3 changes that will re-make your analytics166 views
by Andrew Edwards
It’s 10 p.m., do you know where your tag container is?
That’s one of the three things you need to address to improve analytics right away.
Many organizations have lots of tags already on their sites before they get around to working with their analytics tools. Tags from advertisers, optimizers, internal bots, old implementations, and somebody’s pet experiment all crowd the page.
Oftentimes no one is brave enough to start pulling them out, lest they loudly hear from someone who really needed the data that tag was collecting.
This makes for a glut of tags on the page. How often have you gone to a site and, despite your otherwise adequate bandwidth, you watch as the page loads umpteen objects before rendering the content you came to see? Usually these are the myriad tags loading, and they can slow down page load considerably.
It’s likely that you will just leave the site entirely before the page loads, because you are too tired of waiting. Conversely, perhaps the link you wanted does load and you can click through to the next page (while hoping there won’t be too many tags there).
From a user’s point of view, this is a slight annoyance, but at least you can still get to where you wanted to go. However, from a data collection perspective, this could be a real destroyer of accuracy.
1. Tag management
Before recruiting developers to manage your tags, ask yourself:
- How often have you needed to target data collection to make decisions in the near term, only to find that developers are not going to place any tags until the next cycle – which is several weeks away?
- How smart is it to rely on developers to place every tag as needed and then set rules for when each tag fires, especially if data collection is not high on their list of priorities?
Instead of relying on developers, move to a tag management system (TMS). This way you only need the developers to place the tag container once, and you or your agency can manage tags as needed in a more user-friendly environment. You can maintain tags with ease in an environment where the content is always subject to change.
Additionally, you may define tags and allow them to fire asynchronously – kind of like having a string of Christmas bulbs that doesn’t completely go out even if one light burns out.
2. Tag placement
If your tag-object – let’s say, for a tag management module – is situated too far down the HTML page, it may not load in time to count the page view before the user leaves. This results in a zero page view, even if someone did come to that page.
You need to move tags up the page in order for it to load quickly and successfully, or else the advantages of having a TMS will be lost. The best practice is to put tags right after the header.
Some organizations continue to resist this in vain, hoping to accelerate the speed of the page load by burying data collection. Accuracy is a high price to pay for a small page load benefit. So if you value data, move your tags up the page.
Another thing that can influence data collection accuracy occurs when pages loads at a rate that lets the user scroll or even click away. In this instance, you may want to set these events to be “non-interactive” so that they don’t register as user activity.
Failing to do this leaves you vulnerable to inaccuracy in a different way, as the event will take the place of a page view and ultimately skew your page view counts. Again, moving your code to the top makes this eventuality less likely.
3. Data layer use
A TMS work best when paired with a data layer. Data layers are used to define events on a page within snippets of code that are then sent to the tags in the TMS.
While this can be a challenge at first for developers, it creates an environment of enhanced accuracy. This is far more reliable than having to scrape activity data from page actions without any direction. Your data layer becomes a carefully defined conduit for feeding complex data to your tags and ultimately to your reporting.
Intrapage events such as form submits, video views, downloads, and cart updates are especially suited for inclusion into a data layer. Without a data layer, you are asking your tags to listen for and notice things without any real guidance from your business strategy. All too often, this results in misleading data.
Accuracy drives adoption
How often have analytics projects been put off to the side because the analysts don’t trust the data? Frequently, inaccuracy occurs when tag placements and data layers are overlooked and tags are mismanaged.
Getting serious about data collection and accuracy will drive internal adoption towards making an impact on your business. Therefore, it is imperative to evaluate these three elements and make any necessary adjustments to your TMS management process, the tag placement strategy, and the focused clarity of data layers.