Getting Good With Google Analytics

Google Analytics

Where did it come from?

Web analytics initially resided on the server, which returned the http requests to the clients browser.  There were problems due to security and storage requirements.  Now a days, Google Analytics and more modern analytic packages use embedded javascript on pages to let the Google data servers know details about the traffic.

The way modern Javascript based analytic packages work, it takes about 3 hours for data to flow from the page to the server to the reporting interface.  Now a days, we have real time tracking.  Google Analytics is superior in 3 primary ways:

1. Accuracy
2. Cost – no additional IT resources for your company
3. Global – We don’t have to worry about tracking visitors, Google does this around the world, so our server is faster

Views

It is often recommended to have many different ways of looking at your data.  Views allow for this utilization.  It is often useful to have a few different types of views, including;

• Primary View – The view you use when working with data
• Unfiltered/Raw Data – A failsafe so you have access to all your data
• Test/Sandbox – A test view where you can try different filters without corrupting anything important.
• Only PPC – Segment out Pay Per Click Traffic

Defining and Comparing Date Ranges

By default, our dashboard data over time graphs show the previous week.  If we use a very large time period, change the graph by to week or month to show smoother data.

We can compare date ranges with the date picker.  Select the checkbox for compare to past to show how your data compares to what you recorded during previous periods or years.

Annotations

Analytics can tell us what happened, but as analysts, we must divine the why.  This requires context and background.  To help with this, we can document large events or changes in our analytic package with annotations.

To find these, select the drawer button at the bottom of every dashboard graph and you can create new annotations or view your existing ones.

We can make notes either shared or private.  They can be used to show when new goals are made, when Adwords was started, any major change.  It is useful for providing context and meaning to the data to help future analysis.

 

When working with Google Analytics in order to assess the current state of your website search engine optimization, you need both Google Analytics and Webmaster Tools (Search Console).  These tools give you tons a data with which we can assess your marketing.

Which Metrics Matter?

• Ranking
  • We want to know where we rank as compared to our competitors
  • We also want to know our search market share
    • use Webmaster tools to find this
• Links to Your Site
  • Referrer Links
  • Most Linked To Content
• Organic Search Traffic
• Bounce Rate
  • watch the bounce rate for specific landing pages
    • if a popular page has a high bounce, tackle this immediately
• Page Load Speed
  • This is a ranking factor that is especially important for mobile
• Indexing
• Crawl Statistics
  • From webmaster tools
• User Experience (UX) Segmented by Device
  • UX is a ranking factor, and is especially important for mobile
  • determine this using average time on site by device
• For competitor research, you’ll want to use AdWords tools

Ranking

Find ranking information in Webmaster Tools.  Go to Search Traffic tab, then Search Analytics.  Make sure to toggle position.

Links To Your Site

Also found in Webmaster Tools under the Search Traffic tab

Organic Search Traffic

This is found under Google Analytics, in the Audience tab.  Click on Overview then add segment to separate out organic traffic and mobile traffic.  Using these reports, you can see how different users interact with your website.

Segmented User Experience

To find the time spent on landing pages, per device, go to Behavior > Site Content > Landing Pages.  We then want to segment this by device, giving us information on how different users interact with landing pages.

Indexing

Head to Webmaster Tools and Select Google Index > Index Status

Crawl Statistics

Webmaster Tools > Crawl > Crawl Errors.  Go here first to determine if any technical issues are preventing Google from working on your site.

Search Engine Segments

You will also want to know how different search engines are interacting with your website and sending traffic.  We’ll want to work with a bunch of data, so expand the date range to atleast 6 months.

Goals

There is no point in using Analytics if you aren’t watching for certain goals that you wish for the users to complete.  These can be signups, video views, purchases, etc.

Custom and Advanced Filters and Views

Filters are used to either exclude or include data sets from your current view.  They are very powerful and can be used to slice up your data into smaller chunks for you to analyze.

Create Views just like you create goals, and take a look at the Advanced radio button to drop down some really powerful options.  We can now extract and combine data fields into a filter that provides a single output.  Check with Google to see the wildcard characters and syntax used for these extract and combine commands.

Custom Dimensions and Metrics

Google Analytics is based on dimensions and metrics.  In fact, there are over 400 dimensions and metrics that can be used when creating custom reports and advanced segments.

