The Chatterbox

Analytics and Tracking Visits From Twitter Tweets

Today we're going to take a look at how businesses can make use of social media services like Twitter to support their marketing efforts. Concentrating on the main pitfall these businesses run into when utlizing such marketing methods, that being how to track the impact or your efforts to make sure you're receiving appropriate bang for your marketing buck. Or really if the time you put into social media marketing is worth the effort.

Needless to say, making use of social media is a far better fit for some markets than it is for others. As a for instance, if you were consultant of some type, where building your reputation of expertise was important, having a presence in social media would stand a pretty good chance of being useful to your marketing efforts. On the flip side, if you sold say Children's Toys you should probably expect that it'll have less impact to your bottom line as a general rule.

That said, one constant problem remains. You HAVE to be able to track users who see your Tweets, figure out how many click on links you provide and how many of these ultimately up convert into paying customers. Without the ability to track these actions you're basically shooting in the dark. You have no way to tell if one subject is more popular than others. Nor can you tell if addressing one subject resonsates better with users than another, thus producing higher or lower conversions.

A couple of ground rules before we dig into the meat of the subject.

First I'm going to be concentrating on Twitter for the purposes of this tutorial. It's not that the methods described below won't work on other social media platforms (they will!) it's just that I want to keep things as straight forward as possible.

Second, I'm not going to address how to get users to follow you at all. This document isn't about becoming a well known, highly followed Twitterer. There are already lots of documents out there addressing this issue. Instead this tutorial is solely about how to track what your followers are doing with the recommendations you make in your Tweets so that you can hopefully gain some insight and make better use of the system for business and marketing purposes.

Why Twitter? Well, the answer to that one is pretty simple really. First, I freely admit I'm not a big social media user myself. There already aren't enough hours in my day and I know from past experience how time consuming these types of things can be. Heck, I used to spend hours and hours playing different MUD games that were really nothing more than online social platforms. I just don't have the extra time to devote to these types of things these days, so I'm not a dedicated Tweeter myself.

The real reason I chose Twitter for these examples is that a friend of mine uses Twitter as part of his business. Specifically he started out using an application called Get Satisfaction that is basically a Twitter-enabled customer service/pre-sales tool. He uses it to answer questions people visiting his consulting site might have.

After he got into the whole Twitter thing he started getting some followers so he started blogging and then Tweeting about interesting things he ran across each day in his travels. Like many he'd blog about these things he found, then Tweet a link to his blog.

Seeing the amount of traffic these pages received from his followers he wondered how he might use it to further grow his business through a new traffic channel. Which was a good idea on his part. It's always a good thing to explore and open new traffic channels IMHO.

While it's a good idea on the surface, several problems arise immediately. The largest one being how does one track these hits that come from all over the place so that you can determine how effective the Tweet Campaigns really were?

Let's pause a moment to touch on why tracking these hits is so darned difficult. The reason is both a blessing and a curse.

First, if the clicks were coming from people who were viewing his profile page on Twitter it would be reasonably easy to track. The referer (sic, but that's the way it's spelled in the logs) field in the users browser would be easy for him to track in the Google Analytics stats for his site.

The blessing and curse is this: According to most studies only about half of Twitter users actually use the viewing mechanisms provided at twitter.com. Which means about half of Twitter users utilize one of several third party applications that have been built to make Twitter easier to use.

These additional applications or clients can be SMS (read cell phone) text messaging, desktop applications like twhirl or TwitKit that pull your Twitter feeds onto your desktop for easier reading, light-weight web apps built for cell phone usage over the 'Net, etc. This amazing flexibility, the ability to utilize one of several Twitter clients to access their system, is one of the features that makes Twitter so popular with its users. It is a true blessing for users, but a curse for marketers trying to track movement.

The long and short of the curse is that clicks on Tweet links from these other applications carry no referer information in the browser. They're often not browser based, so there is no referer. For all intents and purposes these visits will look just like Direct hits in your site stats, meaning you'd have no earthly idea how someone may have found your blog post.

