miercuri, 25 martie 2015
Advanced Content Analysis in Google Analytics - Moz Blog
| Advanced Content Analysis in Google Analytics Posted on: Wednesday 25 March 2015 — 01:15 Posted by Jeffalytics We analyze the performance of our content every day. Sometimes it's subconscious, like when we check the number of tweets we get from a new blog post. Other times, we make more conscious efforts, like reviewing performance metrics in Google Analytics. This feedback—both formal and anecdotal—informs what we do next. It influences future blog posts and validates our strategies. Reviewing content performance on a regular basis has been key to the growth of many online publishers. We should all be taking note of these successes as we build our own content marketing efforts. Paying attention to which of your content efforts are working well is the cornerstone to data-driven marketing. The companies that make these investments can produce tremendous results. For an in-depth analysis on the importance of being data driven, here are two recent articles that inspired me:
These articles show how taking data-driven approach to producing content can produce great results. Exponential traffic and revenue in these cases. I don't know about you, but exponential traffic sounds pretty great to me! But we will never get there without taking a methodical and data-driven approach to our efforts. We will never get there if we are only counting page views. It's time to take things to the next level! Using Google Analytics Content Groupings and Dimensions to inform our content strategyFor many of us, Google Analytics is the tool of choice for analyzing website performance. It's free, easy to use, and extremely powerful. But because of the free and easy nature, most users do not explore the more advanced features of the product. One of the more advanced features that you have at your disposal is content grouping. Content grouping allows you to gather your content into common themes to create a more meaningful analysis of your data. For example, you can group your blog posts by the type of content that they represent. This grouping is helpful if you cover many topics on your website or sell many products. This is something that I have been doing for years on my own site. It helps me understand which topics resonate the most with readers. It also helps understand which topics drive organic search visitors. In the past, I would have to do this in a manual fashion. It involved exporting data into Excel and grouping content by the presence of certain words in the page URL. This was an ugly manual process that I would not wish on anyone. With content grouping in Google Analytics, we can get a view of this data with little effort involved. Here is a screenshot of traffic performance by content groups, based on common topics that I cover on my blog. Configuring content groupings in Google Analytics There are three ways to set up this feature. The easiest way to do it is by creating rules to define your groups. Rules work like advanced segments in Google Analytics. Set the criteria for your groupings and Google Analytics will do the rest of the work. Note that these rules work only on the page URL, page title or screen name (for apps). Here is an example of how to configure groupings matching words found in your page URLs. The definitions work as a waterfall. If a page url/title fits in your first definition, we exclude it from each future definition. For this reason, we want to be specific with our first rules and then leave the more general and "catch all" rules for the end. Notice how I used a regular expression to define what makes up a PPC Page. The pipe (|) symbol serving as an "or" statement in the expression. You can also use the "or" statement on the right, but this can get unwieldy fast. For long regular expressions, use the extraction method for content grouping. This works wonders for complex regular expressions with several criteria to classify posts. Using code to define your content groupingsThe above options use the data that you already send to Google Analytics with each page view (page URL and page title). While this works well if we have search friendly URLs and titles, it is also limiting in our ability to perform analysis. If you would like to analyze beyond words in your content, then you will need to use code to push this data into Google Analytics. While this sounds daunting, it is not too bad. I was able to get this code working in less than 30 minutes to provide a proof of concept. What are some groupings that you might want to use for measuring content performance? How about the length of your content? Many of us have seen studies on the importance of the length of our content. Is it worthwhile to write longer articles, or is that just a "best practice" that does not apply to your site? Let's measure it! How about the date that you published your content? If you put the date of your post in the URL, you can use rules to build these content groupings. I don't include the publishing year in my URL, so I would need code to get this done. Here is how I configured Google