Right and Wrong Ways to #Hashtag on Twitter

A consideration of the pointfulness and usefulness of hashtagging.


Tagging, before hashtags, was a way of creating a many-to-many, theme-based alternative categorization of content. In particular, blog software such as Wordpress was designed to have a blog post reside in one category of posts (e.g. recipes), while tagging would allow you to use several keywords to relate similar blog posts together more fluidly (e.g. bacon, breakfast, health), either to help readers hop from a blog post to similar ones, or to generate a kind of third, middle ground, option in discovery somewhere between keyword search and category browsing. The tag cloud is a good visual way of displaying frequent/important topics and encouraging browser discovery of interesting content, and using tag clouds nicely blurs browse vs search navigation and discovery.

Tagging blog content became popular not just because of helping human users read more: tagging content into directories such as Technorati used to be an effective SEO tactic because of the frequency of search bots spidering big topic tags and hence coming back out of those major hubs to discover and index your content faster, meanwhile crediting it with good on-topic pagerank. Spamming broad terms was effective because of this pagerank effect. It’s worth bearing that in mind when you see how some marketers instinctively approach hashtagging.

Why tag with #?

Originally, adding a # symbol to a keyword served the purpose of making that keyword unnatural and unique. The symbol itself was not important, just the ability to distinguish “#SEO” from “SEO”, where a search for “#SEO” only brings up something people have purposefully tagged to be included in the results for that search, not all the other mentions of ‘seo’ (which is also a Korean word, which in a huge pile of content like Twitter mixes up lots of irrelevant stuff in searches).

The power of hashtags is that they allow users, without any special permissions, to create virtual categories inside systems like Twitter, without any official curation. The implication of an exact keyword match for a hashtagged term allows for opting out of whatever broad matching the normal keyword search system offers. This in turn allows users to add back in items to a keyword which would otherwise be missed, for example a post about how to optimize your website for Bing does not mention SEO, but you tag it #SEO to say it should appear in that topic search. In a way this is a kludgy way to improve search into a “things not strings” effect.

The usefulness of hashtagging

Hashtags, in Twitter and many other places, are (now) honoured by the software and automatically linked to a search for that hashtagged term. Assuming “correct” hashtag usage, this makes the hashtag link a very visible navigation cue, leading to a de facto topic feed with content that is possibly a better match than using normal category browsing or a bare search term. Google+ attempts to add hashtags automatically (if you don’t add your own), but generally in Twitter and elsewhere, hashtagging implies human intelligence: content producers/sharers adding the tag manually, and we hope (see rest of article) thinking a bit about how to assigning a topic to the content, should produce good results.

In understanding these origins of the hashtag, we can see that its main usage is about improving the chances of discovery of content within a closely related network of similar content.

“Improving chances of discovery” imagines three routes:

(a) people reading your content click through the hashtag (i.e. perform a search with that hashtag) and discover useful related/similar content;

(b) people do the same on other people’s content and, via the hashtag search results, find your content;

and (c) people initiate their search on a hashtag with a deliberate intent to discover related content.

Of these, I believe the latter is the most plausible situation, where the user has an organic reason for wanting to browse many/all tweets on a hashtag, e.g. to follow breaking news, participate in a Q&A, add jokes to a meme thread, see what others are saying at a conference, etc.

A corollary of the above is that hashtagging which is ambiguous is not useful according to the original point of hashtagging, i.e. you tag with a broad or ambiguous term where lots of unrelated content could be tagged like that, so the assembled content when searching with that hashtag lacks a strong topic or audience connection. An overwhelming quantity of search results can remove usefulness just as much as poor relevance of search results.

Here are three common types of poor-relevance hashtagging:

1 — Very broad location

e.g. #London

It’s not (yet) a normal situation to stack hashtags in searches, so there is no “refinement” in adding #searchterm + #london … and the result is that the hashtag #london search is cluttered with a huge volume of different stuff that makes using it pointless.

In this example you can see that it would have been more effective to use a more specific hashtag like #londonboroughmarket, #boroughmarket, or #londonbridge (area)… or at least a location+topic tag like #londonshopping #londontourist or #londonfood

(On the other hand, if the location is broad but still not that frequently tagged, it could still be effective, e.g. #seychelles)

2 — Very broad reaction term

e.g. #LOL

As you can see here, not only does #lol produce a massive stream of stuff that has virtually nothing in common, but it also catches in other stuff completely — in this case League of Legends twitterer who is unaware that hashtags on Twitter are case-insensitive.

It obviously seems intuitive to a lot of people to add hashtags to reaction, exclamation, or “emoticon” type words, e.g. #FAIL #facepalm, but when you reflect on this usage, it is hard to find any actual rationale or usefulness.

3 — Very broad content category term

e.g. #selfie

As you can see from the timestamps, the mass of content tagged on a generic content-type keyword like #selfie makes the tagging pointless.

