Web Analytics
2
LESSON
LESSON SCRIPT
@Circus_St | [Link]
WEB ANALYTICS LESSON SCRIPT
LESSON 2
1. Introduction
Welcome back! Last time we introduced you to the vast world of website analytics.
2. Data Frameworks
By now you should know how to collect user data from your site, but the thought of
actually opening up your analytics tool and figuring out where all that data lives might still
seem a little daunting. We’re going to take a look at the different reporting areas within
analytics tools, and show you which sections your users’ data can be found in and – more
importantly – which questions each area can help you to answer.
We’ll start with real-time reports. These pages in your analytics tool show you live
information about visitors who are currently on your site - for example, which pages
they’re on, where they came from, and some demographic information. This is in contrast
to most other analytics reports, where the data updates on a 6 - 24 hour basis.
Live reporting allows you to continuously monitor and optimise campaigns and
promotions, and to take advantage of current trends and opportunities, like visitors
responding to news stories, product reviews, sporting events or celebrity endorsements.
© Circus Street Ltd. 2
WEB ANALYTICS LESSON SCRIPT
LESSON 2
The ability to see real-time activity on your site and the potential for fast response can
be very seductive. However, before you leap into action, there are a couple of things to
watch out for:
• Do you have enough visitors showing similar behaviour to make your conclusions
valid? Be careful about acting on the evidence of a small sample alone, since that
might not be representative of your whole audience.
• Are you really familiar with your business goals and KPIs, and what your campaigns are
trying to achieve? Fast decisions can lead to great results, but they’re also dangerous if
taken too quickly, or with insufficient understanding of the implications. Say you started
running a new social media campaign, but noticed it wasn’t performing well. You might
be tempted to cancel it - but it could be that the audience for your new campaign is
mostly active at weekends, and you just haven’t waited long enough to see that yet!
Next up, we’ll look at the Audience section. This is where you’ll find information about the
characteristics of your visitors, including whether they’re new to your site, their geographic
locations, their ages and interests, the language that they speak, and technical information
about the devices and browsers that they’re using to access your site.
Within the audience section, you can find answers to your questions about geographic
location: What proportion of your traffic comes from the area your business actually
operates in? Are those visitors converting at a reasonable rate? Do you receive significant
traffic from any other locations that might suggest you need to make a change to your
overall business model?
These reports can also help you understand the makeup of your visitors from the
perspective of age, gender, hobbies and purchasing patterns – in other words, their
demographics and interests. For example, you can see whether your traffic skews male
or female, younger or older, or what types of products they typically buy. You can then use
these insights to refine your content, brand messaging, and advertising copy, making sure
that you have the needs of your specific audience in mind.
And then there’s mobile devices. Your audience section will tell you what proportion of
your traffic accesses your site using a smartphone or tablet compared to desktop usage.
How are your mobile conversion rates? Are visitors on mobile devices actually likely to
complete a purchase? That isn’t to say that you shouldn’t work on your mobile digital
presence, but your analytics data might help you to assign some priorities when
allocating budgets and resources.
© Circus Street Ltd. 3
WEB ANALYTICS LESSON SCRIPT
LESSON 2
The third reporting area we’re going to look at is the Acquisition section. This holds
all the information about how your visitors get to your site. All of the various marketing
channels that you might use, like organic search, email, social media and mobile
marketing are shown in a report that measures the relative performance of each one.
In order to make the most of this section, there are two acquisition terms you need to
be familiar with. Direct traffic refers to visitors who came “directly” to your site, without
clicking on a link or going through a search engine. This implies that they already knew
the URL of your site or had it bookmarked, so it’s generally assumed that they are
familiar with your business. For this reason, direct visitors can be an important factor in
considering brand awareness. Referrals, on the other hand, are visitors who arrived at
your site by clicking on a link from another website. The referral path in the analytics
report shows where they came from.
Your acquisition reports will help you understand your advertising return on investment.
Obviously, if you’re spending money on advertising, it’s important to know which
campaigns and media are effective in generating conversions. If you’ve configured your
analytics goals correctly, you should be able to review both macro and micro conversions
that are associated with paid visits.
You can also use these reports to look at goal completions and engagement rates for all
channels, helping you to identify which ones are over- or underperforming. Where do
your most valuable visits come from? Are there any channels that seem to be either much
more, or much less successful than the others? If so, do you have any clues as to why that
might be? And how should you readjust your spending and priorities as a result?
