Three tools to measure Twitter success

Many companies look at hits on a website, clickthroughs from advertising, numbers of “friends” and mentions in blogs (to name just a few) to measure the success of online activities. I’m not sure that these are meaningful, but how can you evaluate success without some sort of metric? You’re a very savvy bunch and probably know all about these, but I just came across three great websites/services that can help break through the noise, at least on Twitter. The sites are relatively primitive so far, but have the potential to help marketing managers figure out whether their product is a topic of conversation and, to some extent, if that chat is positive or negative. The first is tweetvolume.com. Type in a few terms and it shows how often they creep up on Twitter, all in nice chart format. For example, type in PLM, CAE, CAD and CAM, and you get the following: (Image is a screen capture from tweetvolume.com) Of course, such short terms mean nothing in this context; who knows what other words can be abbreviated using these letter combination? (Also keep in mind that Twitter is language-agnostic, so a search for “cae” also turns up Spanish tweets, where it means “falls”.) One can play around (and waste lots of time) narrowing search terms to more meaningful levels. For example, type in product names and the game changes: (Image is a screen capture from tweetvolume.com) Interpreting these results isn’t so easy. It could be that SolidEdge users don’t tweet, that Inventor is a noun that represents both a product and a personal accomplishment or that tweetvolume’s search engine can’t correlate SolidWorks and SW, a frequently-used acronym for SolidWorks, and SE to SolidEdge. At this point, tweetvolume can really only be used to gauge if one term is being tweeted about more or less than it was before. (The site has flaws, such as not allowing chart customizing or set a time period over which this volume is discovered, but for tracking a term over time, would be quite useful.) backtweets.com provides another popularity metric: how many tweets have linked back to a particular website. Running its search engine through autodesk.com, solidworks.com and solidedge.com finds that roughly 5.5 pages of links exist for Autodesk and 4.5 for SolidWorks for the period from October 1 2009 to October 23, 2009. (SolidEdge had no hits — Twitter really doesn’t seem to be a focus for its users.) Interpreting, again, is the key: who is tweeting and are they original or retweets seems more relevant than pure volume. Too, the number of hits alone may mean nothing — Autodesk supports lots of products via autodesk.com, so the comparison to solidworks.com is perhaps unfair. But backtweets could be helpful in judging changes over time during a particular campaign. Perhaps the most useful site is tweetbeep.com, which purports to track not only mentions but also sentiment. Type in search terms and see how often they have recently turned up, either realtime or through an email alert. Most interesting is that the site will convey if the tweets in the search convey a positive :) or negative :( attitude or are asking questions. The site’s help wiki doesn’t describe whether sentiment is gauged solely by emoticon [ :) ] or if some sort of sematic algorigthm is employed; if it is only emoticons, the sentiment component won’t be terribly useful for PLM marketing. After all, how many adults use emoticons when typing even 140 characters about a work topic? But the utility here is in knowing that something is being said and gaining the ability to respond in public or in private to the conversation. I still maintain that raw numbers of tweets are rather pointless, as are numbers of friends, clickthroughs or links. But when you have no information to gauge if a campaign is at least reaching people, these tools are better than nothing.