Twitter Analytics: Presentation from Social Media Masters

For anyone who is interested, my presentation from Social Media Masters is available for viewing on Slideshare:

Twitter Analytics: Michele Hinojosa, Red Door InteractiveTopics include: Twitter research for competitive intelligence, hashtag and network analysis, download, export and backup tools, Twitter account measurement using Twitalyzer and Klout and tracking Twitter links back to your website. Originally presented at Social Media Masters Twitter workshop in San Diego (2/11/2011)

These are a few of my favourite things (about Twitalyzer)

For those of you who don’t know, Twitalyzer is one of my favourite Twitter measurement tools. Why? Well for one, they let you see the math. As someone in the digital measurement industry, I find it hard to put a lot of faith in a “secret sauce” approach to measurement. (“Just trust us, this is your score.” No thanks.)

I have been using Twitalyzer for months now, first through a free option, and now through a paid subscription. Why did I pay? Because it’s worth it. (And FYI, the individual subscription is cheaper than a cocktail. That was a very easy sell for me.)

So (Von Trapp style) … here are a few of my favourite things.

1. Customisable metrics report

Over the months, the number of metrics included in Twitalyzer reports has grown. Great, but for me, it’s approaches a “TMI” level. Therefore, I love that I can actually customise not only what order different metrics appear, but also whether they appear at all. If I decide something doesn’t help me manage my little corner of the Twitterverse, I can choose to hide it, and simplify the metrics report. To me, this is especially useful in simplifying the world of social media to those who may be newer. Choose the key metrics you need them to see, and perhaps add to those over time.

Now, to be honest, I’m looking forward to evolution of how you customise the metrics report (currently it’s a little more cumbersome than I’m sure it will be in the future) but the customisability alone wins big points with me. (FYI, it also transfers across accounts – so if you have a number of Twitter usernames connected to your Twitalyzer accounts, you don’t have to repeat the process every time.)

So what do I care about?

I’m still playing around with this, but here’s what I am currently showing on my Twitalyzer metrics report:

Impact: My impact is defined by how many followers I have, how many references there are to me, the frequency at which I’m uniquely retweeted and retweet others, and the frequency at which I post. (I call this one the Twitalyzer Ubermetric.) 

Engagement: The ratio at which people reference me vs. those I reference.

Influence: Is the likelihood that a Twitter user will either retweet or reference me.

Retweeted: How often I am retweeted. I find this useful to help me understand what content of mine my network feels is worth sharing.

Signal: Whether my tweets contain a link, mentions or retweets another user or include a hashtag. Aka a way to measure how many of my tweets are of the “Ohhhh, my cat is so cute!” or “Yum, I just had pancakes” variety vs. something actually interesting. I’m sure in some networks, depending on your contacts, low signal isn’t a big deal. But I’m conscious of not shoving too much “status” type stuff out there. (I figure general “here’s what I’m up to” is what Facebook is for.)

Generosity: The percentage of time that I retweet others.

Clout: The likelihood that I’ll appear when searched for in Twitter. I won’t lie – this one still makes me go “Hmmm …”

Followers: Number of followers I have.

Lists: Number of lists I’m on. Often it’s a good idea to look at what the lists are (e.g. I’m on some lists of “People who RT-ed me” … probably not one that the list creator even reads!) but it’s definitely a “nice to watch.”

Velocity: The frequency at which I post.

Referenced: The number of times I’ve been mentioned on Twitter.

Potential Reach: The number of followers I have + the number of followers those who retweet me have. This attempts to estimate how far a tweet I send can be viewed.

Effective Reach: Takes my retweeting followers x my influence (the likelihood that I’ll be retweeted.) This attempts a more realistic estimate of how far a tweet I send could go.

Klout Score: Comes in through Klout. (The “secret sauce” tool.)

So why do I care about these metrics, and why don’t I care about others?

My reasons for Tweeting in the first place, and for tweeting about the topics I do, are fairly simple: to learn via reading what others post, to contribute to the discussion, and to build relationships in the community.

Measuring my signal ensures that I’m not posting junk that might alienate my Twitter contacts. (Sure, my cats are cute, but I’m pretty sure nobody cares.) Monitoring retweets, followers and lists is a good indication of whether what I’m posting is considered of any value by my contacts. Monitoring references to and by me, and my retweets of others, is a good indicator of whether I’m engaged with the community.

