The vast majority of Google Analytics accounts are set up on a single profile. All the data, testing, and filters impact a single source of data. This can cause searious problems as your business grows. Before we dive into the profiles that your account should have, we need to cover how Google Analytics produces profiles.
Google Analytics does not store raw data about your traffic. By the time you see data in your Google Analytics account, it has gone through several steps:
- Your Google Analytics Tracking Code sends visitor data to Google Analytics’ Servers.
- Google Analytics’ servers compile data on your site.
- About every 3 hours, Google Analytics processes your data using your settings. Filters are applied, conversions counted, and site search is processed.
- Each day, Google Analytics dumps the data from the previous day.
- Once all the data is processed, it’s stored in a database where you can access it through your Google Analytics account.
The key part of this process is that the data you have access to is not raw data, it’s data that’s been processed using settings you’ve applied to your profile. Once the data has been processed, there’s no going back.
This is why a new goal cannot be applied to historical traffic. Google Analytics processes goals only one day at a time.
In other words, the settings we apply to our Google Analytics profiles will change the data permanently for each day that they’re applied. If you have a typo in an exclude filter, you could irreversibly corrupt a portion of your data.
To avoid such pitfalls, there’s several profiles that you’ll need when using Google Analytics.
Raw Data Profile
This profile does exactly what you think it would, it houses raw data. There’s no goals, no site search, no filters, no custom reports, nothing. The purpose of this profile is to provide a fail safe in case all your other profiles become corrupted. If you lose all your other data, you’ll still be able to pull reports from this profile. The data won’t be polished but you’ll be thankful it’s there if you ever need it.
Simply set up a profile, name it Raw Data, and then never touch it. The sooner you do this, the more data you’ll have backed up if something goes terribly wrong.
Test Profile
This is where you’ll test all your settings to avoid any accidental data corruptions.
Before applying goals, filters, or anything else to your other profiles, test them here first. You can watch the data for several days and see how your reports are impacted. Once you’re sure that everything is working precisely as it should, you can apply the same setting to the profile the setting was designed for.
Master Profile
When I’m combing through reports looking for useful insights and evaluating the success of my campaigns, I’m looking at my Master Profile. It’s the default profile on all my accounts. It’s also the most critical. This is where all my filters are applied, my goals and custom reports are set up, and site search is enabled.
Since I’m using it to do all my analysis, I want to make sure that any new settings aren’t going to skew my data and lead me to the wrong business decisions.
Once you have your Master Profile set up with whatever settings you prefer, you’ll generally want to leave it alone. If you’re going to apply new settings or tweak existing ones, run them on the Test Profile first.
Custom Profiles
Create other profiles to do whatever your heart desires. Here’s some common custom profiles to give you some ideas:
- Access-Based Profiles that limit people in your organization to only certain pieces of your Google Analytics data. Just filter out everything you don’t want them to see.
- Source-Based Profiles include traffic from a single source. This will allow you to have direct access to how that traffic behaves. This is especially useful if you’re spending a lot of time and energy on a single source for traffic (Facebook, Twitter, Youtube, etc) and want to evaluate the success of those efforts.
- Location-Based Profiles let you focus on a single location. If you’re expanding into an overseas market or focusing on a single location, you’ll want to easily see how that traffic behaves.
Aside from Accessed-Based Profiles, all of this data will be available in your Master Profile. You’ll just have to drill down to the right report or create custom reports. Segmenting everything into a new profile simply makes it faster to access what you want. If you’re going to be spending months (or years) evaluating a specific portion of your data, it’ll be worth it to set up a profile. Otherwise, just grab it from your Master Profile.
What other profiles do you use?
Hey Lars
What are your thoughts on having different properties across your different assets? I.e. Your blog, Your main marketing site and your application?
This is the way we have it setup and it seems to work ok, but was surprised to see you didnt recommend that here. Is that the approach you take at KM?
Aaron
Hey Aaron!
If you combine multiple subdomains/assets into a single GA profile (or view as they call it now), it’s very common to split them out again into individual profiles. This would give you one aggregate profile and individual profiles for each asset.
The other option is to just keep them separate with different properties (tracking codes). This is easier to setup since you don’t have separate traffic out into different profiles with filters. But you won’t have an aggregate profile that includes everything.
The direction I go usually depends on how heavily the business relies on GA data. If they use it for most of their analysis, I’ll use the same tracking code for everything and spend the time filtering assets out into different profiles. But if they use other tools for their data, I’ll keep things simple and split it up with properties.
i used “Filter: Include only Mobile Traffic” but my profile did not have any data, why this happen? how can i fix it?
Hey Lee!
Filters in Google Analytics are pretty powerful. If you set them up incorrectly, they can delete all your data. I’d double check all your filter settings to make sure nothing is being deleted accidentally.
Hi Lars,
Let’s say I have had one view (titled Master view) for years collecting unfiltered data. Then, after reading your article, I added two additional views, titled Test and Raw. Is there a way to copy the data collected in the Master view, and use it in the Raw and/or Test views? Or can views only use data that they have collected since creation, even if they are created much later than the default view (Master view)?