Dimensions describe the data.  They are categories of information.  They usually answer the question “what” that we ask when looking at a report.  When we want to know “what page is the most popular”, the page is the dimension.

Metrics are then elements of that data.  Metrics are the numbers of your data.  So, if we ask “what page is most popular,” the dimension is the page and the metrics are the data associated with that page, like how many times it was viewed, for how long, and by whom.

The All Pages report gives you different dimensions (pages) that are on your site, and some metrics (data) associated with the pages.  However, not everything is offered by default.  For example, how many of the page views were by logged in users?

The best way to understand how all this works is to use the Google Analytics Debugger (Chrome Web Store Link).  If you inspect a page and look at the console, you’ll see a wealth of information pertaining to the page.

You’ll see language, screen colors, page title, viewport size, etc.  These are all dimensions of the page.  Take special note of the cliendId, or Visitor Id.

The Visitor id is a unique ideantifier that GA uses to track you throughout your session.  It is also a Primary Key that is used by GA to connect different scopes.

Scopes

You’ll use these to track anything that analytics doens’t automatically track.  We can use custom dimensions to track the most popular author on a blog.  AN ecommerce company could track the ongoing value of a customer, beyond just the initial purchase or signup.

We can track how people whom have completed a contact form, or leads, interact with the site as opposed to those that dont.  You can track how often newsletter subscribers come back to the website.

When you create your new dimensions, you set a name and a scope.  The scope defines whether your custom dimension will be applied to the hit-level, user-level, session-level.

• Hit level is also called Pageview and it consists of data related to each individual pageview
  • An example would be the language of the page.  Google sends this data (metric), which is then scoped to the hit, so each page has a language.
    • language is a hit scope metric
• User-level is all the data associated with a particular user ID (provided by Google Analytics).
• Session-level is the group of pageviews that happen within a 30 minute window
  • An example would be bounce rate, which is the percentage of sessions with only one pageview within the 30 minute window.
    • Bounce Rate is a session scope metric

Use the Google Dimensions Explorer to better understand dimensions and metrics in GA

Weighted Sort

Cartain times, the dataset you retreive will contain insignificant data for you to use.  for example, if you sort Landing pages by bounce rate , you will get some dated for sessions of only one item. It’s difficult to going too much information from a single session so choosing waited from the sort option allows you to only show data from significant events.

Where did sort isn’t available for certain metrics like average session duration, in this event you can export the data texell and be able to filter and sort to your hearts content.

Using Google analytics to understand user intent

Google analytics office great tools for understanding user intent . They located in the behavior panel um I’m doing behavior flow for cohorts.

Another great tool we can use instead kwari our site search. Whenever a user searches for anything on our site, we will show analytic dated for that search. This is turned on in the admin panel . You wanna go to view and then select site search to turn it on.

What to measure?

What about the wedding site search look for which keywords in such phases are being used. It is also important to know whether there are any patterns or top performing searches.

We also want to know a pond which page are users are searching. that is, we want to know the page this searching from.  Could we improve that page to answer the question? Or is there anything else we can put on the page to avoid that search?

We want to know whether any of the search is on the site that any conversions on the site as well. How do users who search different from users who do not search as far as conversion?

What is the page depth? that is one the user searches, how far down the search results do they go before they click an item.

What is the bounce rates on search pages? this tells us, how satisfied The user is with what they find on the search page.

When people search on our site, what are they searching for? What was their reaction to what they found. Did they do what we want or do we does that desire different Outcome.

A neat trick that you can take advantage of if you have a WordPress website, is that you can look at search behavior a different way. would you wanna do is go to behavior and all content and you want to choose as a secondary to mention the previous page. Once this is done into the search query parameters or quit pramit are in the search box to then refine down the data to only query string if you were press website the query parameter is usually S equals.

Google analytics plugin for Chrome

Used to Google analytics plugin for Chrome in order to better visualize what people are doing on your website. Windows then start in the Chrome Web Store, you are able to view detailed statistics on Exactly what people are clicking. Furthermore, you are able to segment this data to look at mobile traffic returning users in all different sorts of segments.

Contact grouping

Google analytics allows you to group content in order to view statistics just for that grouping. the grouping is of course yours to choose. It would work well in he, setting if you have multiple departments like men’s or womens Department. Do you also work well on for black if you have multiple different types of content I’m like businesses versus users versus my cousin category of Content that you are writing.

