Google Intelligence

Google Intelligence is one of the options on your Google Analytics Dashboard. As you can see in the screenshot above, the button for it is just below the "Dashboard" button.

If you click on it, you'll see a line graph for the length of time you're examining (default is 30 days), and below that a bar graph showing all the times Google got a surprise. For many sites, you'll find that you have no alerts, because the site has performed just as Google thought it would. Google has lots of data to work with, so it gets pretty good at predicting.

Our example has had as many as 11 alerts in a day over the past month. On the day we're looking at, Google got surprised by the total traffic, the number of returning visitors, the number of visitors from the United States, and the amount of referral traffic the site received.

You can see that Google was expecting 0-52 visits from referral traffic and the site received 74. That's 46% more than Google was anticipating, based on past performance for this site and Google's experience of sites in general.

This site is surprising Google mostly just because it's getting traffic increases faster than is typical. We don't really need to take any action when we see these figures, and we could even set our alert to a lower level of sensitivity.

Sometimes, though, you want to do something in response to these alerts. For example, a client recently had a 324% higher than expected number of visits from Illinois -- an extra thousand in one day. For something like that, we need to find out why and do it again. Another client saw a higher than expected bounce rate among new visitors on several occasions; we need to examine that site to see what changes we ought to make to help new visitors feel more at home.

You can also set custom alerts. If you're doing a seminar in Illinois and hoping to recreate that spike in traffic from the Prairie State, then you can ask Google Intelligence to watch particularly for increases in traffic from Illinois. You can also watch your bounce rate for people entering from Facebook or the amount of time new visitors spend on a page.

Just click on "Create custom alert" on the Intelligence page and use the form to set the parameters.

Google Intelligence can be a time saver, alerting you to data that could be important to your strategy. It doesn't take the place of human analysis, but it can be very helpful.

Reading Your Google Analytics Reports

When I write about Google Analytics, I generally write as though you were going to your own analytics account and reading the data there. However, many business owners receive a report. Your webmaster or SEO professional chooses the elements which are most useful for you and instructs Google Analytics to prepare a report, which can be exported as a PDF or for a spreadsheet. The shot above shows what the PDF report looks like.

Lets examine it in more detail.

The top half of the page tells us about the visitors. The blue line shows the past 30 days and the green line shows the 30 days before. This is the default option; you might get reports showing different lengths of time. A glance at the line graphs for traffic show that traffic is increasing steadily, which is what we like to see. We can look more closely for more specific information, and other patterns might prompt us to check for other details, but in this case we just see a fairly new site getting increasing amounts of traffic.

Here's the section between the two line graphs. Reading down the left-hand column, we see that visits have increased by 234.63% and pageviews by 245.11%. This means that the total number of visits to the site increased significantly. Pageviews counts the number of times individual pages were seen, and that also has increased. Below that we can see that the average number of pages visitors look at has increased slightly.

The right-hand column starts with the bounce rate. When visitors look at only one page and leave, they "bounced." Almost 71% of our visitors look at just one page and leave. We usually like to see a bounce rate of 40% or lower. This site is a blog, so it's normal for people to come look at one page -- the page, typically, that showed up in their search results -- and to leave. Still, we'd like to see people stay and explore. Our bounce rate decreased by about 2% since last month, so it's going in the right direction, but that's something we should work on.

Below the bounce rate is the information for the average length of time people spent on this website. It was just over one minute, an 18.97% increase over last month, but we'd like to see that number increase. The last metric in this section is the percentage of new visits: over 90%. This is a new site, and the traffic is increasing rapidly, so we're not alarmed by this, but we do want to see more returning visitors in the future.

The bottom half of that first page shows us the number of unique visitors. We see that this site received 2,744 visits, but there were 2,513 visitors. This shows that some people came back more than once. The bottom half of the page also shows that most of our visitors came from the United States. It also shows an overview of our traffic sources: the overwhelming majority of our visitors came through search, with referring sites the next more popular method, some direct traffic, and "other." Google says that "other" simply means that the database is full and Google wasn't able to report all the data. It's widely believed that "other" traffic is from email or ad campaigns; this site isn't using email or ad campaigns, but people may be sending one another links to it.

We can get more detailed information in reports, too. If your webmaster or SEO professional feels other reports may be helpful, or if you ask for them, you can get reports covering any of the data in analytics. In this case, we might want to look more at our traffic sources.

This is a report on traffic sources:

The report shows the places traffic came from during the past month, compared with the previous month. In the detail below we can see that direct traffic and referral traffic represent a lower percentage of the traffic this month than they did last month. That doesn't mean the number of visits has decreased. It simply means that visits via the search engines have increased more than the other types of traffic.

