Digital Analytics
Digital Analytics
Digital Analytics
According to one of the pioneers of digital analytics, Avinash Kaushik, digital analytics can be
defined as…
“Digital analytics is the analysis of qualitative and quantitative data from your business and the
competition to drive a continual improvement of the online experience that your customers and
potential customers have which translates to your desired outcomes (both online and offline).”
With the right skills, processes, and technologies, you can uncover the vital information about how you
engage with your customers, effectiveness of your engagements and give you the data on which to
take action to improve the outcome – business results”.
Digital Analytics is, therefore, not only about the statistical analysis of data but the process of taking
that data and translating it into actionable insights that improve the experience of your customers.
Digital Analytics is an on-going process that should be run continuously alongside digital marketing
activities.
● The process of Digital Analytics can be broken into four general stages:
○ Measurement
○ Analysis
○ Reporting
○ Testing
● We will examine how these four stages of Digital Analytics can be applied to two distinct
areas of understanding visitor behavior:
○ Website Interaction
○ Conversion Tracking
However, since testing procedures are the same for Website Interaction and Conversion, Testing will
be covered at the end for both.
Website Interaction refers to monitoring how visitors typically behave once they reach your website.
Visitor interactions with the website can be described by several different statistics including,
average page-views per visit, time on the website, bounce rate (which is defined as the percentage of
visitors who view only one page of your website and immediately exit), among many others.
● By analyzing these types of statistics, you can determine the pages of your website that
are the strongest performers, as well as the pages of your website that cause visitors to
exit. Using this information, you can then take action to improve your website.
Conversion Tracking refers to the process of understanding what types of digital marketing (e.g.,
display ads on other websites) are the most successful for your business.
● To get the most out of your digital marketing campaigns, you need to understand which
campaigns are actually driving visitors to take the desired action (e.g., subscribing to a
newsletter) or to be converted into buying customers.
● Beyond understanding which digital marketing campaigns are driving the most
conversions, Conversion Tracking can allow you to determine which ads within a specific
campaign are driving the most conversions.
The combination of this information allows you not only to spend your marketing dollars where they
will drive the most conversions, but you can also understand the specific aspects of your ads that
are generating success.
Digital Analytics - 2
● When a visitor reaches a page within your website, this code collects information about
the visitor and sends it to Google Analytics.
● Google Analytics then provides a reporting dashboard, which allows you to segment
visitors and view their information in many ways.
○ Some examples of these segmentation and reporting options provided by
Google Analytics include demographics, location, and type of device, among
others.
In addition to being able to segment visitors and track basic website interactions, Google Analytics
can be configured to track what is referred to as a “Goal Funnel.” Typically, most conversion
processes contain more than one step.
Another useful measurement tool Google Analytics can provide with advanced integration is
E-commerce reporting. Rather than track the number of conversions by simply the number of
page-views of a confirmation page, a website can inform Google Analytics of the exact dollar value
of a purchase and what items are included in it.
● This provides an analyst with significantly more information about a transaction than
tracking goal conversions through page-views alone. For example, one digital marketing
source may convert a high volume of transactions, but all for meager dollar amounts. In
contrast, a different digital marketing source may drive significantly fewer conversions,
but each of those conversions could be worth a much higher dollar amount.
● E-commerce reporting allows a marketer to see this information and the ability to
analyze the value of a digital marketing medium.
Digital Analytics - 3
Overall Dashboard
Goal Funnel Visualization
Ecommerce Reports
Demographics Overview
There are also some technical limitations on how systems like Google Analytics can collect
information. Google Analytics collects information about visitors by executing a small piece of
JavaScript code when a visitor reaches a page.
● Therefore, visitors who have disabled their JavaScript will not have their information
collected.
● While the number of users who have disabled JavaScript is small and has decreased
from roughly 5% to 2% over the past decade or so, it still represents a segment of visitors
Google Analytics cannot track correctly.
Google Analytics also relies on a second technical tool called Cookies. Cookies are information that
a website can store on your browser so that it can track your information.
● This allows websites to keep users logged into a website on subsequent visits and
allows Google Analytics to determine if you are a new or returning visitor.
● Like JavaScript, however, Cookies also have their limitations. Visitors can disable their
cookies (approximately 5%), and they can also manually clear them from their browsers.