Contact groupings are in the admin panel. you want to never get there and choose the view you like to you. And then cutting grouping is one of the tabs below the view. You can install can’t in groups in one of 3 ways. You can install a modified tracking code on the page is that you want to group. You can use extraction Which creates worlds best on the yourell of the page. for instance if you have a WordPress site with everything in categories, you could use extraction to pull a full category in to a group.

Leslie, you can find your own custom rolls, but that’s pretty powerful and can lead time predictable results.

Regex

These are regular expressions that we can use in conjunction with Google analytics to set up filters , goals, segment, audiences.  You always want to keep your regular expressions as simple as possible.

Regex Metacharacters

• \ – This works as a bridge between regular expressions in plain text. the backslash is the most used character in regular expressions. It says that what follows the backslash is to be interpreted literally and not as a regular expression character .
• | – This is the other or character. the vertical pipe can be used to separate expressions with either or logic
• ? – This is a little more abstract. Do use a question mark if the preceding character might not be there . For instance, if you want targeting a word that can possibly miss be misspelled, You can use the question mark to denote that letter might or might not be there.
• () – These are used to match the incose characters in that exact order.  Parentheses can also be used group expressions as well.  The match makes sense, the grouping doesn’t make as much sense.  If you have multiple characters that might or might not be there , you can still have the using multiple question marks you would group there’s multiple characters with apprentices. And then add a question mark at the end that’s making the containing characters optional .
• [-] – Square brackets matches any of the characters in any order anywhere in the string. The hyphen then create a range of characters within the brackets that you can match to string .
• . – The duck matches any character on your keyboard even the whitespace .
• + – The plus sign matches the preceding characters one more time
• * – It looks confusing, the stars East match the preceding character zero times or more.
• ^ – Used to match any adjacent characters at the beginning of the string .
• $ – The dollar sign is used to match the Jason characters at the end of the street  .

Advanced Reporting

If you go to the Customization panel, you get access to unique customization tools.  Click the dashboard link gives you access to your Google Data Studio 360!  I’ll study this in-depth and write back with what I learn!

Dashboards

We want to show the members of the team the data they need to make decisions.  We can do this with custom dashboards that can show different dimensions to different users.

Custom Team Reports

We can set up new custom reports, or we can search the gallery of reports and edit them to meet our needs.

Site Linking or Cross-Domain Tracking

If you session path carries the customer over multiple domains, you will want to link those domains to track your traffic as you bounce from one to the other.  To enable this, modification must be performed on your Google Analytics account.

CDT is also great if you use a lot of sub-domains.  Also great for 3rd party shopping carts like PayPal.  Use them as well with iFrame content that is located on other websites.

Follow these steps to easily perform the modifications: Google Analytics Cross Domain Tracking Help File.

Fault Finding

Use the comparison tools for comparing periods of time to get a stronger understanding of how your clients and customers habits have changed.  When trying to understand why things are happening, comparing to past periods is useful.

Conversion Rate Analysis

Which metrics are important?

• Page Timings – High Bounce rates are often correlated with slow page loads
  • usually look to optimize website and images
• Referrers – Good referrers have good conversion rates.  If your referrer isn’t converting for your, then its often the bounce rate.
  • Time on page is also important for referring websites.  It helps you know if your content was a good match for the referral.  Make sure you have big data to get good averages.  Time on page is a quality metric.
• Exit Pages – Compare the rates of exits on specific pages with the bounce rates for the same pages.
• For PPC, you want to know your best converting keywords and have the highest ROI

Page Timings

To find Page Timings in Analytics, you’ll look at the left menu > Behavior > Site Speed > Page Timings.  The dashboard shows overall data, but looking at actual pages is whats important.  The table below shows which pages are most popular, then the right-most column shows how long those pages load compared to your average.

Also be sure to segment out your mobile and tablet traffic to see if your responsive coding is providing a consistent experience when compared to the desktop version.

If you want to dive deeper into why a page is slow, look to the left menu and select Speed Suggestions, which will perform a pagespeed insight on the give page and provide recommendations to improve the performance of the page.

