Let’s begin our irregular series of articles explaining how to do things with a simple one: how do I track page loads?
The underlying questions your friendly marketer is asking include:
- How many Visits and Visitors do I get (to a certain part of my site)?
- What are the most popular pages on our site?
- Where do people go after they landed on a specific landing page?
- At what stage of the checkout process do people drop off?
- Does anyone ever look at our Italian pages?
- Do people read news and sports articles in one visit?
A long list of questions, possibly never ending. Your friendly marketer will ask these questions. You may work with an analytics team that tries to answer some of them and come up with even more. (Note: if they do not come up with more interesting questions, they are missing out!)
But don’t despair!
All of the above can be answered if you only track the pages of your site when they load.
Now you’ll need to track some more than
s.pageName to be able to answer non-trivial questions, but there are only three basic techniques for how you track them:
- Track server-side
Let’s look at the pros and cons of each of these techniques.
What are the advantages and disadvantages of tracking like this?
- Tracked in browser — catches reloads, not thrown by caches or when people load saved pages. In other words: Rather than looking at what your server delivers, you can see what people consume.
- Captures some automatic data like OS, browser type, screen sizes. This is mostly technical information.
- Plugins — you can use the library of available plugins as well as write your own. Makes the live of an implementer so much easier.
Not so good:
- Tracked in browser — everything said above applies here as well. You are tracking where it matters.
- Size — no libraries to load, no big scripts, just an <img> tag with a URL. Loads nice and fast.
Not so good:
- Tracked in browser — everything said above applies here as well. You’ll never get 100% accurate data.
- No plugins — no libraries to load, no big scripts, and no plugins! You’ll have to code all fancy tracking yourself. Using one of the server-side AppMeasurement libraries can alleviate that pain a little bit.
You could argue that Redirect tracking falls into this category and I would agree. The pros and cons above apply to Redirect tracking, plus you can track things with it that you otherwise could not.
You can send data into Adobe Analytics via the “Data Insertion API”. You basically send tracking calls to a URL as you would normally, but you send them from your server rather than the browser.
See the online documentation for an introduction. We will also pick up this topic at some stage.
- Accuracy — because your server sends the tracking, you can get very close to the numbers in your transactional back-end systems. It is not uncommen to get to an error < .5% using this method. The business users love that!
- Stealth — can track things without your visitors seeing it.
Not so good:
- Stealth — can track things without your visitors seeing it. Yes, I think that is not a good thing. Be transparent, your customers want you to be!
- Plugins — you’ll have to rewrite any plugins that you’d want to use.
- Some things can not be tracked — things like “how far did people scroll down the page?” or “which form field was the last they filled before they walked away?” or even “how far into the video did they stop watching?” just can’t be done without access to the browser.
- Cookies / State — you can’t track anything that depends on cookies only. If you want to handle state, you will have to build it on your server.
Firstly, you are reducing the dependency between site functionality and web analytics. Release cycles won’t get in the way as much and marketers will be happier.
Secondly, you can use the functionality provided by your TMS to add very powerful customizations to your tracking.
The technical limitations of cookies are an issue, of course, as are the feelings people have about them. But there is nothing else out there, currently, so we still use them as far as we can.
So, how should you track a page?
- If you need accuracy better than 5%, your only realistic option is to use server-side tracking. I would suggest you go for the hybrid approach here and only track the really important bits server-side.