I simply love web analytics.
It’s one of those things that most people know nothing about, but make a huge difference.
I say that because I was one of those people.
Like anyone who has had a website before, I knew the basics of Google Analytics, and the basic definitions of things like pageviews, goals, time on site, and so on.
But when it was time to really analyze my website, I was simply useless.
I didn’t know how to do that. In fact, I didn’t even know that you had to analyze a website.
Fortunately, last year I made the decision to learn as much as I can about web analytics, even though I can’t remember the reason why I decided to do so (I think it was to get a job in Buenos Aires as a web analyst).
The thing is, I took the only serious and professional course there’s available online: Market Motive’s Web Analytics course.
It’s taught by Avinash Kaushik, one of the most famous and important people there are in web analytics.
The only caveat? It costs 300 dollars a month.
However, I must warn you that you don’t have to pay 300 dollars a month to learn what I learned.
Fortunately, you have a much cheaper option. I’m talking about Avinash’s highly successful book Web Analytics 2.0.
Basically, Market Motive’s course is Web Analytics 2.0 made in video.
When I realized that I was quite disappointed, because I had already read that book before taking the course, and I was expecting much more for the 300 montly cost.
However, I learned much more with the video course, because (and this is the key) I had to take notes. Lots of them.
Listening to Avinash talk made everything easier to grasp.
Today I’d like to share with you the 10 key things I learned throughout this course.
Disclaimer: This post doesn’t contain affiliate links of any sort. Nobody paid me to do this, neither I do it for any particular reason. So don’t be a jerk, relax, and enjoy the lessons. Word.
Market Motive’s Web Analytics Class Notes
Lesson #1: Web Analytics Framework
The first thing I learned was the web analytics framework.
This framework is also know as the Digital Marketing and Measurement Model (for those who studied for the GAIQ test, you’ll know what I’m talking about), or Web Analytics Measurement Model (WAMM).
In words of Avinash himself:
The root cause of failure in most digital marketing campaigns is not the lack of creativity in the banner ad or TV spot or the sexiness of the website. It is not even (often) the people involved. It is quite simply the lack of structured thinking about what the real purpose of the campaign is and a lack of an objective set of measures with which to identify success or failure.
I’ve developed the Digital Marketing & Measurement Model as a simple, structured, five step process to infuse this much needed thinking.
The web analytics framework consists on defining:
- A business goal (the reason why you exist)
- Specific goals (the different goals or objectives your business should aim to achieve)
- KPI’s (see later)
- A target for each KPI
- Segments for each target (see later)
A web analytics framework looks like this:
Pretty cool, isn’t it?
What Avinash taught me is you must develop a WAMM with any website you’re working with.
Not only it will give you clarity of thought, but it will also help you analyze the effectiveness of your campaigns and activities.
You’ll use framework throughout the whole marketing campaign, so be sure to create one properly.
Lesson #2: KPI’s Importance
The most important question is: Why does my site exists?
As you’ve seen in the framework above, you need a benchmark to analyze how well (or bad) you’re doing against your specific goals.
That’s where the KPI’s show up (holla!).
KPIs are an acronym for Key Performance Indicator, which indicate the key performance of your goals (see what I just did there? 😉 )
Jokes aside, a KPI are measures that help you understand how you are doing against your objectives.
They should be:
- Instantly Useful
For example, if you had an e-commerce store, and you wanted to increase your sales, your KPI’s could be any of the following:
- Average order value
- Conversion rate
- Days to purchase
- Visits to purchase
It’s important to note that you shouldn’t have too many KPI’s. Usually 2 or 3 KPI’s are enough.
Lesson #3: Segment or Die!
I think this is the most important thing I’ve learned in this course.
Learning about segmentation completely changed the way I think about web analytics, SEO, online marketing and marketing in general.
Segmentation is SO important and at the same time so underappreciated.
By segmenting data, you can find useful insights for your business. Your entire attention should be on: “What can I segment? What kind of commonality can I find on my visitors and in their behaviour in my site? What can they accomplish so that I can understand, and from which I can extract insights in which we can take action?”
How to Segment
1. Drilldown into standard reports: Drill down a little bit in the standard report, and find simple commonalities among the data.
2. Use standard (built-in) segments: Use the segments that come with the analytics program. Among the best segments in GA you have:
- Non-paid searches;
- Bounced visits;
- Mobile and Tablet visits;
- Multi-visit Users.
