On-Demand Webinar Available Now | Cardinal Path Blog
Blog

10 COMMON MISSTEPS UNCOVERED IN AN ANALYTICS AUDIT ~ NOW AVAILABLE ON-DEMAND

During our recent webinar, Don’t trust your data? 10 common missteps uncovered in an analytics audit (and why you need to get them fixed yesterday), many of our audience members asked very informed questions. We want to share the Q&A portion of the webinar with you.

View On-Demand Now

 

Watch the on-demand webinar and come back for the Q&A session summary below.

Question: Does using Google Tag Manager add to my loading time and add more server calls?

Absolutely not, GTM is using asynchronous loading, which means the browser will load it as soon as possible, but will prioritize other synchronous scripts first. Tags firing from GTM are also asynchronous, further guaranteeing there will be no performance impact. It won’t add more calls either, given you are tracking the same things. On top of that, the code for GTM and when you use native tags is optimized.

GTM has been really well engineered and they’ve done a great job ensuring that it won’t negatively impact page load time. In fact, if you have many tags on your site, it will likely speed up page load times because now their following a sequenced and optimized firing process. Also, if you are using the old GA classic code, you’ll notice a big improvement simply because UA (and GTM) are just much more streamlined. In fact, there were many posts after the launch of UA in Beta that raved about how UA would “speed up the web” because the code was so well optimized.

Question: Where should the GA tag should go in the page code?

If you use Google Tag Manager (highly recommended), the container code should go right after the option BODY tag. If you integrate Universal Analytics directly on the page, it should go in the HEAD part of the page. If you still use Google Analytics “classic”, you should definitely upgrade.

Question: Where should the redirect go with respect to the GA code, and if the traffic is redirected before the GA tag is fired, will parameters from the initial link be carried over to the final page after the redirect?

The best way to do redirections is through server-side .htaccess syntax with a 301 status, and this will typically pass on all parameters. If you are coding a redirect directly on a page using JavaScript, then your code needs to pass on all the parameters (and this is by far a less preferable approach).

While the best way to do this is server side, if you don’t have that option (and there could be many reasons why this is the case), then you may want to investigate why you have that roadblock before you just simply opt for on-page scripting approach. Why? Because sometimes you get push-back when it should be you pushing back. If you are running many campaigns and spending lots of budget to acquire traffic, you absolutely want to know how effective this is. And particularly if you have an Agency managing your marketing spend: the failure to fully pass the parameters with the redirect will mean that you will not see this visit as being acquired through paid traffic. However, your Agency will be sending you a bill each month based on clicks. You’re looking at visits, they’re looking at clicks and it can get nasty pretty fast when the gap gets big. If this is your situation, push for the right solution: server side redirect.

Question: For Google Tag Manager with dataLayer can the dataLayer be in <body> or should be in <head>?

The dataLayer can be anywhere – what will tell GTM something is happening is when you add an entry with the special “event” attribute. GTM watch for those and this is how you would create a rule/trigger. When the GTM event occurs, GTM looks at the latest value of all attributes in your dataLayer and “exposes” them to your container as macros/variables.

Question: Will GTM interfere with my capture of e-commerce and transactional data – which is on a 3rd party website?

No, GTM will not interfere – view it as an empty shell until you actually add tags to it, such as a GA page view or event. As such, you will also want to use proper cross-domain configuration so your 3rd party website receives the right info. And if you are lucky, you should deploy GTM on that 3rd website too. Visit https://support.google.com/tagmanager/answer/3561401?hl=en

Question: Do you think a day we will be able to trust Digital Analytics data like we trust the databases data?

The question is “why would you need to trust your DA data as much as your database data?” Database (taken in the sense of back-office data such as customers, sales, inventory, etc.) needs to be accurate and precise. DA is in the realm of statistics – it needs to be good enough with an acceptable margin of error so you can make decisions. It actually doesn’t have to be perfect. Through an audit, we can evaluate how far is the discrepancy between the absolute truth (your back office) and what is seen on the front-end (your website).

If you properly implement (and particularly with the new features in enhanced ecommerce from GA), you should be able to have equal confidence in either data sets. To be clear: garbage in/garbage out is the rule for *ANY* data set. The issue is that it has always been difficult to control for all the variability’s in the digital environment.

Question: How do you know how close you are to the GA standard 10M hit limit?

The easiest way is that you’ll get a note from Google informing you that you are exceeding the TOS for data collection. Otherwise, you’ll have to an estimate based on the average number of hits per page (note the page counts as 1 hit; every event on a page counts as a hit; transactions count as hits; social shares [if clicked on] count as a hit per click) and multiply this by the number of page views per month. So, if regularly have 2M page views per month, and you know that you have 2 events on every page + an average of 0.5 social shares and 0.5 transactions per page, then you have 3 hits / page or 3x2M = 6M hits / month.

Question: How does the AdWords upgraded URL’s effect the tracing of AdWords data in Google Analytics?

