miércoles, 17 de julio de 2013

How to get started with Twitter analytics

Twitter Analytics

Twitter rolled out its built-in analytics to all users fairly recently.

If you're an advertiser with Twitter, you'll have had access to Twitter analytics for a while, but for everyone just hopping on board, I thought it would be useful to take a spin through the various features and look at the insights you can (and can't) glean from them. 

EDIT: Some users may not be able to access the ads and analytics features just yet. Twitter appears to be rolling these out to the public. For more details, see the Twitter developers blog. 

First things first; Twitter's offering isn't really designed to give you massive insight. It's a fairly top-level view of your Twitter activity, but with a little lateral thinking there's a lot of information you can extract.

As always, try not to rely on information from any single source when using analytics, and make sure you cross-check with your site data if possible. 

First of all, you'll need to access your analytics. To do this just head up to the top of your profile and select 'Twitter ads' from the drop down menu: 

http://assets.econsultancy.com/images/resized/0003/5344/1-blog-full.png

This will open a new window and you'll be prompted to reenter your user name and password to access analytics and ad campaigns. 

Once you are in you'll see there are three analytics options available from the drop-down at the top of your page:

http://assets.econsultancy.com/images/resized/0003/5346/3-blog-half.png

  • Timeline activity
  • Followers
  • Websites 

Here we'll mainly be looking at the 'Timeline Activity' and 'Followers' sections. 

Timeline activity

http://assets.econsultancy.com/images/resized/0003/5347/4-blog-full.png

There's a wealth of information available here that lets you see just how well each of your tweets are doing.

You can filter these by clicking 'Best', 'Good' or 'All' at the top of the 'Recent Tweets' section.

By scrolling down you can clearly see the number of favourites, retweets and importantly, replies each tweet has received.  

Examine these figures carefully as they give you a great, straightforward way to optimise your content.

Which factors did your most retweeted content have in common? For example, does a particular headline style do better than others? 

Twitter helpfully points out content that performs particularly well too: 

http://assets.econsultancy.com/images/resized/0003/5358/13-blog-full.png

It's worth remembering that 'reach' isn't directly tied to the number of retweets you get.

You could get 100 retweets from people with 20 followers each, or one retweet from someone with 1,000,000 followers, which will make a big difference, so you'll need to check who is retweeting to make the most of this information. 

Similarly, check the reply rate to see which kind of content receives the most replies. In some cases this may not seem valuable (Some of our best-performing tweets have been entirely off-topic), but in fact they give you a very clear picture of how engaged your audience are at any given time, allowing you to build better relationships with users.  

Click on the date under each tweet in analytics and check who has retweeted and favourite: 

http://assets.econsultancy.com/images/resized/0003/5357/12-blog-full.png

If you see lots of similar faces in here then you can use this to easily build a list of potential prospects and brand ambassadors and get in touch with them. 

Next to each tweet you'll also see 'number of clicks'. This figure matters, big time. This shows how many tweeters are following your links. 

Now, there is room for confusion here so be careful when examining these figures. 

First of all, Twitter is displaying the total number of clicks on a particular link.

This means that if you send the same link out twice, this figure will display the cumulative number of clicks it has received. If however you send a link to the same page, but append it with different tracking data or use a different link shortener, this figure will be for each unique link. Here's one I posted recently as an example: 

http://assets.econsultancy.com/images/resized/0003/5359/14-blog-full.png

The second time I posted that, I used a different shortener, so in fact 256 (182 +74) people have clicked through in total. 

Also, remember that one 'click' does not always mean one visitor to your site.

There are all sorts of ways that click can be interrupted before the user hits your content, so check figures using your site analytics. These numbers will never quite match, but you can combine them to get a relatively clear view of things.

I've written about the difficulties you may face in the past if you are interested. 

All these figures should help you to optimise your content, so check them regularly and take these insights on board. Types of content that perform well now may not be as popular in  a year's time, so spend time thinking about why. 

Timeline graph

http://assets.econsultancy.com/images/resized/0003/5348/5-blog-full.png

This is relatively straightforward, displaying follows and unfollows over the last 30 days (Pacific time).

Hover over each day's graph to see a breakdown, including number of mentions for that day.  

Twitter doesn't archive this so it's worth downloading the data once a month.  

You'll note that with our content, content that gets a lot of follows also gets the most unfollows, which (probably) means... you can't please all the people, all the time.

If you start getting more unfollows than follows, then it's time to take a serious look at your content. Unfortunately there's no ability to drill down into these figures, so you'll have to do this manually. 

Followers

http://assets.econsultancy.com/images/resized/0003/5349/6-blog-full.png

Again, we've got a wealth of data to play with here, all with specific uses. 

The graph at the top displays your followers over time.

This data doesn't update in realtime, so you'll need to give it a few days to catch up on any specific campaigns.

Ideally that graph should be a nice steady upwards curve, with a few spikes around specific campaigns. Remember, Twitter engagement is an ongoing process, so you don't want lots of bursts around campaigns and nothing in between them. A steady climb will show that you're keeping people interested. 

In the section below we have a large block which gives rough figures on the location of your audience.

These are rounded off so don't put all your faith in them, but they do let you know how well you are penetrating in target markets.

This is particularly useful if you run regional accounts, or are considering doing so. The reason Econsultancy doesn't have dedicated US/UK accounts is because we have large amounts of followers from both already, so forcing them to switch seems unnecessary.

You can drill down into this easily by clicking on any given content block. Here are our followers from California: 

http://assets.econsultancy.com/images/resized/0003/5351/7-blog-full.png 

Hello Sunshine-state-ers! 

Something to consider here: There are a number of reasons you may be doing well in any given territory. We do well in NYC and Cali not only because there are a lot of digital marketers there, but also because there are simply more people living there, and more Twitter users.

This is a small factor but could be very important to consider if you're struggling to gain traction  in a particular territory. 

The Followers display also highlights gender demographics, and gives you a handy list of the competitors your followers follow.

If 20% of your followers like a rival, then maybe the rival's other followers will also like you. Time for some outreach! 

Similarly, Twitter provides Interest lists to the left of the page. These are fantastic for topic and keyword research, and can help you easily identify any topics you should be covering but aren't. Again, it's not detailed, but it's a good starting place for some gap analysis and in-depth keyword research.

Bonus: Twitter Ads insight

If you head to the top right of this section, you'll see a small drop-down menu. If you are running Twitter ads, this will display the number of new followers you received as a direct result of a given campaign:

http://assets.econsultancy.com/images/resized/0003/5352/8-blog-half.png

(Yes, I ran a campaign called 'President Scroob'. It made sense at the time). 

You can check this data against the insights provided in your ad analytics: 

http://assets.econsultancy.com/images/resized/0003/5354/10-blog-full.png

This information displays your overall engagement rate (Users who took any action - including favoriting, replying, or retweeting - as a result of your campaign), so if your goal is getting more followers, you can cross check these figures as a helpful extra indicator. 

Finally, there's the 'Websites' section. This provides code allowing you to link your site directly to Twitter

http://assets.econsultancy.com/images/resized/0003/5353/9-blog-full.png

This will provide you with an extra layer of data displaying traffic coming from Twitter to your site, and information on the effectiveness of any Twitter buttons you may have on your site. 

There are full instructions on how (and why) you might want to do this here

We've been using Twitter analytics for a while at Econsultancy, and as with any analytics tools, it isn't perfect, but it does give valuable extra data about all our users and overall Twitter activity, as well as helping to identify opportunities for growth.

I always recommend that you use as many different sources of data as possible when examining your social activity, but for Twitter users this is a must-have. 

No hay comentarios:

Publicar un comentario