL of the tags... tags point to a generic page there, and the Technorati tags the same.... For this tool to be 100% useful in our application of tags & tagging, the URL would have to be customiseable to point to an iniidual user's acct. Other than that, though, it is great to see tagging (and trackback) integrated into the posting process.

<    >
Blogger Greg said...
Aditya: I agree it can be a little confusing with near-synonyms like "tag", "tagging" etc. But I was just following the house style! (ducks) :-)

Seriously, John has previously admitted to being guilty of tagging with wild abandon - presumably to maximise exposure to tag-searchers. This raises the question: is it okay to "turn off" your post footer tags (perhaps via "display:none" or writing them in the background colour). After all, most of these are included for the benefit of non-humans. OK, they also have the benefit of showing the reader examples of what "how-to" or "humour" might mean to the publisher. But when is enough enough?

Singpolyma: I think that between anchor tags (for feeds) and tag bundles (for pages), you should be able to get away with one account and "keep organised". Sure, there's a performance hit for forcing delicious to do tag intersections all the time, but with their sparkly new Yahoo servers it's not really noticable. Do you have a usecase for when multiple delicious accounts keep it neater? I'd be interested to know, in case I need to design around it.

<    >
Blogger Aditya said...
im sorry if this is a newbie question, but what and how does customizing tag url make a difference?

<    >
Blogger John said...
Adi: Customising the tag URL is what (at least in my mind) changes these over from tags to categories. If your URL is then when you click the tag on the post, you see everything tagged chickens by any user. If you can customise the URL so that it says then you're only shown results from my acct, and therefore only the stuff that I've categorised. The custom URL makes it possible to only display the posts from your acct.

Technorati recently went halfway towards the same sort of system, and if they'd give us a sidebar cloud or a tag feed, would have a rival....

<    >
Blogger RPM said...
I have started from my oldest post and am tagging bit-by-bit every time I get some time. My steps:

1. Log in to my blogger account in Firefox.
2. Since blogger dashboard does not allow me to go all the way to the bottom, I actually open a new tab and go to my blog and open up archives and go to the earliest period.
3. Once I have all the posts for that archive, I open each post with the 'edit post' pencil icon in a new tab.
4. I am using the greasemonkey script for tagging to delicious (through the link field). So when I open the post for editing, I can simply enter the tags and publish.
5. Then I link to delicious using the link generated after publishing.
6. Close that tab and go to the next post and repeat.

It is quite slow and tedious, but I don't try to do too many at once. And I am in no rush to get all done.

Earlier, I had simply done searches through my dashboard for certain common topics/keywords that I write about, and tagged them separately. So I know I have taken care of some of the important posts.

<    >
Blogger Singpolyma said...
RE: anchor tag -- I've actually never used an anchor tag, but my original reasoning for using multiple accounts (which is always what i've done) was that I don't want people subscribed to my main del feed to be seeing all my blog posts all the time -- if they want blog posts they sub to my blog, if they're subbed to my del they want bookmarks. At least that's my logic.

<    >
Blogger Greg said...
Fair enough. It's never occurred to me that people would subscribe to delicious links!

<    >
Blogger Mary Whisner said...
I appreciated your distinction between tags and categories.

When I first imagined having some sort of indexing, I thought about having general categories (like the labels I used for a long time: "Case," "Legislation," "News," "Event," etc.). (My blog is about trial practice issues.) But then when I searched for a way to add indexing to a Blogger blog and found FreshTags, I saw how easy it was to add lots of descriptors.

So I have the hybrid you describe. There are general categories that frequently recur, and there are also tags for finer indexing. A case about the admissibility of evidence under Rule 404(b) is tagged with the broad category "Cases," the fairly common "Evidence," and the quite specific "404(b)." A post about a particular judge will have the general "Judges" and the individual judge's name.

I call everything categories simply because that's the way the bookmarklet labeled them and by the time I figured out that I could change it, I'd already invested time in tagging a lot of posts. (Now my bookmarklet puts in "Filed as," which gets away from "Category.")

If I wanted more purity (in terms of information science), I would distinguish between categories and tags (or index terms), but I'm trying to just get the job done.

I've been running my blog for over a year and I feel familiar with the range of posts I've written, but I've found as I've gone back to tag old posts that some topics recurred more than I expected. I like having the tags to pull them together. For instance, there have been several posts about sports cases, and some people might like being able to retrieve them easily.

I'm aiming for tags that would be useful for navigation within my blog, rathern than trying to find tags that would work for people outside. That means I don't have some tags that, in another context, would make perfect sense. For instance, I don't have a tag for "trials" because the blog is about trial advocacy and almost every post has something to do with trials. I don't have a tag for "Washington State" because the focus is on Washington State.

When I summarize cases about evidence and trial procedure, I stick to tags that are about those topics. The case might be a trial for murder or child abuse, but the reason I'm summarizing it is to for the evidence question. (Putting in a lot of tags about the aspects of a case about sex with children might draw visitors to the site, but they wouldn't find the site interesting and would leave quickly anyway.)

It's been an interesting exercise. And I think the tags will make the blog much more useful.

eXTReMe Tracker