3. Create and use custom segments. There are 3 main custom segment categories.
Custom Segment Categories
1. Acquisition: All segments in this category represent all the activities that you undertake as a business to make people come to your website. These activities could be blogging, PPC, etc. Among the different segments, you have:
- Specific location (country, cities) of the people you are attracting;
- Search engines (SEO or PPC);
- Referral Websites (Social media, earned links, specific websites);
- Campaign traffic (Display, social, email, affiliate);
- Conference Presentations, Tradeshows;
2. Behavior: All segments in this category represent all the people who take actions once they arrive to your website. Among the different segments, you have:
- Really engaged traffic that have visited more than once, and stayed longer than X time;
- Bounced visits vs. non-bounced visits;
- People who did X action during their visits vs. those who didn’t;
- People who visit once vs. those who visited more than X times;
- People who use internal site search vs. those who don’t, and which keywords they use;
- Visits with abandoned carts and/or checkouts vs. those who don’t.
3. Outcomes: All segments in this category represent all the activities that add value to your website/business. This segments not only help segment macro conversions (like conversion rate), but also micro conversions much better. Among the different segments, you have:
- People that order over the average order value;
- People who make purchases (macro conversion);
- People who complete specific goals (micro conversions);
- People who buy more/less of an specific quantity, revenue, goal value;
- People who buy a specific product or complete a specific goal;
- People who buy multiple times.
Applying the Three Segment Categories to Standard Segments
1. Acquisition Segments:
- New Users
- Search Traffic
- Paid search Traffic
- Non-paid Search Traffic
- Direct Traffic
- Referral Traffic
2. Behavior Segments:
- Bounced Visits
- Non-bounced Visits
- Performed Site Searches
- Single Visit Users
- Returning Users
- Multi-visit Users
3. Outcome Segments:
- Make a Purchase
- Visits with Conversions
- Visits with Transactions
4. Acquisition or Behavior Segments:
- Mobile Traffic
- Tablet and Desktop Traffic
- Table Traffic
- Mobile and Tablet Traffic
Lesson #4: Macro and Micro Conversions
One very important lesson that I learned from this course was the difference between Macro and Micro conversions.
Basically, a macro conversion is any metric that represents an Outcome, or that represents an activity that is directly tied with an Outcome.
What’s an Outcome?, you may ask.
An Outcome it’s the most important activity in your site, the one that makes it or break it.
In an e-commerce site, it’s # of transactions and/or average order value (among others).
In a lead-gen site, it’s leads.
In a SaaS, it’s trial sign-ups and/or paid sign-ups.
The Outcomes are directly related to your revenue.
If your site doesn’t have any Outcomes, it’s not profitable. And, as far as I know, you can’t pay for rent with kindness. You need to make money!
Fortunately, there are ways to have your site “create” more Outcomes.
See, your Outcomes (I write it with capital o just to make a point of its importance) don’t happen in a void (even though that would be awesome 🙂 ).
Your Outcomes happen because there are other activities that support and “nurture” that Outcome.
Those activities are your micro conversions.
[…] even though the macro conversions are the one that bring you the money, all the smaller steps that lead to that conversion are as important, if not more, than the conversion itself. […] the micro conversions have a bigger impact on the conversions on the long-term; they enhance the lifetime value of the customers. They make the economic value of the website, while the macro conversions make the total revenue.
In an e-commerce site, a micro conversion would be the #of products added to cart.
In a SaaS, a micro conversion would be the # of email suscriptions.
Micro conversions can be detected by looking for all of the non-convertible activities that are tied to the existance of the website/business.
If you want to know more about defining and measuring your micro and macro conversions, just head to this Avinash post.
Lesson #5: Web Data Analysis Process
In the first lesson, I talked about the what’s the Web Analytics Measurement Model (WAMM), and its importance.
This is the meat of the course, were you’re going to pull the big analytics guns.
In order to apply the WAMM to a website, you need to follow a 10-step process:
- Visit the Website, Define Business Goals
- Diagnose the Sophistication of the Acquisition Strategy
- Analyze Visitor Orbit Strength
- Aggregate Marginal Gains
- Identify Valuable Content
- Question the Search Strategy
- Focus on Outcomes
- Optimize Marketing and Campaigns
- Focus on the Funnel
- Analyse Unkown Unkowns
If you want to learn how to apply the WAMM to your website, check out this article I wrote: How to Apply the Web Analytics Measurement Model.