Analytics to track word counts and publishing year of articles. First, you set a new definition for your content grouping in the admin section. I selected indexes 4 & 5 to avoid any potential conflicts. I use Google Tag Manager on my website, so I pushed data into the system using the data layer functionality. My code looked like this for tracking word count, word count range and year published: We trigger this code on every page of my website using native functions from WordPress. If you are using Google Tag Manager and WordPress, I would be more than happy to provide you with the code that I used to build this data layer. Next, I created a macro in Tag Manager to recognize these variables. Once your definitions are in place, you will see your groups listed in the admin section of Google Analytics. You can define up to 5 unique content groups per view. Viewing this data in Google AnalyticsOnce your definitions are in place, Google Analytics will start to push this data into your account. Note that these definitions do not work retroactively—only on data moving forward. Unfortunately that means that you will need to wait a few days for meaningful analysis of this data. But when the data starts to come in, it's beautiful! Let's start with the content grouping definitions for post topic type. I have had these in place for a while, so this data is already providing meaningful insights. Here is what we start to see when looking at website visits by topic type. This high traffic volume, low page value metric helps me draw two conclusions:
By grouping my content into themes, I now have a fresh perspective on the effectiveness of my content. Instead of choosing the topic on my mind on any given day, I may benefit by only writing about Google Analytics. This level of insight is not possible without content grouping. Content grouping is incredible when you have this data tied into the goals you have already set up with Google Analytics. Checking in on our code-driven content groupingsAs you can see, content grouping provides excellent insights into your content strategy performance. If you have thousands of articles on your website, content groupings will help you sift through the noise and go right to the signal. You can gain insight into other aspects of your content strategy through this same method. Let's check in on the groupings that we set up through code earlier in this article. Please note that this is a proof of concept with only a small amount of data to support the groupings. Over time, your picture will start to become more valuable as you see conversions rolling into your account. How many page views are we getting for the content we produced over the past 4 years? This is easy to view with our content groupings. What about by word count? Not surprising, lower word count pages (like the homepage) are getting the most traffic. Applying segmentation to our content groupingsWe have grouped our content by length of the article and when it was published. Now we can measure how these factors impact our organic search traffic. We can do this a few ways. My preferred method is to look at the medium of organic search and then use a secondary dimension of content group. If granular keyword data were still available in Google Analytics, we would be able to segment brand/non brand traffic. But alas. We can do this same analysis by year as well. Content grouping makes everything better! Where else does content grouping make Google Analytics data shine?Many of your favorite Google Analytics reports get better with content groupings. The behavior flow report comes to life with your content groupings. We no longer need to look at this report with several branches of data hidden from view. Now you can see how people visit your site based on your pre-defined content groupings. Custom Reports You can also use custom reports to combine several fields together. For example, try to view organic visits by the year you wrote the content and the topics into a single report. Several years ago I wrote a post about investing in SEO for YouMoz. The basic premise is that SEO investment does not fit into normal budget constraints. For example, you may budget for all your SEO efforts in 2015, but there is a revenue impact of these efforts for years to come. A custom report by post year can help you better understand the continued return on your SEO investment over the years. What other content groupings make sense to explore?Once we start grouping our content for analysis, many possibilities become available. Here are a few more ideas for what we can measure for content groupings:
You can measure pretty much anything with content grouping. The only real limitation being your imagination AND Google's current limit of 5 content groups in each view. You can even get around that by using multiple views if you want. What type of questions can we answer with content groupings?With content groupings in place, we can answer more business questions than standard content reports. Here are a few business questions I can start to answer with the content groupings we have already discussed.