Hey Darren!
Unfortunately, there’s no way to copy old data into new views. Each view will only show the data that has been collected after the view was created. This has to do with how the data collection and processing works. GA doesn’t store raw data, it only keeps it on hand for a few days. Once the raw data is processed and filtered through the view settings, it’s finalized. This is what you’re looking at when you run a report. But the raw data isn’t stored anywhere. So there’s no way to go back to the raw data and re-process it with settings in a new view.
Hey Lars, great article and a sensible approach to setting up Analytics structure. If you don’t mind, I’m researching for a little 7,000 word article on common analytics errors, I might include a link back to you? 🙂
Regards, Larry
Absolutely Larry! I’d be honored. 🙂
How do you create the 3 views??
Hey Ginevra!
You’ll need to log into your Google Analytics account and then click on the “Admin” button in the top navigation. Then in the right-most column (the one that says “View”), click on the dropdown at the very top. This dropdown normally lets you select between different views. There’s also an option at the bottom of the list to create a new view. Not quite sure why the GA team put the new view link there, it’s super hard to find. 🙂
When creating the new views you should start by copying your default view for each property. A mobile view and PPC view are also good ones to create asides from the three mentioned here.
Agreed 100%, both are great tips David. 🙂
Great article! Do you suggest renaming the “All Website Data” view to one of the 3 mentioned here? If so, which one (Master, Raw or Test)?
Yeah, I’d definitely update the name Jamie. The finished structure should look like this:
For the view that your property starts with, I usually change that to “All Website Data – Master” and then add the other two views.
This naming structure also makes it easy to include other specialized views like “Mobile Data” or “Paid Data.”
Hi. this is very interesting!
Please could you advise whether it is better to copy the existing view (Master) which already has some goals etc in it when creating the Raw view?
Thanks!
Alice.
Hey Alice!
I don’t believe that there’s an option to “copy” one of your views. Each new view is created from scratch on your Google Analytics property. To “copy” a view, you’d have to create a new view and then set up all the settings again so it processes the data in the same way as your other view.
But since you’d be creating the Raw view, there’s not settings to configure. I wouldn’t edit the account in any way so that it’s truly your raw data.
Hello Lars,
Thanks for this. Just wondering if you think it worthwhile or necessary to set up a test view for each separate view eg mobile view and mobile test view. Or does one test account fine for a property?
Setting up a test view for each individual view is probably more effort than it’s worth. I’d only do it if you have a lot of specific debugging on a particular view that your normal test view wouldn’t cover.
Hi Lars!
Thanks for this great article! Complete marketing newbie over here.
Can it be that analytics changed the term “profiles”? Would it be users now or view? https://support.google.com/analytics/answer/1009618?hl=en
Trying to understand the hierarchy of the different layers…
Cheers,
Alex
You’re absolutely right Alex! Google Analytics did change the name of their “profiles” to “views.”
Hey Lars, great post! Do you generally exclude dev traffic, or internal office traffic in your master profile?
Yes, I’ll usually exclude dev traffic and internal office traffic if there are easy IPs to exclude with. But most of the orgs I work with are heavily remote and the sites are pretty large (1 million visitors/month and above) so it’s not something to worry too much about.
Hey Lars, Great article! I have 3 views on my website. Master, Raw, and Test. I’m about to start utilizing Adwords, but which view do you recommend I link it to? Master?
Hey Edgar!
I’d connect your AdWords account to your Master and Test views. Test should be a mirror of Master so you can easily check any changes you’re unsure about in the future. And for Raw, I prefer not to set up anything beyond the basic script install.
Hey Lars, great article, but I have one question. Should I enable Bot filter in the Raw data view? Also, should I enable Strip query parameters out of URL, from the View settings?
Thanks and best regards!
Thanks Harry!
I wouldn’t set up anything on the Raw Data views, it should be as raw as possible to minimize any chance of messing the data up. That’s it’s real purpose.
I’d apply those items to your Test and Master views.
Hey Lars!
I accidentally created a filter to ‘include only’ one of my office IPs to test if only that computer’s data is reflecting in “real time” section of GA. BUT I did that in “all website data” view
but then I read about why we should create new view to create filters. I’m afraid that will my previous data be impacted because of this? I immediately deleted the new filter that I created.
Do I need to do any further steps to be sure that previous data isn’t impacted. I had the filter on for about half n hour before I discovered a wrong approachand deleting it…
While the filter was live, it would have impacted any data that was sent by your Google Analytics tracking script. Based on your description, it sounds like it would have excluded all traffic that didn’t come from your office IP. So all of that data will now be lost, there’s no way to recover it.
You definitely did the right thing by removing it so quickly. Since it was only live for about 30 min, I wouldn’t worry about it. The impact on your site data will be relatively minor. I’ve personally been through MUCH bigger data issues with Google Analytics. 🙂
In the future, I’d set up a test view to try things like this.