Needless to say this is the type incomplete data that can severely skew data. Especially when you're trying to determine the Source or Channel to sort out if your Tweeting is bringing you good, converting traffic.

So that's where we start the exercise. We have to see if we can come up with a way to track these visitors, and do so reasonably accurately, tracing them back to the original click, and then continue to track them all the way through conversion.

Talk about an interesting puzzle! The fact that there is a data disconnect at the point of the first click is a MAJOR stumbling block.

So how do we do this without having to resort to a paid service? And still make use of some of the url shortening services that are already out there for Twitter users?

Well, I don't want to spoil the plot for you, but I will give you a heads up that it can be done. And once you have the basics set up the first time, it's not something that's difficult to do daily. So let's get on to it!

Here's what we'll be using to track clicks and conversions:

1. Google Analytics (hey, it's pretty good and best of all free!)
2. One of the avaiable url shorteners like Tweetburner (You can use any of these services, not just Tweetburner. The trick is to pick one that integrates well with the Twitter client you use. I'm using Tweetburner because it's what my friend already used and because it works well with his twhirl Twitter Desktop Client.
3. A little magic that was intended for Adwords advertisers, but can be used by anyone with a Google Analytics (GoAn) account.

The Beginning

So in the very beginning we're going to create a web page or make a post on our blog that contains some juicy information, something that people will want to read about. Pretty standard stuff here. Millions of people post stuff like this every day.

The only trick, if you want to call it that, is that you usually need to find and use a URL shortening service if you're going to include these URLs in your Tweets. The reason being that Twitter and similar micro-blogging platforms place a limit on the number of characters you may include in any message. So the shorter the URL the better, as a general rule.

Most people who already Tweet are going to be well aware of this fact, so I'm not going into it in any depth. Suffice it to say you need to keep your URLs as short as possible, so you'll want to use a shortening service.

The Analytics Part of the Equation

So now we have a couple of tools that will allow us to make our Tweets and shorten the URL address. But we still don't know how to track the clicks.

Enter Google Analytics. (As stated before, this same sort of concept would work with many other stats programs, but GoAn is free, quite good and has everything we need built into it already. Namely URL Tagging.)

First, a quick word about the URL Tagging feature of Google Analytics. URL Tagging was originally devised to allow Adwords advertisers to better define and determine which of their ads or ad groups were sending traffic to your web site. It's a pretty cool feature for Adwords advertisers so that they can test and tweak their ads, eventually leading to better quality traffic. And if used properly, traffic that converts.

We're going to be using a few of these URL Tagging fields to help us define our Tweet links so that we can tell what we're looking at when we review our Google Analytics stats.

There are a couple of ways you can get your URL Tagging up and working for you.

Manual URL Tagging

If you want to do it the manual way, after you've made your blog post --or created whatever page you're going to promote via Twitter-- you can use Google's URL Tagging Builder tool. What you do here is plug in the URL address of the page you want to send folks to in the top section. Then, since we don't need all of 'em, we'll put some info in each of the required URL Tagging fields. Namely, Campaign Source, Campaign Medium and Campaign Name.

You can put pretty much anything you want into each of these, but what I suggest is to use something nice and generic for some of them and possibly something more specific in one or two so that we can gain a finer degree of control. You'll also want to avoid spaces in your tags so that you don't worry about URL Encoding issues. As you'll see in the screenshot below that is an example based upon this post I've chosen to put Twitter as the Campaign Source and Medium, and used a Campaign Name of TwitterAnalytics01272009.

FTR, the Campaign Name is just a visual clue I'm setting up for myself so when I see it in GoAn I'll know it was a Tweet I sent out on January 27, 2009 and the subject of the Tweet was this discussion about enabling some web analytics tracking capabilities for Twitter induced visits to this page. A screen capture of using the URL Tagging Builder is below.