○ Additionally, since Cookies are stored on individual devices and browsers, it’s
impossible to identify visitors as return visitors if they visit your website
multiple times on different devices.
There are newer programs referred to as Ad Blockers designed to stop visitors from having to view
unwanted advertisements, and some of these Ad Blockers also block Google Analytics. Like these
new Ad Blockers, other new programs are being released, designed to stop web tracking programs
such as Google Analytics specifically.
As a result of these technical and user limitations, a study by the website http://www.quantable.com
shows that in 2016 approximately 10% of website visitors cannot be tracked by Google Analytics,
meaning that roughly only 90% of visitors can be tracked.
● This, along with the inability to sometimes associate new vs. returning visitors, can make
the data collected feel somewhat unreliable.
● To overcome this unreliability, it is recommended that rather than monitoring exact
numbers provided by Google Analytics, one should monitor trends and changes in user
activity and behavior.
Several business factors can have a major impact on how your visitors interact with your website
and the performance of your primary business goals. It is vital to understand these aspects of your
business, as they will have a drastic effect on how your digital analytics should be performed.
● For example, if your business has extreme seasonal tendencies doing a month-to-month
analysis within a single year is unlikely to yield actionable results.
○ Instead, it is more likely to be useful to compare one month's performance to
your performance of the same month of the previous year.
● The same can hold true for businesses with large variability depending on what day of
the week it is.
○ Understanding any seasonal or weekly variations in your business will give
you the correct frame of reference to perform your analysis.
It is also important to analyze the online environment of your marketplace against your chosen
metrics. This is important because almost all of the information gathered about Website Interaction
needs to be placed in a correct context to have real meaning.
● Knowing that your website is generating a specific number of page-views for a keyword
doesn’t inform you of how well you are performing compared to the rest of the industry.
● Knowing that you are getting more visitors for a particular keyword than a competitor
gives significant insights.
This type of comparative information is important for almost all website metrics, even conversion
rates. In one market converting 2% of visitors may be considered a very poor conversion rate, while
in a different market, it may be an outstanding conversion rate. While it is difficult to get these exact
numbers about a competitor, it is possible to approximate or benchmark this information by
undertaking tasks such as analyzing keyword rankings or collaborating with other industry
professionals.
Digital Analytics - 5
https://ga-dev-tools.appspot.com/campaign-url-builder/
This allows analysts to use Google Analytics to segment visitor traffic by the links they have tagged,
thereby providing powerful insights. The next page has an example of conversion tracking
measurements as provided by Facebook.
Digital Analytics - 6
Account Overview
Campaign Analysis
Ad Sets Analysis
Ad Analysis
Digital Analytics - 7
Many of the concepts, e.g., seasonality, industry-level changes, of Website Interaction Analysis, apply
to Conversion Tracking analysis.
The following are some of the most important metrics when determining if a Digital Campaign is a
positive investment:
● Number Of Clicks: This metric reports to a website owner the total number of clicks
generated by a digital marketing medium. While it is not necessarily the most important
digital marketing metric, it is still very important as it represents the overall reach of the
advertising campaign. Even if the rest of the campaign's metrics are strong, it may not be
worthwhile if the campaign's overall reach is minimal.
● Cost Per Click (CPC): The amount of money paid to a digital medium per click. This is
the most basic metric provided in conversion tracking reporting. It informs the website
owner how much they are paying per visitor the digital marketing medium is sending to
their website.
● Cost Per Acquisition (CPA): Cost per acquisition informs the website owner how much
they pay per visitor who completes a conversion on their website. This metric is superior
to CPC because it factors in that visitors complete the desired goal, rather than just
visiting the website.
● Return On Investment (ROI): Return on investment is the most advanced and is the true
core metric to conversion tracking. If the Digital Marketing medium does not explicitly
provide the ROI of a campaign, it is best to calculate it internally. ROI factors in not only
the number of conversions but also their overall value and inform you exactly how much
revenue a campaign generates. As a marketer or analyst, you should be wary of any
Digital Marketing campaign for which you cannot calculate the ROI.
Digital Analytics - 8
Two main types of tests can be used to determine if the changes are an improvement, and these are
“Split Tests” and “Time Comparison Tests.”
● A Split Test is when a website creates a new version of its content and runs it alongside
its current content. The website directs a certain percentage of its visitors to the new
content and the rest of the visitors to the website’s old content. After collecting sufficient
data, statistical tests can be run to see if one set of contents is significantly better than
the other set.