Interestingly, there was a lot of talk in the GA Authorized Consultant community about this very thing. A number of people were reporting a lot of (not set) values in the AdWords > Campaigns report, which was thought to be a result of longer GCLIDs. Apparently, this has now been addressed, so the answer going forward should be, “it doesn’t!”

Question: How often should an org audit their analytics tracking?

It really depends on the type, size, complexity and change velocity of the site. After a first audit we can more easily determine a sound plan that will increase your confidence in the data while not going crazy with it (which would be too expensive and time consuming).

Financial statements are prepared by companies and audited yearly by external 3rd party organizations. Why? Because we need to make sure that we can trust those financial statements, particularly if we’re making investments in the company. There is no reason why, in the 21st century, where the online channels are generally the #1 touch-point with consumers, that organizations don’t view their digital data the same way they do financial data. If you have a large site that is being worked on by many divisions, or you do a lot of ecommerce then you should audit yearly. Otherwise, I’d say only as you do major roll-outs or upgrades.

Question: I added Google Tag manager code on my site and I noticed that my bounce rate dropped significantly and my page views doubled. I have since removed the tag manager code. Could this issue be a result of installing Tag Manager code in the wrong place?

From the symptoms you are describing, you likely deployed GTM with a page view tag sending to the same web property (UA#). In GTM, set a macro for your UA and send the data to a test web property while you do your tests. When you are ready, change it to your production one. There are many other useful tricks, such as using a table lookup macro so you send to your test web property when in “debug & preview” mode, and to your production one otherwise.

Question: I have been told to track every action on a page, like links and downloadable documents, forms, etc. You all say to not track too many events. What is an example of an event you shouldn’t track?

If there’s no business decision and no optimization possible, don’t track it. Especially with GTM, it’s easier to turn on/off tags so you collect data when there’s really something that can be done. It’s a matter of balance – and this can more easily be done when you start with business objectives in mind, and map out the custom lifecycle and indicators demonstrating the visitor is moving toward conversion.

Question: I recently installed GTM on my site and also noticed an increase in page views but I am using 2 different tracking codes, one for testing and the original how do I test which one is correct?

If there is a discrepancy between your testing (new) and your existing (old), some factors could be where the code is on the page, or some code not firing on some pages, etc. Use your “raw/unfiltered” view for both (so there’s no filters applied), use the same timeframe. Go in your All Pages report and download the data for both – load in Excel, sort on page name, and you should be able to do a side-by-side comparison and see which page is radically different. This is one way – or use WASP do crawl the site and see if there are any missing tags.

Question: If another agency is governing the websites analytics implementation, is there anything we can do to double-check the tracking/implementation? I know you mentioned 3rd party audits; however, if we don’t have access to the tag manager or all aspects, how would we be able to do an audit? It’s hard to trust data when not involved in set-up and knowing the capabilities of the people setting it up.

You could use WASP.inspector while you browse the site and you will easily see which tags are firing on which page, and what data is being sent. You can also use WASP.crawler to crawl the site and do a bulk check – all of this without even having access to GTM.

Select 100 pages on your site and run a crawl. WASP will look inside GTM to find all the tags, so no problem there. It will also tell you if the tags are on every page, and if so, are they consistently deployed. Any lack of uniformity in the data within these 100 pages is a red flag for what you’ll find on the rest of the site (if you were you to scale up the crawl). This is the type of data you’d bring to your agency and ask them to tell you what’s going on

It is YOUR website, it is YOUR data, and you definitely should own your digital analytics and GTM as well. If ever the agency refuses to let you at least “View” GTM, you should consider alternatives. We’ve seen agencies refusing access to AdWords to their clients – and in our opinion, this is unacceptable.

Question: Is it possible a self-referral comes from the redirect from domain.com to www.domain.com (or vice-versa)?

Yes, this is quite possible, especially if the redirection is not done correctly. It should be a 301-redirect handled in a .htaccess file. See https://support.google.com/webmasters/answer/93633?hl=en

Question: Is there a way to copy annotations from one view to another?

No, currently this isn’t possible. One approach is to keep technical notes in the raw/unfiltered view, and marketing-related notes in your production view.

Question: We are a company still working on moving to UA. What are some caveats to simply using the auto-transfer-to-UA function? I’ve heard things about needing to migrate custom events separately but it’s still a bit unclear for me. What things should I look out for?

This tool will help you identify where there’s more risk and things to consider: http://www.cardinalpath.com/your-universal-analytics-migration/

Question: What is the “best practice” way to tie domain & subdomain tracking on same website to not lose the original source? Two different profiles? One profile? Cross domain or just exclude the referral?

Simply send all the data to the same web property. With Universal Analytics, the sub-domain will ‘automagically’ be handled.

Question: Why does GA show my own site/domain as a referrer?

There can be many reasons, but the most common one is untagged pages. Although this article applies to “classic” Google Analytics, there’s a lot of relevant info: https://support.google.com/analytics/answer/3198398?hl=en

 

Copyright © 2016, All Rights Reserved. Privacy and Copyright Policies.