Through content grouping, we can find answers within our pre-defined points of analysis. We no longer have to look at individual posts and pages to find answers. We provide the taxonomy that works for our business. Then we use this taxonomy to show how visitors reached our website through acquisition reports. We see how they performed on the site through conversion reports. Now Google Analytics starts to think a lot more like our business. It uses our own words to describe content within a structure we define. Plus, we have the tremendous processing power of Google Analytics to handle our queries. Bonus: Use custom dimensions to make these reports even more usefulIf you were paying close attention to the data layer variables I showed earlier in the post, you will see a third variable. This third variable is the exact word count for each page. This variable was added to the data layer as I was starting to do analysis on the content groupings. I found that some analysis may become easier if I have the exact word count available in Google Analytics. In Google Tag Manager, I set a custom dimension of Word Count using my third data layer variable. Now, I can view post topic by word count of the article in Google Analytics. You can add up to 20 custom dimensions per web property in Google Analytics. It only works with the Universal Analytics version. What type of content analysis are you going to do now?Groupings are like a cheat-code for content marketers to take their analysis to the next level. You get to push your own data into Google Analytics. You get to use your own definitions within the tool. There are really no limits to what you can measure. What is it going to be? I would love to hear your ideas in the comments section. Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read! |
This simple screenshot is quite revealing. It shows which topics resonate the most, as well as content deficiencies. And these reports get even more valuable once you start to segment your data. More on this shortly.
Before we can get into deep analysis of our content, it makes sense for us to talk about how we can configure this report in Google Analytics. 
As soon as you have defined your grouping, Google will give you code snippets to use for tracking in Google Analytics. There is code for Classic and Universal Analytics. 
I gave a default value of 0-200—in the event that a word count is unavailable from WordPress, it will list 0-200 words. Then in my Universal Analytics tag, I set content groups in the tag configuration options. My indexes correspond to the groups we set in the Google Analytics interface. The words in the {{}} brackets represent the macros we defined above. 
After publishing, every page load will send content grouping data into Google Analytics. Pretty awesome! 
For even more on the topic of setting up content groupings, here is an awesome article by Michael King on 
While WordPress pages drive the most traffic, they have relatively low value per page view. This does not count any affiliate revenue, but it is indicative of the traffic brought in by this topic. High traffic volume/low value. 
This is a traffic pattern that I had assumed in my mind (I wrote much more in 2013 than 2014). Now, I have the numbers to prove it. 
This data will get even more interesting over time.
Again, we see that our shorter articles are driving the most search traffic. This is for two reasons. 1) The default content range is 0-200, so this includes articles with no word count defined by WordPress. 2) It includes our home page, which often ranks for branded search results. 
Notice that the current year is receiving the most organic traffic. I can only assume that this is again due to branded traffic. 
You can also start to add your conversion data in place and understand the value of the content that you have produced over the years. 
Useful? Definitely! There are many times when you need an exact number available to conduct analysis. | You are subscribed to the newsletter of Moz Blog sent from 1100 Second Avenue, Seattle, WA 98101 United States To stop receiving those e-mails, you can unsubscribe now. | Newsletter powered by FeedPress |
| FeedPress is a service edited by Beta&Cie, www.betacie.com | |
Seth's Blog : What is customer service for?
What is customer service for?
Customer service is difficult, expensive and unpredictable. But it's a mistake to assume that any particular example is automatically either good or bad. A company might spend almost nothing on customer service but still succeed in reaching its goals.
Customer service succeeds when it accomplishes what the organization sets out to accomplish. Google doesn't have a phone number, doesn't want to engage with most users. McDonald's doesn't give you a linen napkin. Fedex used to answer the phone on one ring, now it takes 81 seconds for them to answer a call. None of these things are necessarily bad, they're merely examples of alignment (or non-alignment).
Organizations don't accidentally run ads, don't mistakenly double (or halve) the amount of cereal they put in the box. They shouldn't deliver customer service that doesn't match their goals either.
Here are some uses of customer service:
To create a significant competitive advantage by engaging with customers in a way that others can't or won't. This is what the over-the-top customer service approach of Zappos did. They went from being a commonplace (you can buy shoes from anyone, we're anyone) to a customer delight company that happens to sell shoes. Rackspace does the same thing with technical support.