Overall, this means that our SEO efforts are going well. We like to see increasing results from search. However, I usually aim for a more even distribution. With fewer than one quarter of our visits coming from referring sites, it's clear that we could use a good linkbuilding campaign.

Another section of the report shows the keywords people used to visit: the search terms they typed in at Google which brought the site up for them.

We can see that "classroom theme" is a good keyword for this site, and we should keep using it.

Obviously, I wouldn't show you analytics for someone else's site, so I am of course the owner of this site. I don't really send myself reports. If I did, though, I could read this report and conclude that I needed to make some changes to encourage my visitors to spend more time exploring the site and to come back again. I would also want to do more linkbuilding in order to get an increase in referring traffic.

If your reports don't make sense to you, ask the person who sends them to you for help in understanding them, or ask here in the comments. If they don't contain all the information you want, ask to have other metrics included.

Traffic at Your New Site

How can you expect traffic to progress at your new website?

Assuming that you're doing a good job with your SEO, you should see a pattern like the one above: some dips and peaks, but overall a steady upward climb. What do the dips and peaks mean?

The example above is a site directed toward teachers, and it shows a slump at the end of school and the Memorial Day weekend. Following that spell when teachers don't want to think about school, it gets back onto its upward path.

This example, a recent redesign, has a much spikier pattern, both because the total number of visits is smaller, and because there's more variation in the number of visits each day. But we do see a gradual upward trend. As time passes and the traffic continues to increase, we'll see a clearer pattern.

What if you're not seeing that pattern?

This example shows a site that recently launched a redesign. We see on the left the typical few visits a day the site has had for years. After the launch, we see some peaks -- and valleys. The average number of visits per day has more than doubled since the launch, but it's not yet showing the consistent upward trend we want. In this case, too, there are real-world factors: this is a business affected by weather, and the troughs coincide with rain.

We'll keep an eye on this. If, at the end of the month, we don't see an upward trend, we'll know that we need to give this some more juice to keep the improvement in traffic coming.

The thing to look for is that upward trend. You may want it to speed up, and you may want to fine-tune the type of traffic you're getting, but the rising line is the sign that all is well with your site.

SEO Makeover Success

We recently did an SEO makeover for Bill West Roofing, a roofing contractor in Kansas City. Initially, the company contacted us thinking about PPC or magic behind-the-scenes optimization. They liked their website (see below), which they'd built with the help of a friend, and didn't really want to change it. They just wanted it to perform better.

Here's the thing: online marketing with a well-optimized site always works better than attempts to shore up a poor site.

So we gave the client an estimate for making minor changes, but we also sent him a proposal to optimize the site properly. Sensibly, he decided to do it right. Here's what we did:

• First thing, we added Google Analytics to his site so we could see how things were going. We checked his rankings with the search engines, researched his competitors and the patterns of search in his town, and got to know his business. You have to  know where you're going before you set out on the journey.
• Using the results of the research, I re-wrote his content. I kept the feel of the original content, but wrote it with search and conversions in mind. I made it scannable and made sure the essential selling points were foremost. Content always includes meta tags, too, so I made sure that those were as they should be.
• Designer Tom Hapgood updated the design and code. We used the same research, plus some testing for usability issues, to inform the changes. The client had lots of input, and Tom tweaked the design to suit. The new design is similar to the old one, with the same colors and a skyline graphic, but it's more eye-catching, more readable, and more attractive. Search engines don't care about attractiveness, but human visitors sure do. 
• Josepha did a linkbuilding campaign. The full results of that won't show up for a while, but it's essential to get the site off on the right foot.

We gave it a couple of weeks to settle in, and compared the results since launch with the same amount of time before launch.

Total visits are up by 87.3%, PageViews are up by 117.53%, the bounce rate is down, time on the site is up, and we're seeing 90% new visits. Looking just at the company's service area, Missouri and Kansas, visits have more than doubled. (Wisconsin visits are up by 700%, but that doesn't do Bill West Roofing much good.) Search traffic is up by 50%, with visitors arriving via 51 different keywords since the May 18th launch.

This isn't just bragging. Of course, I'm excited by these results and happy to share them with you. But this is what I'd like to suggest you take away from this story:

• Good, honest SEO gives good results. You don't need to look for magic.

• Changes that improve user experience are also likely to improve traffic and rankings.

New Website Lab Report: Benchmarks

The new website I just launched -- this is an interior page, with Earth Day lesson plans -- has two days' worth of analytics now. The first day it had five visits and the second day it had seven, for a total of 12 visits from 6 unique visitors. 8 were direct visits, 3 were referrals from this blog, and 1 came via search, looking for "freshplans."

Yesterday the site was at #18 on Google for "Fresh Plans" and over 100 for "FreshPlans." Today, it's at #14 and #28 respectively for those terms. 