The acquisition area will also show you the search queries used in search engines
where your site was listed in the results. Since Google usually shows the corporate site
in response to searches for a company name, it’s usual to see the name of your business
in the search query listings. Queries that include a company name are referred to as
“branded searches”.
By definition, visitors searching for your business name have already heard of you.
But there are probably many more potential customers who aren’t familiar with your
company. A good search engine optimisation strategy should also include “non-
branded” keywords focused around your products and services to attract this wider
audience. By reviewing your search queries report, you’ll be able to evaluate the current
strengths and weaknesses of your organic search positions for both branded and non-
branded searches.
© Circus Street Ltd. 4
WEB ANALYTICS LESSON SCRIPT
LESSON 2
And finally, there’s social media. If you include social media as part of your marketing
mix, it’s important to know which platforms produce the best results for you. Again, look
at the conversions and other goal completions, as well as engagement metrics such as
time spent on key pages. This information should help you to tailor your social media
campaigns towards your most valuable audiences.
Now to the Behaviour section. This tells you about what visitors do once they’re actually
on your site, with a focus on activities other than purchases - these are covered in the
Conversions section, which we’ll talk about in a moment.
Reports in the Behaviour section include the specific pages visited, how the site
performed in terms of loading speed, details regarding usage of your onsite search
engine, and information about any events that you’re tracking.
You might start off in this area by looking at visitor loyalty. The ‘frequency and recency’
reports tell you how many times visitors came to your site in a given timeframe. These
measures of loyalty can tell you a lot about the effectiveness of your branding campaigns,
whether your content is engaging, and how compelling your incentives for repeat
business might be.
When viewing these reports, you need to identify a reasonable expectation around
returning visitors for your specific business. For example, a news site might want visitors
to come every day, whereas an online grocery might expect visitors to place orders once
a week.
Content usage reports tell you how specific pages on your site are actually being used.
That helps you identify consistently popular content and ensure that you’re taking full
advantage of these pageviews. You can also tell which pages are rarely being seen. If
you have pages that don’t draw visitors, you might consider whether that content is simply
not interesting to your audience, or whether your site could do better in drawing them
to that area. You can also identify ‘entry’ and ‘exit’ points – the pages where users arrive
and leave your website. If you’ve got a high bounce rate, looking at the content on your
entry page might help you work out why. Similarly, if the content on your exit pages is a
logical exit point – a thank you page after a transaction, for example – there’s nothing to
worry about. But if visitors are consistently leaving in the middle of the shopping process,
there’s probably an issue you need to fix.
© Circus Street Ltd. 5
WEB ANALYTICS LESSON SCRIPT
LESSON 2
Next up: the Conversions section. This is where you’ll find information about how well
your site is driving transactions, leads, and any other interactions (macro and micro goals,
and events) that you’ve defined.
Most companies tend to focus on this section, because it has the most obvious impact
on the business bottom line. However the key role that Acquisition and Behaviour
play in generating conversions means you should really pay equal attention to those
sections too.
In this section, you might want to check out goal or event completions. This shows which
visitors complete your goals and events, which content and pages best motivate them
to do so, and in what order. Look for goals and events that are rarely fulfilled as well,
especially if they’re important to your business. Are there clear obstacles or other
reasons that might prevent visitors from completing their tasks?
The conversions section is also where you’ll go to see your funnel analysis. In order to
convert on your site, visitors will have to take a number of specific steps along the way.
For example, in the case of a purchase, the customer will need to place an item in the
shopping cart, enter their delivery and billing information, confirm the order, and pay. In
general, far more visitors will start the process than will finish it, and the funnel allows
you to see where they drop out. In theory, if you can locate and identify the issues that
cause them to leave, you should be able to improve conversion rates and therefore
increase revenue.
Bear in mind that you don’t always know the visitor’s motivation for doing something,
though. They might have added an item to their cart just to research shipping costs,
without an immediate intention to purchase. Visitors may enter your funnel at an
unexpected point, or skip a step somewhere. The funnel can be a great indicator
of patterns in visitor behaviour, but don’t interpret funnel analysis too literally!
The final part of the Conversions section we’re going to talk about is multi-channel
attribution. Understandably, marketers need to be able to evaluate which channels
and tactics are generating the highest conversion rates, and assign budget accordingly.
However, it’s pretty unusual that a visitor to your site will convert on their first visit. In fact,
it may take several sessions before a user converts, and it’s likely that each time they visit
your site they’ll come from different channels.