You'll see this produced an incredibly long URL address of:

http://www.randycullom.com/chatterbox/archives/2009/01/analytics_and_t.html?utm_source=Twitter&utm_medium=Twitter&utm_campaign=TwitterAnalytics01272009

This is the main reason we need to use a url shortening service. The tracking url itself ends up being almost as long as Twitter allows for your entire message! So head over to Tweetburner or whichever URL shortening service you use and plug in your Tagged URL to get the considerably shorter version.

In my case Tweetburner shortened this exceedingly long URL to:
http://twurl.nl/8zv28p

Much better!

Okay, that's all for the manual way to create Tagged URLs. If you don't Tweet URLs much, the manual method might be easiest. In which case you can skip the next section. If you Tweet very often though you'll want something a bit more automated. Thus I'm including a much easier and faster way to produce these Tagged URLs.

And the easier way is?

A Bookmarklet

Hopefully you're already familiar with what Bookmarkets are, because that explanation is outside the scope of this post and I don't want to take the extra space explaining it. Basically though, bookmarklets are relatively small bits of Javascript code that you can drag to your browser's Bookmarks that then do something when you click them while on a web page.

In the case of this bookmarklet it processes a bit of Javascript to automate a few things. It...

1. Grabs the URL of the page in your current browser window.
2. Pops up a little message to let you set the Campaign Name.
3. Produces a Tagged URL that Google Analytics will understand and process.
4. Sends it all off to Tweetburner, which in turn gets you the shortened url address to use in your Tweet.

The text version of the Tweetburner Bookmarklet is:

javascript:var%20tweetletCampaign=window.prompt('Enter Campaign Name','ChangeMe');var%20tweetlet='http://tweetburner.com/links/create?url='+escape(location.href);if(location.href.indexOf('?')>0){tweetlet+=escape('&')}else{tweetlet+='?'}tweetlet+=escape('utm_source=Twitter&utm_medium=Twitter&utm_campaign=');tweetlet+=escape(tweetletCampaign);window.open(tweetlet,'_blank','width=500,height=320,scrollbars=yes,resizable=yes,status=yes,toolbar=no,location=no');tweetlet.focus();

Or if you want a version you can click and drag to your bookmarks: Tweetburner Bookmarklet

As mentioned above, this same concept will work with all sorts of url shortening services. Though you'll have to tweak the above bookmarklet to work with other such services. Also, most of these services will provide you with additional click data if you sign up for an account at whichever service you choose to use. The above Tweetburner Bookmarklet has been tested in Firefox 3, IE6, and Chrome. If you use another browser and have difficulty with the bookmarklet let me know and I'll try to sort out any problems. No promises, but I'll give it a whirl.

FTR, don't bother asking about IE7. MS overreacted to a nefarious bookmarklet that exposed a bug in IE some time ago and put some security measures in place that defeat the usefulness of bookmarklets. It's not worth the effort to get bookmarklets working in IE7 IMHO.

If you use Cli.gs, Snurl.com, TinyURL or Is.gd as your URL shortening service I've created a bookmarklets for each that can be found here. I'll add more of those for other services as people suggest them and as I can find time.

Pulling it all together

Okay, so if used the bookmarklet you now have the shortened URL. Same goes if you used the manual method. What's next?

Now we need to start distributing the information via Twitter. How to do this depends upon if you use their web appliance or one of the various client applications. Whichever it is in your case, write your message and drop in your shortened URL, then send it to the ravenous masses.

Whichever method you choose, it will automatically get Google Analytics collecting the info you want. And not just from those who see it in your Tweet either. If they share the URL with others (via tweets, in their blog, in email, whatever!) the hits will all still show up in your GoAn as TwitterAnalytics hits. Though in some cases you may also see a referring URL.

Your done with setting up the basics of incorporating web analytics into your Twitter usage. Many will need do nothing else, and will automatically start receiving the enhanced data. It just happens.

That said, the following questions beg to be answered: Where does this show up in Google Analytics and How can I make use of it?