We've done nothing at all, except to filter my computer out of the results and mention the site here. So this is what you might expect if you simply build a good site and put it up and don't do anything in the way of search engine marketing. I'm going to give it a week before I take any steps, to establish proper baselines.

Analytics and Your Web Site Redesign

Industry expert Dave Taylor says that he's never arrived at a meeting about a site's SEO redesign to find clients with analytics data. Me neither, unless I bring it with me.

What kind of evidence do people bring with them?

Well, here are some comments from initial SEO site redesign discussions I've been involved in recently:

• "We should have more pictures of people. People like looking at people."
• "We need to do what [competitor] did."
• "We need something edgy."
• "We need it to pop."
• "Perhaps we need Flash?"

These may all be true statements about what these websites need. But, as Erika Andersen has pointed out, starting your meeting with statements about needs means you're jumping ahead to strategies before defining the problem or the desired outcome. It's the equivalent of saying, "Hey, let's try this!"

Starting with analytics allows you to see what's happening and compare that with what you want to have happen. For one of the sites in question, the owner defined the problem and the desired outcome by referencing the bounce rate (the percentage of people who leave the site without exploring further) from one source. He wants to keep those people hanging around the site longer. He's arranging for four pages with slight variations so he can test them.

For another, I've just installed analytics, so I'm not yet seeing a lot of information, but I can see that this 18 month old site appears to be averaging only one visit each day, and none from his service area. We can easily identify his immediate problem. Once we start bringing him some traffic, we can look at other issues, but making his site findable is clearly the highest priority.

So on the one hand we have

"In order to increase the ROI on our ad campaign with Site X, we need to improve the bounce rate for people coming from our ad by a) testing the effect of a suggested design change, and b) following through with the design that gets the best results" 

and on the other hand we have

"We need it to pop!" 

I hope the difference is obvious.

Filter Your Google Analytics

Sometimes I analyze websites' analytics for their owners or for other companies. Sometimes I analyze current websites, including their analytics, so I can improve their performance or make sure that their new site does what they need it to. Sometimes I follow clients' analytics as part of my work with them.

In all of these cases, I need to be alert to the possibility of internal traffic -- the sites' owners or staff visiting the website. You should be alert to this, too, if you look after your own analytics.

Not only can numerous visits from your own workers make you look as though you have more traffic than you really do, they can also throw off other aspects of the information available to you:

• Where are your visitors coming from? The percentage from your service area may be smaller than you think, if a lot of your local traffic comes from your own office.
• How long are people on your site, and how many pages do they look at? It's great to keep your website open at your brick and mortar shop to give people information about your products and services -- but it can skew the averages on these important metrics.
• How do people use your site? The way you navigate through your own site is bound to be different from the way other people do.

Filter your own computer, your office or shop computer, and your workers' computers from your site's analytics. Then Google Analytics won't count any of your own visits.The little video above shows you just how to do it, and just how easy it is.

Analytics Results Can Be Deceptive

I think you know that I love web analytics. With data about your visitors, you can make strategic decisions based on information, not on guesses. You can tell whether your new promotion works or not. You don't end up making decisions on the basis of a conversation with one person who feels strongly (and I think we all know that businesses do that with startling frequency).

But that doesn't mean that analytics always give us a full and accurate picture.

Leave aside for a moment the question of whether anyone is actually trying to deceive you or not. Most businesses -- or at least most of the businesses I work with -- don't see a lot of malicious messing around with their web results. Instead, they see anomalous results requiring some kind of explanation. Here are some of the surprises you might see, and what they might mean:

• Surprising direct traffic. Direct traffic always deserves a closer look. If you have a simple, obvious URL (lucky you!), people may be just as likely to type it into the address bar at the top of the screen as to type it in at a search engine -- even if they've never visited you before. My friends at Onsharp (Onsharp.com) get lots of direct traffic, and it's fair to assume that many of those visitors are guessing correctly at the web address of their local web firm. But lots of direct traffic, or surprising patterns or changes in direct traffic can also mean that your staff hasn't been filtered out of your analytics properly, or that someone has been working on your website. Ask around the office before you start formulating any new strategies in response. 
• Self referrals. We've seen several examples recently of sites getting a lot of referral traffic from themselves. In one case, there were thousands of visits a month, so it was worth tracking down the path. Usually, you can safely ignore it. It's usually a shopping cart, a place customers check in -- some part of your website that involves some engineering.
• Surprisingly limited visits. One case last summer really stands out as an example of this. A website reported receiving exactly the same number of visits every day for months. It looked to me as though the code was installed on all the pages, and indeed when I kept saying, "This can't be right," the engineers all reported that everything was as it should be. After weeks of tweaking, we discovered that the analytics were picking up only the activity in the administrative part of the site -- where the business owner had a very methodical routine. More recently, we've been struggling with a site that seems to have visitors only on the homepage. Further examination shows that this is not the case -- so we're having to look at the analytics and figure out what the error is.