© Circus Street Ltd. 6
WEB ANALYTICS LESSON SCRIPT
LESSON 2
Multi-channel attribution reports offer the ability to display all of the channels that the
visitor used leading up to the transaction, and to assign different values to each one so
that you can evaluate the impact of each channel on the end result. Each business will
need to work out the attribution model that best suits them, and your analytics tool should
be customisable to represent the model of your choice.
Last but not least, we come to the eCommerce section. As you might have guessed, this
is only applicable to sites that actually sell products directly.
This section of your reports will help you answer questions on product performance.
What are your best-selling products? Are there any surprises here that you could take
advantage of with new or further promotions? If you have physical stores too, are there
any differences between the top sellers online and offline?
You can also use these reports to look at shopping behaviour. Do visitors look at a lot of
products, but add very few of them to their shopping carts? Perhaps you need to improve
your product descriptions, or add some reviews.
And finally, you might want to look at checkout behaviour. How many people abandon
their shopping cart, and when in the process is this most likely to happen? For example,
do visitors exit after seeing shipping charges or other additional fees?
As you can see, there’s quite a lot you could be measuring – and we’ve only just
scratched the surface! Rather than trying to look at it all at once, make sure you identify
the areas that correspond to your goals, and use this information as a guide so that you
know where to look when it’s time to fire up your tool of choice.
© Circus Street Ltd. 7
WEB ANALYTICS LESSON SCRIPT
LESSON 2
3. Segmentation & Reports
Now we know how to identify what data is useful to you, how to collect it and where to
view it, we need to think about ways to drill even deeper into that data, and come out with
some really meaningful insights. And where better to start than with segmentation.
Visitors come to your website for many different reasons, from many different places, and
with many different expectations. If you understand this, then it should be clear that you
can’t just look at all of your data as though it comes from one homogenous audience.
The segments function of your analytics tool allows you to choose specific criteria that
divide your visitors into smaller, more contextualised groups.
For example, returning customers are generally more likely to make a purchase on your
site than first time visitors. They’re already familiar with your products, and they need less
background information and reassurance about your returns policies. On the other hand,
they’re more likely to be interested in your loyalty club. For this reason, it’s helpful when
reviewing conversion rates and onsite behaviour to look at these two groups separately.
A good analytics tool will provide some commonly used “default” segments which are
already set up for you, as well as allowing you to create your own custom segments.
Google Analytics’s defaults can be used to segment your visitors according to the
Audience, Acquisition, Behaviour and Conversion categories - and they also offer
a ‘gallery’ of free to use segments that other users have found helpful.
Default segments are great to get you started, but with a clear understanding of your
business goals and KPIs, you’ll want to develop your own custom segments.
Let’s say your site has a high overall bounce rate. You’re also aware that some of your
visitors seem to be from a country where your products aren’t available. By creating
a custom segment to exclude these ‘unqualified’ buyers, you may well find that your
bounce rate improves immediately!
As always, don’t forget it’s not just about user activity on your site. If you’re using tags to
track your advertising campaigns, you can divide your audience into custom segments
here too. Say you want to look at whether a specific campaign worked better in certain
cities. You could do this by segmenting by city, and comparing the responses of visitors
who saw the campaign with those who didn’t.
© Circus Street Ltd. 8
WEB ANALYTICS LESSON SCRIPT
LESSON 2
So you’ve had a look at your data within your analytics tool, and implemented some
useful segments that help you identify patterns in what’s going on. The next step is to
create a report.
When it comes to producing analytics reports, it’s easy enough to just fill page after
page with numbers, graphs and charts that look complicated and impressive. The real
skill in web analytics is in understanding what those numbers tell you, and using the
evidence you’ve gathered to provide commentary on the current situation, insights and
recommendations for next steps. Remember that an executive summary of the data itself
is not analysis!
All of your reporting should focus on the objectives, goals and KPIs of your business, and
the potential effects of your findings on your overall bottom line and customer satisfaction.
You may well be called on to write long reports with lots of analysis, but if you’re
presenting to time-poor decision-makers you’ll need to answer the following questions
clearly and concisely:
What have you learned?
What do you want us to do, and why?
What benefit might we gain from taking action on this?
What might be the cost of not doing something about this?
Suggesting business-critical changes can sometimes be nerve-racking, because you’ll
never be able to determine the exact outcome before you start – but if you don’t act on
what your analytics data is telling you, there was really no point in spending all that time
collecting and analysing it. There’s no need to rely on ‘gut instinct’ though. At the bare
minimum, your analytics data should help you demonstrate the current losses or damage
being caused by the scenario that you’re flagging – but there are options for providing
evidence-backed propositions for change.
