On February 23rd, 2012, Pure Visibility’s own Michael Beasley will be speaking on Michigan State University’s campus, at 6:00 p.m. The presentation will cover how to cut through the complexity of all the possible things you can measure with web analytics, to find the things that are really important and useful to measure, from a business and a user experience perspective.

This talk, which is the first Michigan Usability Professionals’ Association meeting of the year, brings together the areas of user experience and web metrics in much the same way Pure Visibility brings together these two fields within the work of our own Measurement Team.

For more information or to register, please visit the Michigan UPA website.

At Pure Visibility, one of our core competencies is Pay Per Click advertising (PPC). While some may think that navigating the interfaces at Google is where paid search management begins and ends, we know it is only the tip of the iceberg. One of our core values, “Driven to Improve,” plays out constantly in our interactions with our paid search clients.

In the paid search arena, landing page testing is one area where Pure Visibility gains extra leads and conversions for clients. It’s exciting, and we can see the results of quality landing page design, immediately.

I have been managing a very exciting project at PV since I joined the company. One of the sub-projects in that engagement, which I have really enjoyed, is a true team effort. Some of the stakeholders and team members include client representatives and their third-party website designers.

It has been an exciting project from a team building point-of-view, as the client manages both agencies. We all work closely to make sure the results meet our expectations and the client wins.

Good landing page design takes into account what the potential customer is looking for, website usability, and a compelling sell or call-to-action. Before the PPC part of the business can do its magic, the landing page has to be designed.

We design the blueprints (or wireframes) for the web design firm to create. After the page is designed, copy is written, and everything is perfect and approved, the new landing page is moved to a production environment where Pure Visibility’s PPC Pirates (a.k.a. Analysts) can use and test it.

Google has tools that make A/B testing very simple to understand. Using Google Website Optimizer, Pure Visibility’s PPC analysts can set up a test which sends users to the existing, control version of a landing page (A), 50% of the time, and then to a test version (B), the other 50% of the time. If there is a statistically significant improvement in conversion rate between A and B, the interface will let us know and the test will be over.

At that point, we can roll out that winning landing page to a production environment and use it 100% of the time. Since the client has several brands, this landing page design can be scaled and we see gains across all brands. Excellent!

In our last landing page test, the new version of the page was so compelling that it showed a 37% improvement over the older page with a 99.9% level of confidence.

The next time you search the web for something and you soon find yourself on a company’s landing page, stop and think, “How can they do better?” I always do.

I bet you thought that once you graduated school that you were done with having to worry about people evaluating and grading your writing. Well, think again! Google evaluates and classifies the reading level for each of your pages into one of the following three buckets:

• Basic
• Intermediate
• Advanced

The classifications are intended to describe the level of reading comprehension needed to understand the content on the page. For example, scientific research papers would be more likely to end up classified as Advanced, while you would expect a website targeted at kids to be classified as Basic.

Google does this for many reasons, but the main purpose, like most of the things Google does, is to better serve its users. By being able to evaluate reading level, Google can better identify poor quality content and filter out spammy low quality content from their search results.

In addition, they can allow user to search for things based on reading level. For example, lets say you just bought a new iPad and you were looking for a tutorial. You could search for “ipad tutorial” and filter sites that contain basic reading levels to ensure that the tutorials you review are easy to read and understand. Not sure how to do this? Don’t worry, just keep reading.

Filter Your Search Results By Reading Level

Here is a step-by-step guide to filtering your search results by reading level – it is actually quite easy!

Step 1

Enter your search query in the Google search box as usual. Then click the “Show search tools” link in the bottom of the lefthand sidebar. In this example, we will search for iPad tutorials.

Step 2

After clicking “Show search tools” the menu expands. Under the “All Results” menu, click “reading level.”

Step 3

After clicking the “reading level” the search results should refresh and you see a chart that lists the percentage of sites that fall into each reading level. In addition, the reading level is displayed for each individual search result.

Step 4

The last and final step is to click the reading level that you want to filter the results so that only sites that match that reading level are shown. That’s it, you did it!

In the example below I clicked on the Basic reading level. To revert back to normal search settings simply click on the “x” in the upper right hand corner of the reading level box.