When seeing this kind of odd behavior at your analytics reports, be sure to consider that you might have technical issues. Once that possibility has been eliminated, dig deeply into your data-- and your real world information sources-- to find the explanation.

Your Website's Traffic

I've been doing annual reports. Traffic is the starting point for all of them, and of course everyone is happy to hear about their increases in traffic. Me, too. My traffic is up 198.92% over the same time last year. But don't stop there. Ask yourself a few more questions:

• Is it business, or is it just traffic?

I have one client who had an increase of 370.95% in 2009, compared with 2008. But he only serves local clients, and his local traffic increased by a mere 44%. That extra traffic is fine -- there can even be side benefits, such as a general increase in web visibility and prestige that could increase conversions -- but his basic traffic info suggests a higher level of success than we're really seeing.

• Are you seeing the trends?

The client below had a nice percentage of increase between 2008 and 2009, but really it's better than that. 2008 was essentially flat, while 2009 shows an upward trend (ever since they hired me) that is likely to continue if we continue making good decisions. The reality here is better than the percentage of increase would suggest.



The client below has a fairly new site, and the percentage of increase isn't that impressive yet. But the line on the graph is heading upwards. We might want to speed the process up, but the general trend suggests that we're on the right path, and shouldn't make a complete change in strategy.

• Have you broken it down?

Here's my chart for traffic from search engines.


If I look at my traffic over the whole year, I have a fairly smooth and steady increase, like the ones earlier in the post. But breaking it down by source shows a different story. My direct traffic is relatively flat. Search traffic shows a temporary peak in May when I was mentioned in the Wall Street Journal, and then a nice increase between July and August that stayed high till the typical holiday drop --and even then was considerably higher than it was to begin with.

Here's a site, launched this summer, that shows completely different profiles for its three sources of traffic:





I haven't been working on this site since its launch, but if I were, I'd need to be aware of the different paths visitors were following.

When you look at your site's traffic for 2009 and make your online marketing plans for 2010, be sure to look closely enough at your traffic data to get the information you need for strategic decisionmaking.

Need more basic info about website traffic? Here are some posts you might find helpful:

Website Traffic
Learn from Your Traffic Sources
Detective Work at Analytics

Doing Detective Work at Analytics

When you look at your web analytics, there are some basic things you'll want to examine:

• traffic -- you usually want to see it increasing steadily.
• traffic sources -- you'll want to notice what's bringing traffic and what's showing the highest conversion rate, so you can do more of that; you'll also want to catch any surprising results here, and see a steady increase in the number of different keywords bringing you visitors.
• interesting patterns -- you can often tell a lot about who you're reaching by noticing patterns in visitor information, heat maps, and traffic.
• spikes and changes -- you always want to find out what happened to cause these, so you can respond strategically.

But notice that most of these things you notice involve finding out more. And hooking up what you learn with real world information to see the full implications, too.

Recently, Josepha noticed that one of our education-related clients had regular small spikes on the 15th of the month -- teacher payday. Our hypothesis here is that teachers are using personal funds to shop there, not just school or grant funds. This means that some more personal approaches might work well.

Another client showed an apparent drop in visits to a particular page -- until we looked back and found a completely artificial rise in visits to that particular page, caused by the webmaster's failure to filter out worker's visits to the site. The page in question had been the subject of a lot of debate and discussion the previous month, so there were a lot of visits by workers.

Another client had a big increase in overseas traffic -- which turned out to be from a European equivalent to StumbleUpon. After thorough checking, we ignored it, thus saving the company from a lot of wasted time looking into serving overseas customers.

Moral of the story: look further when you see something interesting going on, and make sure that you know what's really going on.

Diary of a Website: Discovering Your Traffic Sources

We're continuing with our series of posts for people with new websites, or who are generally new to Google Analytics. Yesterday, we saw how to get in to look at the data, and had a look at the Site Usage section of the dashboard. It's good to have a look at that information on a regular basis.

The other part of the dashboard that you ought to get familiar with right away is the Traffic Sources section. You can click on the words "Traffic Sources" in the navigation in the upper left of your dashboard, or on the pie chart in the main section of the dashboard, which looks like this:





I wrote a post on the basics of traffic sources a few months back, so I won't repeat it, but I recommend that you click on that link and read it if you're new to Analytics. Essentially, the Traffic Sources section of your dashboard tells you how people found your website.