Basically, I’m choosing to focus on the things that relate to my goals in using Twitter, and as such I’ve currently hidden some metrics in Twitalyzer.

2. Percentile

This is a small one, but when I first started using Twitalyzer, you could only see “Percentile” by clicking into a metric. Now they have incorporated it on the metrics view, visible at all times. I love this! The percentile basically tells me that my score is the same or better than X% of the Twitter users tracked by Twitalyzer. This can be helpful when you realise that a Impact score might seem low, yet still be in a fairly high percentile. (Aka it makes me feel better…!)

A nice future add-on would be a tailored percentile, based on a subset of Twitter (perhaps those you follow, or a list you’ve created.)

3. Goals

Data in a vacuum is meaningless. Twitalyzer lets me set goals for individual metrics, to ensure that I’m actually measuring something in context.

4. Comparisons

I love the ability to compare myself to other Twitter users on a variety of metrics. To me, this is another way to benchmark myself, and perhaps even help to set my goals. (What’s feasible? Well, what is X’s score? Okay, I’ll aim for something similar.)

However, I’d definitely love to see a future ability to pre-create a number of comparisons, rather than doing this individually each time I log in.

5. Recommendations

It’s pretty standard for a tool to tell you how you measure up. It’s less common for one to tell you how to go about moving the needle. Twitalyzer will actually give you recommendations of what is more and less important for you to do to positively affect your impact scores.

6. Network Explorer

Now, to me this is more informative than necessarily something you’re benchmarking against, but it’s still super cool. Basically the Network Explorer is a visual map of the people and topics you are connected to. (Because who doesn’t like to play Six Degrees of Awesome #Measure Folks?)

7. It’s about the community

So you follow these people on Twitter, or they follow you. So what?

Twitalyzer can help you understand:

  • When your network is online (e.g. time of day, day of week) and where they are located. (I definitely think there are some improvements that could be made in the presentation of regional information, but that’s an aside.)
  • This can be very helpful information if you’re trying to decide when the optimal times to post, say, a blog post are.
  • Who the most influential members of your network are, who your most influential “friends” are (based on your communication with some members of your network more than others), and who appears to to influence you.

Network Activity (Time of Day):

8. It’s about me

Curious about what dates or time of day you tweet? Are mentioned? Retweeted? Never fear, Twitalyzer is here:

But what about the content? What hashtags do you use? What communities and discussions do you tend to be a part of? What of your content has been retweeted lately? Which of your links were clicked? You can find all of that out.

There’s even the beginnings of sentiment analysis. Now, sentiment analysis is a toughie. (Especially if you have folks in your network who are, shall we say, sarcastic? Noooooo….) The reality is that there aren’t many great sentiment analysis solutions out there that don’t involve human assessment of sentiment. So to be honest, I tend to view all sentiment analysis (included Twitalyzer’s) as interesting, but not necessarily something to hang my hat on. That said, I do like the ability to play around with what words you consider to be negative and positive:

9. Integration with Google Analytics

I’m also looking forward to jumping more into the integration with Google Analytics – but that might turn out to be a topic for another time (entirely of itself.)

10. Export options

The ability to export data does not go unnoticed!

Last but not least … (definitely not least!)

11. Customer service

I’m not kidding you. Even before I was a paying customer, Jeff Katz at Twitalyzer was always enormously helpful, and though I’m a now only a meager paying customer (with my little individual account) he still goes above and beyond to help solve problems. He rocks.

And as a little FYI, if there’s something you can’t get working right in Twitalyer, take a hop over to Firefox or (shudder) IE. I’ve had very few issues using Twitalyzer in Chrome, but the rare issue I’ve had has been browser related. (Which I discovered because the aforementioned lovely gentleman Jeff Katz helped me out.)

Like this post? Related content:

What time of day do web analysts tweet?

Thanks to @menggoh (and Virgin America‘s wi-fi) I have been thoroughly entertained on a red eye between LA and Boston with Twitter analytics from The Archivist. Which means naturally, I was curious: based on when the web analytics community is active and tweeting (based on tweets to the #measure hashtag) when is a good time of day for me to post new blog posts?