Referrer and Bounce Rates

To look at referral data, head to Acquisition > All Traffic > Referrals. You are able to sort your referral channels by bounce rate to see which ones aren’t performing.  Additionally, you can use the tools at the top-right of the data tables to offer different visualizations.

You’ll also want to look at Source / Medium, also in the Acquisition tab.  Because some referrals are mis-coded as direct traffic, these more general numbers are more accurate.

The problem with bounce rates is that they can be caused by many reasons.  Sometimes, its better to look at Average Session Durations.  These are a better quality score.  If you add a secondary dimension to your data table and search for Landing Page in the Behavior tab, you’ll see exactly where your referrer is sending traffic, and what that traffic is doing on your page.

Exit Pages

Found in Behavior > Site Content > Exit Pages. You’ll want to look at the relationship between exits and page views.  Exits are expressed as percentages on each page.  Make sure to segment your traffic by the non-bounce users to get a look at what people who actually wanted to interact with your content, and how long they did.

PPC Traffic Analysis

When working with Adwords, it’s super easy to use Google Analytics.  There is it’s own tab, giving you a wealth of info about Adwords traffic.  It is most useful to look at these reports in Analytics because we get more metrics thanks to GA.

For other services, like Bing or LinkedIn, the traffic will be put into many different buckets, making tracking more difficult.

A/B Testing

We can do A/B testing with Google Analytics and Content Experiments.  As a side note, when learning about these, I ran across https://www.google.com/analytics/optimize/features/, which is Google tool for testing design optimizations.  We’re not going to go over it right now, but I thought I’d let you know about it!

To create a content experiment, you’ll want to go to Behavior > Experiments.  We measure the effectivness of our experiment against converstion goals, so you have to have some of those already set up.  Then go to Create Experiment to get started!

For this to work, you’ll need separate landing pages, each with their own URL.  Enter the original page into the Configuration panel, followed by the other variant pages.

You then need to add the code to your page in order to authorize the experiment.

Multichannel Funnels

Analytics attributes conversions to the last action the user took before conversion.  The last page they saw gets the conversion. The last search they performed gets the conversions.  The last ad they clicked on gets the conversion.  This isn’t a realistic scenario.

However, all the previous data is still stored and cataloged.  The Multi-Channel Funnel Report shows the entire journey of the customer, from their initial interest to the conversion.

Go to Conversions > Multi-Channel Funnels.  We have a lot of data, so lets dive in!  Take a look at Top Conversion Paths.  These are all the sequences that have led to a conversion.  Make sure to look at the top of the page and expand the date range to get more data for your review.

Click on Assisted Conversions to take a look at how each channel group assisted with conversions of your site.

Time Lag will show you the number of conversions that occurred over multi-day sessions.  This can be interesting to know which users keep coming back.

Check out Path Length to see how many conversions you got for long paths vs. short paths.

Attribution Modeling

To get started with Attribution Modeling, go to Conversions > Attribution > Model Comparison Tool.  You can take a look at the different attribution models available, and ever compare up to three at once!

Here are the default Attribution Models:

• Last Interaction – Attributes all credit to last channel
  • Use it if you campaigns target customers at the point of purchase.  These would be useful to transactional websites that have a short decision making period for customers.
• Last Non Direct Click – The default model, it gives all credit to the last touch point before the conversion.
  • This tool is handy for divining conversions from new customers.
• Last Adwords Click – 100% of conversion is given to the last ad they clicked before conversion
  • used for tracking most successful adwords
• First Interaction – All conversion value goes to first channel.  Its great for new companies who want to build awareness.
• Linear – Conversion value is distributed evenly over each step in the conversion path.
  • Useful to maintain contact and awareness with customer, so all touch points are equally important
• Time Decay – Attributes majority of value to the most recent touch before conversion.  By default, it can track up to 7 days before the conversion.
  • used for short campaigns
• Position Based – Majority of conversion is assigned to first and last touch point.  Usually, 40% goes to first and 40% goes to last.  The remaining 20% goes to the interactions in the middle.
  • useful if you value the first touch point and the last touch point as important

We can also create Custom Attribution Models around rules that we specify.  If we have assumptions that we want to test about our conversion funnel, this is where we do it!  Click Create new custom model to get started!

First, name the model and choose a baseline to take a default model to start from, and then edit.  You have a couple more toggles and you can create custom rules at the bottom, then turn it on and use it for your reports!