The pie chart example above is an established site with a nice amount of traffic. Myra Grayson, who has so kindly allowed us to keep a diary of her new website's progress, doesn't yet have that kind of traffic, and she has a different proportion of sources, too:



The established site has about half of its traffic from search engines, a bit more than a third from direct traffic, and much less from referring sites. Myra's website gets two-thirds of its visits from referring sites, and only a little bit from search engines.

The best balance of sources depends on your business, your goals, and a number of other factors. However, the fact that Myra gets very little search traffic so far shows that the search engines need to be alerted to her existence. We're currently doing a basic linkbuilding campaign for Myra's new site, so we'll expect to see the proportion of visitors from search increase. Visits from direct traffic may increase, too, though for a variety of reasons that might not be a major traffic source for Myra's site.

Let's have a look at her top referring sites:

• graysllandacres.blogspot.com (referral) 21.88%
• facebook.com (referral) 17.97%
• rebeccahaden.com (referral) 10.94%
• art.uark.edu (referral) 6.25%

Myra has 16 different referring sites right now (that number will increase as she gains more links), and four have sent her multiple visitors. Google Analytics tells how many visitors have come from each source, and what percentage that is of the traffic -- in this case, of the total traffic, but you can also see what proportion a given site provides of your referred traffic.

We can see that 21.88% of Myra's visitors are coming in from her blog -- a good blog is a good source of traffic, so we're glad to see that. She has people coming in from her Facebook page, too. It's likely that social media will be a good choice for her site. Even though the amount of data we have right now is small, Myra can already get some sense of what strategies she might choose to pursue to increase her traffic in the future.

She also has people coming from my website and from the designer's website -- those are the other two domains listed as referring sources above. Those people probably aren't customers for Myra. It doesn't hurt to have those visitors, but the fact that they send traffic doesn't mean Myra should try to get more links like those.

As she gets more visitors and more data, Myra will be able to use the information from Google Analytics to make decisions about her website, her marketing strategy, and even about her business.

Diary of a Website: Getting the Hang of Google Analytics

Myra Grayson's new website, GraysLlandAcres.com, has been live for a few weeks, and she wants to see who's visiting. Tom installed Google Analytics for her when we built the site. This is my personal favorite analytics program: it's free and it gives a wonderful depth of information.

I keep track of Analytics for some clients, but others -- including Myra -- want to do it themselves. While I do trainings for local people (like the folks at BabySmart Travel, who made cookies for the occasion yesterday), Myra isn't close enough for me to sit down with her at her computer.

So for Myra, and the owners of BabySmart Travel, and for you too, if you're wondering what all this talk of Google Analytics is about, here is the first in a series of lessons on using Google Analytics for absolute beginners.

Before you start, you have to have Google Analytics installed at your website, and you have to have access to your analytics. Ask your web master to do this for you.

#1: Get into your account. Here's the familiar Google homepage. In the top right-hand corner, you may see the words "Sign in." If so, click on them and sign in, or register. Chances are, you already have an account and you see the word "Settings" up there. Click on "Settings," and you'll get a drop-down menu. Choose "Google Account Settings."





Next you'll see a page like this one:




You'll have a list of stuff you use at Google. Your list may be longer or shorter than this one, but it's in alphabetical order, so "Analytics" will be near the top. Click on it, and you'll see your webpage (or webpages). Click on "View Report."

Now you'll see the dashboard:





#2: Get to know your dashboard. Myra is kindly allowing us to check out her dashboard, so let's look at one particular part of it. Here is the "Site Usage" section.



The blue line shows how many people came to see Myra's website. when you first open the screen, it's set to show you how many visits you've had in the past thirty days. Myra's site hasn't been live for thirty days, so she has a flat blue line until the day her site launched.

Actually, she had a few visitors in the days before it launched, while people were working on it and eagerly checking to see whether it was live and things like that. But on the first day it was live, she has a mountainous peak.

This is where she told all her friends and family to go look at her way cool new website. She announced it on Facebook and stuff like that.

Since then, she has had between 1 and 10 visitors each day. In fact, if we leave out the initial tell-everyone-to-go-look stage, she has an average of 5.5 visits a day, a fact she can discover by clicking on the word "Visits." It's in the upper left below the big blue line graph.

Here are some other things she can learn from the "Site Usage" section:

• Visits This counts the number of visits the site has received.
• Pageviews Pageviews counts the number of times someone looked at a page -- not individual people, necessarily, just the action of looking at a new page.
  
• Pages/Visit The average number of pages per visit shows how many different pages people checked out, on average, when they visited.
  
• Bounce Rate When someone comes, looks, and leaves without exploring your site further, they're said to have "bounced." The bounce rate shows the percentage of visitors who have bounced away.
  
• Avg. Time on Site Average time on site measures the length of time, on average, visitors stay and read. People often spend just a few seconds (12 is the widely-quoted number) deciding whether to stay or go.
  
