How to Debug and Fix tracking errors - YouTube

Channel: MeasureSchool

[0]
- Do you trust the data you are collecting?
[2]
If not, in this video, you're gonna learn a framework
[6]
to ensure your tracking is working correctly.
[9]
I'm gonna give you a five-step process you can follow
[11]
to determine any problems with your tracking and solve them.
[16]
All and more coming up right after this.
[18]
(energetic music)
[24]
Hi there, and welcome to another video of measureschool.com,
[28]
where we teach you the data-driven way of digital marketing.
[31]
My name is Julian, and on this channel,
[33]
we do marketing tech reviews, tutorials,
[35]
and how-to videos just like this one.
[38]
So if you haven't yet, consider subscribing.
[40]
Now a lot of you guys have followed along
[42]
with our tutorials and installed new tracking
[44]
on your website over the past few months.
[47]
But maybe looking at the data now,
[49]
you find that it doesn't seem that accurate anymore,
[51]
or is definitely broken.
[53]
What can you do?
[54]
Well, you could investigate
[55]
and see if you can find the error.
[57]
How?
[58]
That's a question I actually get asked quite often,
[61]
so I thought I'd break it down for you.
[63]
I've come up with a framework you can use
[66]
to find tracking errors, test your existing tracking,
[68]
and get a clearer picture of your data.
[71]
Now, we got lots to cover, so let's dive in.
[74]
(energetic music)
[75]
Before we start diagnosing our tracking error,
[78]
we need to go through some preparations first.
[81]
So step one is writing down your tracking steps and results.
[86]
Before you start testing, you want to have a replicable
[89]
process to go through with a measurable outcome.
[92]
So for example, if your conversion is a transaction
[95]
in an e-commerce store and you wanna test that,
[97]
you should go through the steps of adding a product
[101]
to your cart, going through the checkout
[102]
and using your credit card or another payment method,
[105]
and then landing on the thank you page
[107]
where the conversion actually happens.
[109]
Now the results you are trying to achieve is that
[113]
the conversion actually shows up
[114]
in your e-commerce reports in Google Analytics.
[116]
That's the measurable outcome that will determine
[119]
if your test was successful or not.
[121]
Having steps like these written down will give you
[124]
a clear picture of the variables,
[126]
and determine a definitive checkpoint to go back later to.
[131]
If you are done with this, we can proceed to
[133]
step number two which is getting your debugging tools ready.
[137]
You cannot see what you cannot see.
[139]
Tracking is not directly visible to you
[141]
on the website itself, it works by sending requests
[144]
to the tools in the background.
[146]
So if any errors happened, you wouldn't necessarily
[148]
see them in the browser.
[149]
That's where debugging tools come in.
[151]
Now, it often depends which platform or device
[154]
you wanna try to debug, but for web tracking,
[157]
we have quite good tools to our disposal.
[159]
For example, if you are using GTM to deploy your tracking,
[162]
you have the preview and debug console to see
[165]
a lot of information already about your tag deployment.
[168]
There are also hosts of other different extensions
[170]
for Chrome which can make your tracking visible,
[173]
depending on your tag.
[174]
There's a tag assistant, a tag inspector,
[177]
a Facebook pixel helper, WAST, GA debugger, and more.
[181]
If you need to see the requests that are really
[183]
coming and going from your browser
[185]
and you don't mind getting a little bit more technical,
[187]
you can utilize the Chrome developer tools.
[190]
Other than that, you can also use external proxies
[194]
such as Fiddler or Charles, to see what is being sent
[198]
and received by the browser.
[200]
No matter what you are trying to achieve,
[202]
make sure you have the right tools in place
[204]
so you can diagnose any problems later.
[207]
Which takes us then to step number three,
[209]
which is actually testing our implementation.
[212]
So with all your debugging tools running now,
[215]
you can now execute the process from step number one
[217]
and go through them step by step.
[219]
Note down if you see any weird errors,
[222]
and if you can replicate your end result from step one.
[225]
If you run into any major issues, you can go over to
[228]
step number four, which is debugging and fixing your error.
[231]
And this is really all about asking the right questions.
[235]
So in our previous example, if our transaction
[238]
doesn't get tracked, I would ask myself,
[240]
what needs to happen in order for this to work?
[243]
Luckily, there is a blueprint in the Google documentation
[247]
which explains how things are supposed to work.
[250]
In most cases, transaction tracking heavily relies on
[254]
custom data that needs to be pushed to Google Analytics.
[258]
I would start investigating with the GA debugger
[261]
if this data actually gets pushed,
[263]
and if there are any errors on the page
[264]
in the markup itself that prevent that from happening.
[268]
Yes, this can be quite intensive, technical,
[270]
and sometimes frustrating, but it can really help you
[274]
to come up with a solid treatment to fix your error.
[278]
Then you can go back to your developer,
[280]
re-implement, and start tracking your data again.
[283]
Once it's working correctly, I would go to
[286]
step number five, which is cross-checking.
[289]
What does that mean?
[290]
Well, in our previous example, we did transaction,
[292]
even if they now get tracked in Google Analytics,
[296]
it doesn't necessarily mean that this data is correct.
[299]
But we can utilize external data sources,
[302]
in our case, it would be the back-end order records,
[305]
to cross-check that data, and see if anything is missing
[309]
or if it's in realms of acceptability.
[312]
And by the way, no tracking is always 100% accurate,
[315]
so don't stress too much about
[317]
a small margin of error that can occur.
[320]
(energetic music)
[320]
So there you have it, five steps to test
[322]
and verify your tracking.
[324]
A short recap, first of all, mark down your steps
[327]
and end result you need to go through,
[330]
then get your debugging tools ready,
[332]
then test your tracking by
[334]
going through your testing protocol,
[336]
and debugging and fixing any errors.
[339]
And in the end, cross-check your data
[341]
with available third-party data.
[344]
So now I wanna hear from you,
[345]
what was your most frustrating error
[347]
that you had to debug and fix with your tracking?
[350]
Please leave a comment below and join the conversation.
[354]
And as always, if you like this video,
[355]
please share it to your friends or colleague,
[357]
and subscribe to this channel
[359]
to get more of these videos every Wednesday.
[362]
My name is Julian, 'til next time.