Lesson #6: Qualitative Analysis
You see, not everything you do is about reading dashboards, and manipulating data with advanced segments, and stuff like that.
You need to remember there’s a person behind those clicks you analyze.
Crazy, I know.
One great way to understand the person behind the data you analyze, is by doing a qualitative analysis.
There are three main qualitative analysis solutions:
- Usability testings
- Heuristic evaluations
The first qualitative analysis solution is surveys. The simplest survey you can run on your website is called “the God’s gift to humanity“. It consists in 3 questions:
- Why are you here? This will get you the visitor’s primary purpose.
- Where you able to complete your task? With that information, you can see the level of success, depending on their primary purpose.
- Why couldn’t you complete your task? With this information, you will get a lot of responses, but very few, high-quality insights.
The key goal is to find segments of discontent. You’re trying to find all the people who’re disatisfied with your website, all of those that couldn’t complete their task. You’re trying to find ideas and clues to improve your website.
There are two types of surveys:
- Site-level, which give you broader answers.
- Page-level, which give you information about a specific page. Usually, the page you conduct page-level surveys are those that are very important to the business, like a support or product page.
Then, you can conduct usability testings.
Although they can be pretty expensive form most, UserTesting.com allows you to conduct online and quite cheap usability tests. This process consists of 4 steps:
- You set the target audience (Gender, Age, Country, Yearly Income and Expertise with the Web).
- You set the scenario (mindset) you want the users to have before they start the test, and different tasks you want them to do, and the questions you want them to answer after they do the test.
- You set how many people you want to do the test. At the end, you get the videos of the user’s experience, and the writter answers.
- Let UserTesting.com do the test for you. Review and analyse the test.
To get affordable expert analysis, you can use ConceptFeedback.com. You upload a page/logo/design/information architecture for a website on that website, and then you can choose one of three options:
- Get public feedback (for free);
- Get expert feedback on either copy, strategy, design and usability (for u$s99 for each area);
- Get in-depth user experience and strategy analysis (for u$s999).
In any case, you get expert analysis of your website.
Lesson #7: Experimentation and Testing
This is probably where the couse is the weakest.
Although Avinash doesn’t talk too much about how to develop a test plan, or other specific topics about testing, still he makes a good point that you need to test if you really want to understand your visitors.
The benefits of Testing and Experimentation are:
- You don’t guess anymore
- You give customers real voice
- You increase creativity and idea creation
If you want to learn more about Experimentation and Testing, I’d recommend you read Avinash’s book Web Analytics 2.0, where he explains this topic better.
Lesson #8: Competitive Intelligence
Do you remember when you were a kid and you watched spy movies, like…I don’t know, I never liked spy movies.
The thing is, you can spy on your competitor’s websites by using tools like:
Most of the CI tools are super expensive, so you probably won’t end up using them.
(Unless you were Bill Gates. If you’re reading this, Bill, Id’ say you use them.)
You see, the competitive intelligence is the final step of the web analysis process.
The only people who really need using CI in ther web analysis are the big websites, the one that compete against the “giants of the internet”, like Amazon, Ebay, and those type of huge sites.
Still, it’s nice to know you can spy on other people’s websites.
Lesson #9: Multi-Channel Analytics
In an ideal world, you could merge the online information with the offline information, based on the primary key of each person. So if someone with a particular primary key looked for something online, and then bought it offline, we could search and come up with the specific channel where that sale was coming from, and give the impact to that channel.
The problem is that there aren’t primary keys in the real world.
That’s why you need to track your offline activities online.
Although this may seem impossible to do, there are some ways to track and measure your offline acivities.
Tracking the Offline World
1. Vanity URL’s
When you see an advertisement with a URL that ends with “/tv” (or something similar), if you go to that specific URL, it will redirect you to the correct URL, which has a tracking code of the channel where the person came from (in this case, TV).
To know the impact of a specific offline campaign, you just have to create an advanced segment that includes a Campaign, and contains the name of the campaign itself. Then test it, and then create it. Finally, use it the Goals Report, or other ones.