Reading Level and SEO

Although the reading level can be very helpful for users, it can also be a valuable resource for optimizing content and improving your SEO. Generally speaking, you want to write content that is consistent with the reading level used on the majority of sites that discuss the particular subject matter in question.

For example, 71% of the sites that write about iPad tutorials have content that is at the Intermediate reading level, while only 8% have content that is at an Advanced reading level. This means that in the eyes of Google, an Intermediate reading level is more “normal” and appropriate for this subject matter.

And as a result, Google may favor this type of content in the results because it believes that it better suits the audience and provides a better user experience for the user – or so the theory goes?

If you think about this logically, it actually makes sense. iPad tutorials shouldn’t be at an advanced reading level. You don’t want tutorials to be like research papers; using long sentence structures, a large vocabulary, and requiring advanced reading levels to comprehend the subject matter.

You want tutorials to be written so that your readers can easily understand the material; short sentence structures and simple vocabulary (basic reading level). However, the iPad itself is a technical device and by nature some of the jargon and vocabulary used to understand it is probably above the basic reading level, which is why the majority of sites are classified as Intermediate.

By doing this research ahead of time you’d know the reading level you want to achieve (Intermediate in this case) and the reading level you most likely wnat to avoid (Advanced in this example). In addition, after you’ve written and published the article you can see how Google classifies it and then make the necessary modifications to get the desired result.

Reviewing reading level for keyword targets is not a game changer, but it can help you write better content and it can be an effective and useful tool in an SEO analyst’s bag of tricks.

We’re developing Instructor Resources for our Internet Marketing Start to Finish book. It has been a fun project.

Yet, because several of our chapters describe data analysis, any exercises for students require… well… data. And since we’re describing how to use web analytics data, creating exercises for using Google Analytics seems like the right first step.

All perfectly logical, except it lands us in a quandary. How do we create exercises for these chapters without knowing whether the students have access to web analytics at all? What might the quality of the analytics installation be, on any access they can beg for through a family or academic network?

We came to the conclusion we had to make available a sandbox account where students can view real data. Something where they can play with data, experiment with the interface and learn a few things.

Requirements for a learning Google Analytics (sandbox) account

1. The account should be properly installed (code on every page of the website). One of the first signs of poor installation is a large proportion of self-referrals in the traffic sources, showing that folks are looking like they’re coming from the same website. This can indicate incomplete installation of script-based web analytics services. Web data with configuration issues or gaps in deployment is tricky to impossible to analyze.
2. The account should have over two years of data in it, to shed light on any seasonal trends at work at the monthly level and to provide a long enough timeline to measure year over year change. Because web analytics is recorded from installation forward, it’s not retroactive. So, by the time the course has started, it is years too late to have an interesting data set for students to engage.
3. The account should have goals that have some business value by which to measure success of initiatives and/or success of the site itself.

So, Any Volunteers?

Yet, for businesses that use the Internet heavily for marketing and sales, web analytics data are typically proprietary, so it is hard to envision a scenario where interesting business data is freely available. If you know of such a situation, please leave it in the comments! Otherwise, keep reading for our current, half-baked solution.

Our Current Idea

The Good

The personal blog of one of the authors meets criteria 1 and 2. It has been installed completely on the (very simple) site and has been collecting data since fall 2006.

The Bad

The site and the analytic account is missing several things that would make it a good working dataset – the author is not running paid search to the site or doing a concerted link building program, so the referrals and paid search traffic sources are weak to empty. Worse, it lacks goals throughout its history, to be blunt the site lacks goals beyond experimentation and self-expression. So, it is a poor example for business goals, but a dataset nonetheless.

The Ugly

The thankless and annoying part will be to maintain access for students to the data. Setting up a publicly available login/password seems an invitation to treachery (among other issues, this “open” Google Account would end up having an available email account from which to spam the world).  

So, we’re going to opt for fielding requests from qualified instructors to provide a list of students access for a pre-defined period, opening us up to a lot of silly administration and a fair bit of “tech support” back and forth when student A cannot access her account because she does not yet have a Google Account set up even though she already gave us her email address.

This “access rigmarole” happens all the time with new clients, and doing it for free is not appealing. Yet, this is something that is part of the Google Analytics “setup” process, and is in itself a learning experience, though something I, for one, wish I’d never learned so well!

Let us know if you see a better solution out there or can envision something easier to maintain. Thanks in advance for the input!