Another example of a “content category term” would be tagging the contents of a tweet very broadly…

e.g. #chocolate

…or a style…

e.g. #retro

Obviously the main causes of this kind of pointless tagging are the interrelated problems of not understanding what hashtags are for, thinking that for some reason you have to hashtag your content, and thinking that hashtagging particular words makes you more funny or otherwise with-it.

With general individuals’ “share random stuff with nobody in particular”, this doesn’t matter, but in the case of intentional Twitter usage for marketing and generally grown-up purposes, it is worth questioning whether you have a more useful practice in effect with hashtags. This begins with understanding the opposite of the teenagers’ misconceptions, i.e. (a) you do not have to hashtag anything or everything; (b) hashtagging does not add any value to your content as it stands.

On the basis that good hashtagging should invite and aid discovery of related content that is likely to be of interest to the consumer, and that the results should not be so broad or massive to be unusable, let’s look at some good practice.

10 good types of hashtag usage:

  1. The broad-but-not-too-broad theme hashtag.
    e.g. #contentmarketing
    #photoshoptips
    #html6 [check it out]
    Assign your tweet (or content behind the tweet) a general category, that is not too spammy, and not too over-used to be unlikely to lead to relevant discovery. A lack of thought about whether a topic tag is too broad is certainly the no.1 reason for ineffective hashtagging on Twitter. An excellent tool, Ritetag is available for serious hashtaggers to guesstimate the likely effectiveness of any chosen tag before tweeting.
    Aside from volume and relevance considerations, I think the key third consideration is whether there is true usefulness to the content consumer in clicking through the hashtag to discover more content, or deliberately searching using the hashtag. (If there is true usefulness, then the traffic and attention will be there.)
  2. The longer tail / niche theme hashtag.
    e.g. #twitteradvertising or #twitterads obviously better than #twitter, #advertising, or the unwieldy #socialmediaadvertising
    This would also be applicable to interest themes which are too broad in themselves and benefit from narrowing, e.g. #dryfly vs overly broad #flyfishing
  3. The current affairs news item hashtag.
    e.g. #europeanelections
    #MH370
    As can be seen from these couple of examples, the hashtag is very suitable for tracking long-running or slow-burning news items, as well as breaking/trending ones.
  4. The campaign / cause / movement hashtag.
    e.g. #BringBackOurGirls
    #bringbacklando
  5. The real-life-event hashtag — usually being industry conferences / trade fairs.
    e.g. #infosec14
    #aeexpo
  6. The webinar / MOOC hashtag.
    e.g. #twitteracademy (the assumption is that it only gets used and becomes relevant during live class / Q&A and falls quiet other times)
    #tmmooc — for online learning it may be a good way to help gather together disparate content which at least has in common that the contributors are involved in the same course.
  7. The meme participation hashtag.
    e.g. #AddaWordRuinaMovie
    #asianproblems
  8. The unofficial, community-agreed brand hashtag (with the proviso that it is hard to remain pointful tagging with the bigger company / broad brand keywords).
    e.g. #radio3
    #Wimbledon2014 — this example shows community hashtag using the year to create specificity, better than the official Wimbledon organizers generic use of #wimbledon
  9. The brand-proposed hashtag to create a sub-group / sub-topic.
    e.g. #w3ccommunity
    … which could also include attempt to sub-brand top-down content marketing campaigns: #riskeverything
  10. The stock ticker code hashtag, denoting financial / corporate angle content.
    e.g. #TWTR (but only workable when the stock ticker is unique and most likely to be referred to by people talking about the stock market, cf uselessness of #FIZZ [National Beverage Corp. (NASDAQ:FIZZ)]
  11. and one for luck…
    The very broad term which becomes temporarily meaningful in a particular time because of something happening at this time
    e.g. #guangzhou (trending news item about a knife attack in that city)
    #WorldSnookerChampionship (and really all sports events)

Conclusions?

  • Don’t use hashtags mindlessly;
  • Related content discovery is the original and main point of hashtags: therefore using them to draw attention to words or just for comic effect is discouraged;
  • You don’t need to use hashtags all the time or have more than one in a post;
  • Smart use of hashtags can help your content consumers and you;
  • You can improve the likely beneficial effects of your hashtagging without getting overly analytical or slowing down your process: just consider the core principles of pointfulness and usefulness.

And finally my speculations…

  • Dumb use of hashtags is very often a signal of unimaginative / try-hard brand marketers and works against them even in the eyes of consumers who are not thinking about hashtags as much as us;
  • Nobody knows the SEO value of hashtagging so, in the absence of any serious experiments to find out, just ignore the question and focus on human content consumption;
  • Improving hashtag usage in Twitter will presumably improve your hashtag usage in other places it is relevant, such as Google+, Pinterest, and Instagram;
  • Thinking more about hashtags and how to use them in your content should also lead to good ideas about how to use others’ hashtagging of content, e.g. to group audiences and create lists, to track and discover topical content, and to find trending sub-topics to get involved in ethically, not just hashtag-hijack.

This entire post might just have been to contribute to the spelling of Hashtag as a single word without a space or hyphen. ~ @georgebaily