2. Unique Coupons/Offers
This can be used on magazine, billboards, catalog and newspaper campaigns. You put a coupon or a special offer, and then when people use it online, you can track them. You just have to create an advanced segment, just like with vanity URL’s.
3. Online Surveys
Put a question in surveys to ask how they arrived today to the website. Put offline options, and that’s it.
4. Correlate Traffic Patterns
Correlate your offline and online data to get the most information possible. To do that, see when and at what times you run offline campaigns, and see the impact in the online world.
5 Techniques to Measure Multi-Channel Analytics
1. Measure Offline Calls to Action
Assign a goal value to the “Store Finder” URL. It’s important to put a relevant goal value. Also, in the profile configuration, inside the “Query Parameter”, put the specific parameter of the URL that tracks the ZIP code (or postal code, or whatever you want to call it).
Then, in the Site Search Report, see which ZIP codes are the one that look for the most amount of stores.
That way, you will know how much value the store finder created to the company, and which specific stores have the most traffic.
2. Track Phone Calls and Live Chat
The first thing to track, are the calls to the call center. To do that, you can use two tools: Mongoose Metrics or FreshEgg.
With either of these tools, you put a JS tag in the places where your phone number appears, and then it will give different numbers depending on where the user is coming from. So if someone comes from Google, they have one number, but if they come directly, they will have another one.
With these tools, you can get different reports, where you can see the conversion rates, and other useful metrics that help you measure the holistic impact.
The other thing to track, is the live chat. There’s one service that can help you track this, and it’s called LivePerson.
The only thing you have to do is put two extra lines of code in your GA tag. Then, you have to allow LivePerson to track Chat and Voice Reporting in GA. Finally, in GA you have to create a goal for LivePerson chat. After that, you will see in your reports the LiveChat performance.
3. Use Unique Coupons and Offers
This is similar to what was said before, in the “Tracking the Offline World”. Put a special offer or coupon online, and when someone redeems it offline, you will know that.
4. Use Online Surveys
Put a special survey (maybe an OnExit survey), that ask people wheter if their online experience has had an impact in them or not, so they are more likely to go to the offline store and maybe buy something there. The metric is called “propensity to visit a store site impact”. Also, you can call it “Likelyhood of Offline Purchase”.
5. Conduct Controlled Experiments
You run an experiment where you change some things about your website, and then you show it to a particular set of people (maybe cohort-based, maybe geographically-based), but not to the entire set of customers, for a short period of time (a week, maybe a month). That way, you can see the effect that those changes have had in a particular set of customers, without having affected all the customers. Of course, if the results are positive, then you might want to apply the same changes to all the customers.
Lesson #10: Dashboards
Dashboards are useful, because analysts don’t have any decision power, neither any power to drive action. That’s why when an analyst creates a great dashboard, it has to drive great action.
There are two types of dashboards: tactical and strategic.
Its goal it’s to show aggregated and/or segmentated data to one specific department/silo. It’s made for mid-level leaders and individual teams. An example would be a dashboard to the PPC team, or to the affiliate team inside a company.
This type of dashboard isn’t very difficult to create, neither useful.
Its goal it’s to make business performance analysis. It’s made for senior company leaders so they take action. All of your efforts should be focused on creating this type of dashboards.
Other Things I’ve Learned
As I said at the beginning, I already knew what Web Analytics was all about. Still, I didn’t get it.
It hadn’t clicked yet.
One thing that Avinash does amazingly well, is transmiting his energy and passion for web analytics.
I can’t stress that enough.
He makes you want to love and care about web analytics in a way that’s not easy to get from a book or blog post.
Thanks to Avinash, now I love web analytics, although I don’t have the experience I wish I had.
Also, Avinash is funny. If he wasn’t that funny, probably this topic would be boring to learn, but he makes it so funny, and so entretaining, that you won’t be bored at all.
You’ll love learning web analytics with Avinash, I promise.
This has been my experience with Market Motive’s Web Analytics course.
I shared with you almost everything I learned, including my notes from the course.
If you really wanted to become a web analyst, I’d recommend you to take the self-paced version of the course.
But if you want to become a professional web analyst super ninja, then I’d recommend you to do the coached version of the course.
Sure, it costs $3500, but I’m sure it’s worth the price, if you consider that you’ll be coached by the most famous, respected and funniest web analyst there is.