The cool thing about The Archivist is that they give you tons of fancy charts about who the top users our, our daily tweet volume over the past few months, topics discussed etc. (Seriously, check it out – so much fun!) However, what I couldn’t find was a breakdown of total tweets by time of day. The even cooler thing about The Archivist is that they allow you to download an Excel file of the data. (Insert cry of geek joy here!) So naturally, I took the data and DIY’ed it.

So as it turns out, the peak of #measure tweets is between 6-7 pm. We’re primarily an evening/night owl community (insert HootSuite joke, anyone? Get it? Owl? Hoot? My apologies – I blame the red eye flight…) with much lower activity in the morning. Possibly we work on being a peaceful community by waiting for the coffee to set in before communicating too much!

Here is total tweet volume to the #measure hashtag for 5/11 through 7/2/10.
[Ideally, I’d like to overlap total tweet volume to see whether we follow the overall Twitter trends. I’ll update when I have further data.]

*Note: Time of day is based on United States EST

So what does this tell me? I theorize that I probably don’t want to post my tweets at 6pm, right when a huge volume of tweets are flooding in (too easy to be missed in the midst.) However, perhaps noon/1pm would be a good time, so my tweet is recently posted as we start getting more active on Twitter for the day.

Now, I’m off to test some time-of-day theories …

[This adventure in geekdom was proudly brought to you by The Archivist and Virgin America‘s wi-fi service.]

To follow, or not to follow?

I have been using Twitter for a few years, but more on a personal basis (private account, that was essentially the same stuff as my (also private) Facebook status updates.) It’s only in the last few months that I’ve explored Twitter for more professional reasons. I wish I’d caught onto this earlier. At least for my work, being in the online space, I find that Twitter is an amazing educational resource. There’s an entire web analytics community out there posting their experiences and interesting blogs and articles, some of the gurus of web analytics (@erictpeterson, @avinashkaushik) sharing their thoughts with us without a price tag, and even client-support people sitting on Twitter all day ready to answer questions. (@OmnitureCare)

The one thing I am trying to get my head around is Twitter “etiquette”. As a fairly pragmatic person, here’s how I am approaching the “to follow or not to follow” question.

If someone follows me, great. I’ll check out their recent tweets. If they sound like someone who I want to hear from, then I’ll follow. If not, it’s very nice they’re following me (thank you!) but I don’t feel obligated to reciprocate solely because they followed me.

Things that make me want to follow you:

  1. The majority of your posts are relevant to why I’m on Twitter (web or business analytics, social media, new gadgets, the online space generally, etc.) I’m fine with some more personal stuff interspersed, but I won’t follow you if 9 out of 10 posts are unrelated to my interests.
  2. You share some of your unique thoughts about the industry. I don’t want to follow someone who does nothing but RT others. I’m following you to hear your thoughts, not to hear you habitually regurgitate others.
  3. I am, however, interested in useful RTs, to help me find others I might want to hear from and article posts.  RTing is great, it can help distribute interesting content, and I like reading interesting RTs. And hey, I like being RTed! But per #2, if that’s all you do and share none of your own thoughts, I’ll figure why not just follow the people you keep RTing and take you out of the mix?
  4. You post regularly but not too regularly. I don’t want to hear from you every 30 seconds. I find it hard to believe you can post super frequently and still be posting quality information. Quality over quantity.

Things that will make either pass on following you, or un-follow you (if I already started):

  1. Too many posts. If you post so frequently to take over the feed, I’m likely to unfollow you, unless they’re absolute gems (again, this seems unlikely. PS Why aren’t you working?)
  2. Useless posts. I’m very sorry, but I don’t care if you just ate a ham sandwich, and I definitely don’t care about sports. If you tweet every two seconds of the World Cup, I’m going to un-follow you. It’s nothing personal, I just don’t care about sports. If I wanted to read about it, I’d follow some sports-type-people. (Can you tell I’m such a … uh … sports uh knowledgeable person?)
  3. If you write in a language I don’t understand. This one is absolutely not personal. I actually wish I could follow some people who are posting in German, Spanish, etc. It’s just the reality that there’s not a lot of point in me following you if I don’t know what you’re saying. Blame that one on me for not understanding every language. (That would be nice.)

I would love to hear anyone’s advice of some good dos and don’ts of Twitter etiquette. You can find me at @michelehinojosa. Feel free to follow me, but only if I meet your follow criteria.