One option is to make estimates using external data. For example, if the industry average
cost per click for one of your keywords is less than you’re paying, you can estimate the
potential savings to be gained by improving your Quality Score. Case studies about
changes and optimisation tactics by other businesses can also be a great way
to generate support for your recommendations.
Once you’ve explained your recommendations, you’ll also need to prioritise them
according to levels of urgency and importance. Every organisation will have its own
© Circus Street Ltd. 9
WEB ANALYTICS LESSON SCRIPT
LESSON 2
process for assigning priorities and deciding the order of implementing recommended
changes. However, your report should provide some background and data to help inform
management decisions. You’ll need to answer questions like:
How mission-critical is this issue? If you’ve found an urgent problem for your business -
a broken checkout process, for example - then obviously a speedy response is needed.
If you’ve found some typos on a product page, although it may look unprofessional, this
might be something that can wait.
How much time, money and resource will this cost? Your management will need to know
what it will take to address the issue. As we mentioned before, even if you can’t calculate
the exact costs, you need to provide some framework for a budgetary discussion.
What all this comes down to is that a good report – whether it’s an executive summary or
a deep-dive analysis – needs to be all about identifying issues and proposing solutions,
using analytics data as evidence. A good analyst is great with numbers – but they also
need to be a great communicator for the whole business to benefit from their skills.
4. Dashboards
One way you might choose to present your analytics findings is using a dashboard.
Dashboards are normally one page (or screen) of information, showing you what’s
happening in critical areas, how well your most important KPIs are doing, and how your
results are trending over time – so you can find everything you need in one place.
It’s good practice to create a number of different dashboards, tailoring each one to
a specific audience within your business. This is helpful because each manager (or
department) can choose to see only the information that’s relevant to them. For example,
your Chief Financial Officer might want to see revenue and related spending figures,
while your Vice President of Sales might be more interested in monitoring improvements
to your conversion rates.
So, how do you create one? Your analytics tool might include a default dashboard
presentation – the homepage of your Google Analytics account is a good example.
This is a useful start, but to make sure that you get optimal insights for your business,
it’s best to design and customise your own set of dashboards. Your analytics tool should
allow for this, and you can also export data to one of the many commercial software
programs available.
© Circus Street Ltd. 10
WEB ANALYTICS LESSON SCRIPT
LESSON 2
The key to an effective dashboard is to identify the most critical information to present.
As with all methods of reporting, make sure you’re offering insights and recommendations
with supporting evidence. That means that you need to provide context for the data
you’re highlighting. Never show a single metric by itself without something to compare it
with, or further detail derived from drilling into secondary dimensions or segments.
One of the most important aspects of an effective dashboard are the visuals. Clear,
simple but compelling visuals are key to any good analytics report, especially on a
dashboard, where you’re trying to assemble all the crucial information in one easy to
read place.
Make sure you choose the type of chart that illustrates your data correctly, whether
that’s a funnel showing sequential stages, a pie chart or a bar graph. Whichever chart
you choose, it needs to clarify and simplify the information, not make it more complicated.
A good visual should be self-explanatory. If the visual itself requires a lot of commentary
in order to be understood (as opposed to commentary providing insights and
recommendations), then it’s not useful.
All good visuals have certain elements in common. Chart titles need to be simple and
compelling, and the labels for each axis or area of a graph need to clearly describe
what’s being shown. Using colour well is important too – too many colours can be
overwhelming and confusing, but using contrasting colours appropriately can really
bring clarity to an image.
If you can, try to give something of a narrative to your dashboard and the accompanying
commentary. Many people, especially those who aren’t data-driven, understand and
retain the message of a story much better than they do numbers and graphs. A well-
told story helps your audience see themselves playing a part in carrying out your
recommendations, and increases their engagement and motivation.
And finally, don’t forget that the components of your dashboards aren’t set in stone.
Critical metrics for your business will change over time, so don’t be afraid to let your
dashboard evolve by replacing areas that are no longer important to track with more
relevant ones.
Hopefully now you understand how to turn all those data points into something that’s
easy to read, digest, and action. The most important point to remember is that reports
offer evidence-backed insights, not just data out of context – and if you can provide that,
you’ll start seeing improvements to your activity in no time.
© Circus Street Ltd. 11