• New Visits This metric tells you what percentage of your visitors were new people, and what percentage were coming back to see you again.

That's a good start.

Does Your Website Work Weekends?

One of the great things about your website is that it will go ahead and work for you while you're out doing other things. It will, assuming you've done a good job with it, show the best side of your business to visitors 24 hours a day, 7 days a week.

That is, if you have visitors 24 hours a day, seven days a week.

You can tell by looking at your analytics. Here's your dashboard. You can see the traffic to your site in the blue line.



If you put the cursor over the little blue dots, you can see exactly which day, and day of the week each is.

Look at a couple of examples and see the difference:

This site has minor peaks and valleys. Some of the peaks are on Thursdays, some on Tuesdays or Fridays. The traffic never goes way down. This site has traffic all week.




This site has valleys on the weekends. There is an enormous difference between weekday and weekend traffic every single week. Essentially, this site doesn't work on weekends.

What difference does it make? In general, if your site doesn't work on weekends, your customers probably don't either, and they're only visiting when they work.

In that case, your website can be an all-business kind of place. You can use the jargon of your industry. You should be serious enough that your customers' bosses can walk by and see they're working.

You might also be able to increase your traffic, and perhaps your sales, by posting something interesting on weekends, or even offering special offers only on the weekend.

If you have traffic all week, then people come to you whether they're working or not -- or they work weekends, too. You may want more of a Web2.0 feeling, a community aspect, a fun area. You should make sure that your site accommodates amateurs as well as pros at whatever you do. Or, if you're in an industry where people work on the weekends, you might want to take advantage of it by offering those weekend specials if your competitors don't.

It's just another useful piece of information about your clients.

Browser Capabilities at Google Analytics

Does this look familiar at all?

Maybe not. Even if you use Google Analytics to keep track of traffic to your business website, the Browser Capabilities section may be a part that you ignore.

You'll find this under "visitors" at your dashboard.




The colorful pictures above show the Screen Resolution, one of the choices under "Browser Capabilities." It's chock full of numbers, so if you aren't a numbers kind of person, you'll have skipped the whole thing.

Look back at it now. You'll notice that the largest number of visitors to the example site use a screen resolution of 1440x900. That's what I use, too.

If you're not sure what this means, then go to the control panel or system preferences panel of your computer and you can not only find out what screen resolution you use, but also change it around and see what that means. If you use Windows or Vista, you find the control panel by clicking on your Start icon, which is oddly enough the thing you use to shut down the computer. If you use Mac, your Apple button will take you to System Preferences.

So what? Well, if we look up at the charts again, we see that more than a quarter of the visitors are using a screen resolution of only 1024x768. There are even a few people who are using 800x600, which is like an antique computer of some kind -- or people with limited vision making their screens easier to read. There are also a few people -- those 320x396 people -- using mobile devices.

The most obvious application of this information is to design. On this example site, the designer and I were thinking about the header. There were some issues when I looked at it on my screen -- but when he showed me how it looked at the lower resolution, I saw that perfecting it for me -- and the 31.43% of the site's viewers who match me, plus the other 6% or so who use higher resolutions -- would cause issues for the 60% or so who use lower resolutions.

But this can also tell you something about your visitors. How many of them visit you on mobile devices may say something about the age or affluence level of your visitors. It may also say something about how visitors use your site -- do they look you up while they're out and about to find directions to your shop or to check out the conversations? Are they checking facts while in meetings? Or are your customers only coming to see you from their office computers?

The operating system your visitors use is another piece of information you can get. Here's our example site's breakdown:



The blue part is Windows: just over 64% of the visitors use Windows. A third of the visitors use Mac, and the little orange sliver is iPod.

This looks like an ordinary slice of well-off professional people using their home computers.

Here, for comparison, is an IT company I work with:


The blue section is still Windows -- a huge majority. Macs are the 10.37%, but the orange slice is Linux. There are also some skinny slivers of other things that are hard to see, but they add up to an amazing 80 different operating systems, including not just Chrome and Flock but Android and all kinds of other stuff, too.

The people visiting this website are probably mostly at work. And chances are they're IT guys. So the content of the second site can be much more techy than the content of the first site.

Have a look at your visitors' Browser Capabilities next time you're in Analytics. You may be able to get some insights into your visiting population that will help you not just with design, but with the kinds of blog posts and special offers you need to be creating, and the places you ought to advertise, too.

Learn from Your Traffic Sources

When you look at your dashboard at Google Analytics, one of the bits you should always notice is the Traffic Sources. The first view you have is a pie chart showing how your visits are distributed among Direct Traffic, Referring Sites, Search Engines, and Other.