To streamline the delivery of inexpensive goods produced in an industrial way: This is the model of most fast food places. Deal with the exceptions quickly and well, and keep the line moving. Part of this mindset is to not make it easy for people to complain, and to treat every complaint in just about the same way. When you get a bag of rancid nuts from Planters, sure, you can visit the website, click a bunch of times, fill in a form and let them know, but they don't use it as an opportunity to earn your loyalty. "Here, take four coupons, each good for a dollar off one purchase, thanks, we're done."
To lower expectations and satisfy customers by giving them exactly what you promised, which is not much: This is the model of automated customer service at most big web companies. They'll do just about anything to avoid an interaction with a human, and they're clear about this, meaning that they should only end up with customers who are okay with this.
To raise expectations and delight customers by giving them way more than they hoped for, which was a lot: This is a truly difficult promise to maintain (Apple did it with the Genius bar, but they rarely surprise there any more). The secret is to find a focus, a budget and a scale where you can actually deploy talented individuals to keep this promise.
To dance with customers in an act of co-creation: This is part of 37Signals' secret. From their book to their blog to their clearly stated point of view about platforms and the way they do business, they invite customers to debug with them in an ongoing dialogue about finding a platonic ideal of utility software. They don't promise perfect, they promise engagement. Over-inform. Speak with respect. Be clear about the invitation. This is a very special sort of customer service, and companies often think they're doing this but end up cutting corners and are merely plodding along, disappointing those that would have preferred to engage instead.
To diminish negative word of mouth: Many large organizations resort to this, the last step in a sad journey. As soon as a wheel gets squeaky, they grease it. But that's all they do unless pressed. The problem is that many of your unhappy customers are too busy to get squeaky, they merely go elsewhere, and the ones who you finally do try to help are so pissed off it's too late.
To build extraordinary trust: This is the initiative taken by an institution to do far more than is expected, at a human level, to earn the privilege of serving again. This is the banker who visits you in the hospital, merely because she heard from another customer that you were ill.
To treat different people differently: One way to reward your best customers is to treat the best of them substantially better than others--the word will spread, others will want to join this group, and those in it will be hesitant to switch to a competitor. But if you make that promise, you need to double down on it substantially, continually improving how you treat your favorites.
To race with competitors to lower customer service costs just a bit more than they will: This is the current progression we see among industrial titans who see customer service as a cost, not a profit center. When you measure this, you can't help to want to drive the cost down, and you will do it just a bit faster than your competitors, because to do it too fast is to risk condemnation. Alas, in just about every industry that the internet has sucked the profits out of, we see this cost-cutting race to the bottom. It's not going to end well.
Because you can: This is awfully rare among public companies, but there are many organizations that treat people as they'd like to be treated. Not to grow market share, but because it's the right thing to do.
So it's clear that good customers with urgent problems left on hold by Fedex is a mismatch between what they built customer service for and what they're doing with it. And that a busy startup that doesn't invest as much time as they could in co-creation communication is not serving the goals of the beta fully. On the other hand, the novelist who doesn't invest time in answering reader mail is probably doing good customer service, since reserving her best efforts to write another great novel is precisely the promise she has made.
Every single person who makes budget decisions, staffing decisions and customer service decisions must to be clear about which strategy you picked, needs to be able to state, "we're doing this because it's congruent with what we say customer service is for."
Obviously, you can mix and match among these options, and find new ones. What we must not do, though, is plan to do one thing but then try to save time or money and do something else, hoping for the results that come from the original plan without actually doing it.
Customer service, like everything an effective organization does, changes people. Announce the change you seek, then invest appropriately, in a system that is likely to actually produce the outcomes you just said you wanted.
Make promises and keep them.
More Recent Articles
- Of course it's difficult...
- Things we don't see that much any more
- The tyranny of random numbers
- Companies don't care about you
- How long is your long term?
[You're getting this note because you subscribed to Seth Godin's blog.]
Don't want to get this email anymore? Click the link below to unsubscribe.
Postări