Well, the first place to look at, if you're looking for just basic information, would be in the Traffic Sources > Campaign area. Because we've set the Campaign Name on our Tweet links, Google Analytics will now break these out just as if they were individual Adwords campaigns. You can break out this data by choosing Campaign in the Dimensions drop down, if it's not already selected. Essentially this data can tell you which of your tweets struck a chord best with your Twitter followers. Which tweets had the desired effect of producing a click to your page and which didn't for one reason or another.

That's the basic information available, without having to do anything else. And while it's pretty basic it can also be quite valuable. Think of it as instant feedback telling you what your followers want. Sort of like having them take a survey, but without them realizing they're taking part in a survey.

If you're an e-comm site and you want to get more elaborate tracking data you can also set up Goals and Funnels in Google Analytics. These would allow you to tie conversions back to the original hit on your site, essentially giving you not only a way to see how much traffic a certain tweet brought to your site, but also a way to see how many sales/conversions result from your Tweeting.

Say what?

Yes, you heard that right. Now that we've removed the problem of the data disconnect you can treat Twitter activity (or any other social media platform) as a wholly unique traffic channel, and trace all of your hits and actions that resulted from those hits. If you wanted to you could drill down to minutiae in your social media activity, turning it into a true marketing channel where you can test things and start better meeting your follower/customer/visitor expectations.

You can of course also utilize GoAn's Custom Reports and Advanced Segmentation tools to bring forth some incredibly helpful data. Both of these services are still in Beta, though I understand they've rolled them out to most accounts at this point. Experiment with them and I think you'll find the reporting capabilities quite useful.

In fact, if you're into doing some conversion testing and have a good testing tool you could even show different content to different users who visit via your Tweets (triggering it on the Campaign Name or Campaign Source if you wanted to get really crazy) to track the same message going out to different social media traffic channels.

This post is already way too long so I'm not going to go into all of the possibilities of how you can utilize Goals and Funnels in Google Analytics or a good conversion testing tool to alter the delivered content. Just know that if you can dream it as part of your marketing campaign, you can probably do it!

I would encourage you to do experiment with these possibilities if you're Tweeting things and pointing people to pages that contain a conversion goal of some type. Given the length of this post you might think this would be a difficult and time consuming thing to do. But it's not. Here are the basic steps, from start to finish:

1. Create your landing (or blog) page to which you're going to be sending traffic.
2. Create the pages in your conversion funnel and/or goal completion pages as normal. Or if they're already created, as will be the case most often, simply link to these pages from your landing page.
3. If you haven't done so already, log into your Google Analytics account and set up your Goals and Funnels, per the instructions from the link above. You'll need to add a bit of GoAn code to your Goal Completion page to tie it all together.
4. Use the manual Tagging method or the Tagging Bookmarklet to get your special URL address you'll be pointing people to.
5. If you're using the manual method, shorten your url with Tweetburner or some similar service. If you're using the Bookmarklet method you should already have your shortened URL.
6. Tweet your message and shortened URL out to your followers.
7. Sit back and relax for a couple of days to let the information start flowing into Google Analytics. After a few days review the data, interpret the response and design a new campaign to see if you can improve the conversion response.

There ya go. With a few simple steps you can now not only track your Tweet hits fairly accurately, you can also use your Tweet data to conduct some quick response conversion testing.

How's that for cool? ;-)

You can ping this entry by using http://www.randycullom.com/chatterbox/mt-tb.cgi/48 .

My God man, why haven't you written a book or 10? This is fabulous. If one is going to tweet then one must tract those tweets indeed! Thanks for this easy-peasy tip. Twurl this! is now in my toolbox.

ROFL

Who's to say I haven't written a book or 40 Karri? After all I think I have something over 15,000 posts over at Jill's HRF, and you know I can't answer ANY question in under 2,000 words.

Of course writing books and publishing them are two entirely different things, isn't it?

Glad you enjoyed and got something out of the Tweet Tracking diversion I went on for a couple of days there.

PS Remind me to tell ya about a new little thing Diane (Torka) and I are getting ready to kick off soon. I think you'll like it. ;)