Here's what those terms mean:

• Direct Traffic is people typing in your address or going straight to it from their bookmarks. Sometimes lots of direct traffic means that you have an easy to remember URL (good for you!) and it's just as simple for people to type it in as to search. Sometimes it means that you haven't filtered out your staff. If you have a high proportion of direct traffic, check to see whether it's new visitors, in which case you've got a good URL and people looking for you by name. Lots of direct traffic from return visitors is good news, too, of course, but filter out people working for you before you make plans based on that.
• Referring Sites is what you get when people follow a link to your website. This lets you see what kind of links send traffic, so you can build more of those. It also lets you know when you've gotten featured at Stumbleupon or Digg. And it helps you find sites that have linked to you without your having requested it.
• Search Engines refers to people who found you by typing something in at a search engines like Google or Yahoo or bing.
• Other is anything else, often email (though that can also show up in Referring Sites).

Click on "view report" under the pie chart and you'll get more detail:


You can see that these two examples show different patterns. The one above shows fairly steady traffic from all three sources. The one below doesn't have much direct traffic, but it's fairly steady through the week, while Referring Sites and Search Engines rose at the end.



As with most patterns, changes are often the most interesting thing. When you have your quick look at your analytics each day, changes should be what you're looking for. When you see a change, find out the reason for it. If it's a good thing, do more in that direction. If it's not so good, then it's time to change your strategy.

Bear in mind that the pie chart is about percentages. An increase in search traffic can show that your SEO efforts are paying off, or it can show that more people are looking for one of your keywords, or it can show that your direct traffic fell because you got around to filtering out the people who work on your site.

At the Traffic Sources report page, you can look at lots more data. For example, you'll see the Top Traffic Sources, which is a list in order of popularity of your traffic sources. You'll see the top five on the main page, and you can click on "view full report" to see all of them.

Google is very likely to be your #1 source. For most of my clients, a major referring source comes next. Direct traffic is usually in the top five. For many, social media such as blogs, Twitter, LinkedIn and Facebook turn up here.

Again, changes are what you want to notice. When I see a directory in this list, I suggest thinking about a paid ad at that directory. When I see a new source move into the list for the first time, I run right over and check its conversion rate for the site's goals. The information here can help you make well-founded strategic decisions.



The Traffic Sources report page is also the starting point for lots more detailed information. You can narrow down your focus and see where your direct traffic is geographically, or all the referring sites and their conversion rates, or what keywords people are using to reach you (you can use your keyword data strategically, too). You can also check the performance of your adwords campaigns from this menu.



Explore the Traffic Sources page. You'll find it a useful resource.

Website Traffic

Josepha's been doing basic foundational linkbuilding and social media for a nonprofit, and they've seen a 922% increase in traffic. Another client we're working with has a 215% increase in traffic. Another is up 28% after implementing only a few of my recommendations. We're happy to see these numbers; an increase in traffic is always good.

Yet last week I wrote about a client who has had a 600% increase in online sales in the past year, with only a slight increase in traffic. And I currently have a client whose traffic is down slightly, even though they've moved to the front page of Google for their top keywords.

Increased traffic is good, but it's not the only thing to look at. Here are some questions to ask when you think about your traffic:

• Are your visitors actually your customers? People visiting my website after typing in "internet service provider" probably aren't looking for the kind of internet services I provide -- they're probably looking for an internet hosting company. Increasing their numbers isn't going to do me much good. If your well-targeted traffic increases and your random traffic decreases, you can see improved results without much increased traffic.
• Are your visitors in your service area? International traffic is cool, but your lawn care service won't benefit from it. If you only work with local customers, then you should ignore traffic from elsewhere and look for increases in your local area only.
• Are your visitors taking action? It can take some time for people to move from visiting to taking action, but if you see increasing traffic with no conversion over a long period, then you're not getting the return on your investment that you need. This particular question can be hard to answer if you're not an e-commerce site, but you'll want to notice whether visitors move through your website the way you planned. Make sure that you're taking into account those who visit online and then walk into your shop. And of course with Pay Per Click it's all about conversions -- if you're paying for traffic and they're not paying you, then increased traffic isn't good.
  
• Are your visitors showing seasonal change? It's essential to compare apples to apples, not to oranges. The client I mentioned earlier who has had a dip in traffic is seeing a normal seasonal downturn. The one who has had a huge increase in sales but slight increase in traffic is up 49% over last month -- for Back to School -- but only 12% over last year at the same time.

Increased organic traffic is never a bad thing online. You don't pay for extra staff or higher electric bills from having visitors, even if they're not from your service area or not taking action. Larger numbers of visitors can increase your chances of gaining organic links or of drawing the attention of people who will become your customers. And sometimes there's a gap between when visitors find you and when they begin shopping with you or calling you.