● A Time Comparison Test, by contrast, is a test in which all visitors are displayed the new
content. The new content results are then compared to the results of the previous
content to see which of them performed better on the website’s KPIs.
Unlike Split Tests, Time Comparison Tests are vulnerable to time-sensitive business changes.
● For example, if a press release came out during a Time Comparison Test, it would be
impossible to determine if the improved metrics were because of the changed content or
the press release.
● However, the advantage of a Time Comparison Test is that because the traffic is not
being split, it is possible to complete the tests faster.
In conclusion, if a website has enough traffic to gather data quickly, usually split tests are preferred,
but if there is not enough traffic, a Time Comparison Test should be used instead.
● Both Split and Time Comparison can be used for testing improvements made for
website interaction and conversion. The concept behind an A/B test is to make a small
change to existing content but leave the rest of the content and all other variables
otherwise unchanged.
○ By comparing the KPI performances for both the A (original) and B (new)
versions of the test, you can determine which of the two versions had better
performance.
○ Since there is only a small change and all other variables the same, the
differences in the performance between the two tests can be attributed to
the difference in content.
● To implement an A/B test, you will first need to select a digital marketing medium to
implement the test on. An A/B test must be run within a specific digital marketing
medium, such as Facebook, to control all variables beyond the content being tested.
○ All other advertisement parameters, from targeting the amount you are
willing to pay per click, should remain constant so that any changes in
performance can be attributed directly to the content changes.
● The concept of an A/B test is not limited to only changes in the content of an
advertisement. An A/B test can be applied to most parts of an advertisement, as long as
it is the only change occurring.
○ For example, rather than creating a new advertisement with adjusted content
and testing the content changes' success, you could run the same
advertisement with different target audiences.
○ As long as the only thing changed is the advertisement’s target audience, you
can credit any changes in the ad's effectiveness to the changes in the
targeted audience.
○ Using this approach allows you to test a large number of different
parameters and continuously optimize your advertisements.
● However, the overall budget does not necessarily need to be the same between the two
tests.
○ For example, if you are running an existing advertisement on Facebook, you
could create a new advertisement with the same targeting parameters but
allocate a smaller portion of the budget to the test advertisement.
○ The budget for the new ad should be large enough though to generate
meaningful data.
● While an A/B test's primary goal is to discover content or advertisement changes that
increase effectiveness, it is still important to examine a “failed” A/B test where the
original content outperforms the new content being tested. Often just as much insight
can be gained by determining what caused an advertisement to fail as can be gained by
simply implementing the successful changes.
In sum, the implementation of continuous A/B testing on multiple aspects of advertisements is one
of the best ways to improve overall advertisement performance.
Digital Analytics - 9
There are three primary ways to determine which sources are credited for the conversion: first click
attribution, last-click attribution, and multi-channel attribution:
The following are a few examples of these supplemental analytics tools and how they can be used to
gain additional insights about visitor interactions:
● Heat-maps are tools that create a visual overlay for digital analysts to view where
visitors are clicking within a website.
○ These tools generally highlight areas with more clicks red and areas with
fewer clicks green, similar to how radar shows the strength of weather
patterns.
○ By understanding the areas of a webpage that are drawing the most clicks,
analysts and website designers can attempt to make the more highly clicked
areas more prominent or attempt to make the less often clicked areas more
relevant.
● Follow-up Surveys are tools that ask visitors (usually a small subset of visitors) if they
would be willing to answer a short survey. While these surveys run into typical problems
encountered by traditional surveys, they can still provide valuable insight.
○ These surveys provide an analyst with information about how a visitor felt
about their interaction with the website, which can, in turn, be used to
enhance the overall visitor experience.
○ The most popular of the Follow-up survey questions is to ask a visitor if they
could accomplish their goal with their visit.
● While not strictly a tool for Digital Analytics, Live Chat gives an interesting way for
website owners to interact with their visitors. These tools allow website owners to
interact directly with a website visitor while they are on the website.
○ Many of these Live Chat tools also have the ability to create an automated
program or a Bot that can ask these visitors specific questions, and in some
instances, even answer common FAQ type questions.
■ The ability to ask questions through an automated system can
allow website owners to segment their visitors in ways that
would be impossible through other Digital Analytics means.