But it's important not to focus on that single metric without looking at the others.

SEO Temptations

'The trouble with Google PageRank," said the fellow who had designed the website I was consulting on, explaining why he was resisting the idea of installing Google Analytics, "is that it tempts people to write stilted, unnatural stuff on their websites."

It's true that people are tempted to write stilted, unnatural stuff on their websites. Some even give in to the temptation. One of the local web design companies here where I live has a homepage that says something like, "If your Fayetteville business needs a website for Fayetteville business or any surrounding Fayetteville area business..." It does sound stilted. Stupid, even.

Other firms react to the temptation in other ways. One of the competitors of Fargo web design firm Onsharp has a bizarre paragraph that goes sort of like this: "If you want to find our website, you can find it by going to your favorite search engine and typing in 'Fargo web design, Fargo web designers, Fargo web firm, Fargo...'" This is a more creative version of the above, but no less stilted and no more natural.

Does this have anything to do with PageRank? It has never been suggested that keyword stuffing (that's what you call that kind of content) improves PageRank. Google has never recommended keyword stuffing. Installing Google Analytics doesn't lead to keyword stuffing.

Good web content is written with the search engines in mind. We have to remember that they are robots, and not able to interpret complex allusive stuff. They need to have the keywords -- the things humans will type when looking for your page -- right up there where they can see them and understand them with their robot brains.

We also have to remember that the search engines don't shop with us. When humans come to your page, they don't want to see stilted language. Even if they have never heard of keyword stuffing, they're going to notice that something odd is going on at your page if you're doing it. If they have heard of keyword stuffing, they'll recognize it and know that there's something shady going on.

This is true, but again, it has nothing to do with Google. It has to do with shady practices. Some people are tempted to indulge in shady practices. Since Google's PageRank is a measure of trustworthiness, there's no reason to suppose that shady practices will improve your PageRank. People who take up keyword stuffing can't honestly say that Google tempted them to do so.

What does the picture at the top of this post have to do with SEO? Nothing. Keyword stuffing also has nothing to do with good SEO practices. If you're considering hiring a web firm that uses this tactic, resist that temptation.


Stumble It!

Where Do Your Website's Visitors Come From?

One of the things you can learn by looking at your Google Analytics is where your visitors are when they come to your website. Just click on "Map Overlay" at your analytics dashboard, and you'll see a phrase like "3213 visitors came from 92 countries." You can then look closer and see the region, state, or city your visitors were in.

If it's a personal website you're looking at, you can respond to this with, "Wow, cool, people from Hungary come to see me!" If you've got a business website, this information can be more useful than that.

First, consider whether you can actually sell goods or services to people from 92 countries. If you have local business -- a brick and mortar store, a service that requires your physical presence, things like that -- then it may still be cool to have visitors from 92 countries, but you want the great majority of your visitors to be local. If they're not, then you need to do more linking with local sites, to encourage your actual customers to visit your website.

One company I'm working with right now sells chocolate. They're happy to ship, but not to tropical countries, and not to subtropical states like mine except in winter. So a preponderance of visitors from hot places would tell us that we're not focusing on the right geographical areas in our marketing.

If you have a national or a global reach, you can still benefit from the information. The school supply company I work with serves the entire country, but school calendars differ from one state to another. Seeing when New York's teachers start their Back to School browsing lets us target our marketing and plan for staffing and stocking needs -- if we relied only on the data from the local brick and mortar store, we'd miss those opportunities.

Watch for changes, too. A sudden spike in visitors from Milwaukee? Then you need to find out what happened there -- a radio show? a local mention of your name in the paper or of your product at a workshop? Find out so you can repeat the effect.

Finally, you can look more closely at a particular population's activity once they reach your website. Is the content your visitors from India choose to look at different from that most popular with your visitors from the UK? Do some countries have a higher conversion rate than others -- and if so, might you want to focus efforts on them rather than on the people who look but don't buy? Or do you just need to tweak your message to increase conversion from that other location?

The map overlay doesn't need to be a daily check, but it should be something you look at before your next marketing strategy meeting.

Stumble It!

Another Clever Trick from Google Analytics

I hope you use your Google Analytics regularly. They can tell you lots of useful things.

One thing that not everyone knows is that Analytics will show you what people click on at your website.


This is where you'll find that option: at your dashboard, under Content, with the label "Site Overlay."

Click on that, and you'll be able to see where your visitors are clicking at your site.



At my site, I can see that 17% of my visitors click on my "Clients" page. 10% click on my "About" page,which is just the right number. If I got any surprises here, it would tell me that something was off with my design.

This is easier than going through the Navigation summary, and great for quick checkups.

Try it on your own web site, and see what you can learn.