Creating a Mobile Profile in Google Analytics

Monday, December 12th, 2011

Google Analytics has been improving their mobile reporting but there are two main issues still:

1 – Mobile visits not being captured as mobile.

If the Operating System isn’t captured, it’s coming into Google Analytics as “(not set)” and doesn’t get counted as Mobile traffic. In my experience I’ve seen this come in as less than 1% of traffic to more than 3% of traffic. Not a lot on the whole, but if most of that is would be mobile traffic, and mobile traffic is already a single digit percentage of your traffic, it becomes more significant.

Then more recently, there have been issues recognizing iOS 5, as indicated in this report looking at Mobile > Devices  and drilled down to Apple.

2 – No easy filter to set up a mobile-only profile.

Even if missing a small amount of traffic wasn’t a problem for you, we don’t have a simple way to add a “mobile” filter to a profile. So there isn’t an easy way to create a profile that shows only organic search traffic from mobile devices.

Solution: Multiple Filters to Capture More Mobile Data

So, I’ve developed a set of filters that address both of these issues. In the test cases we used to develop this, the new profile captured between 0.5% and 2.25% more mobile traffic than Google’s Mobile profile. This was significant for one client we were working with, but may not be significant for everyone’s needs.

Here’s the summary (in proper filter order):

SEER – 1) Rename Windows CE
SEER – 1) Rename Opera Mini
SEER – 1) Rename Sony Ericsson
SEER – 1) Rename Nokia Other
SEER – 1) Rename Samsung Other
SEER – 2) Rename Unknown OS with Small Screen Size
SEER – 3) Include Mobile OS All RegEx

These filters rename the Operating System for several variations that should be considered mobile based on browser. Then capture a few more with an updated RegEx for small screen size. Finally, the last filter pulls everything together with one Include filter for mobile Operating Systems (including those we’ve renamed.)

Filter Rename Windows CE

These screen shots are in old GA because you can see more of the filter name.

This filter renames the Operating System to Windows CE when the Windows version is CE so our final include filter can capture it.

Filter Rename Opera Mini

I found anywhere from 10-30% of (not set) to use the browser Opera Mini. This filter extracts those and renames them “Unknown OS Opera Mini”.

Filter Rename Sony Ericsson, Nokia, & Samsung

Similarly, I found many variations for mobile browsers, that also had the operating system as (not set) and weren’t being captured as mobile:

These next three filters rename these as “Sony Ericsson Other,” “Nokia Other,” and “Samsung Other.”

RegEx: sonyericsson*

RegEx: nokia*

RegEx: samsung|gt-s*

Filter Rename Unknown OS with Small Screen Size

We’ve captured a lot already filtering by browser, but there’s still a chunk with both Operating System and Browser set to “(not set)”. So to capture the last of these, I used this method for capturing by screen size. But I updated the RegEx to capture more screen sizes, including the iPhone 4.

(^[1-3]?[0-9]?[0-9]|^4[0-7][0-9]|^480|^640)x([1-7]?[0-9]?[0-9]$|8?[0-6]?[0-9]$|960$)|(800×480)

Then using this filter rename the remaining “(not set)” that are also a small screen size: Unknown OS Small Browser.

Notably, this does not capture the Kindle 3 or NOOK (600×800) or the iPad (1024×768) because these tablet screen resolutions are similar to small monitors. This makes screen size one of the least reliable indicators of mobile traffic and a last resort in this series of filters to capture what couldn’t be captured by Operating System or Browser.

Filter Include Mobile OS All RegEx

Finally, we pull all of the renames, and other mobile operating systems, with one big Include.

Here’s the RegEx:

Android|BlackBerry|^iP|Windows CE|Hiptop|Ezweb|LGE?|MOT|Nokia|NTT DoCoMo|PalmOS|Portable|Samsung|SoftBank|Sony|SymbianOS|Vodafone|other$|Phone|^Unknown

Uses & Limitations

There we go! You can use this mobile filter set on its own to see mobile traffic, or mix it with other filters (keeping filter order in mind) to track other things, such as organic search. With a single mobile profile, you can take a closer look at how your mobile users specifically interact with the site. Some ideas to get you started:

• Conversion reports, the mobile users convert at a different rate or convert on different items?
• Organic Search, do mobile users come in on different keywords?
• Landing Pages, are your most popular landing pages optimized for mobile?

When using these filters, please keep in mind one big limitation: operating systems & browsers change. We started testing these filters back in September, and already had to make one change when “Windows Phone” started to come in as an Operating System on November 15. It’s a good idea to regularly check the profile against an unfiltered profile to make sure you continue to pull in as much or more traffic than is shown in the normal Mobile reporting. On the flip side, the small screen size filter also grabbed additional mobile traffic on Nov. 14 & 15 before Google added “Windows Phone” as an operating system. In addition, there are numerous one-off Operating Systems that may come and go, though if they are only accounting for a handful of visits a year, they shouldn’t significantly impact your numbers.

We’ve tested this profile with a few different clients, but if you test it too, let us know if it works for you.

Posted in analytics | 1 Comment »

Proof That Google’s Secure Search Now Affects More Users

Tuesday, November 1st, 2011

Did Google just take the next big step in rolling out secure search? It certainly looks that way.

Note: Skip to the bottom of the post for the 11/2 update.

Since our team went through the process last week to exclude all ”(not provided)” traffic from SEER’s Search profiles, the change was obvious this morning. Looking at individual sites, we quickly saw the trend of ”(not provided)” traffic increasing. Upon further analysis, “(not provided)” traffic increased for every site, but by varying degrees.

How big was this increase? The graphs below look at the impact of the change on 37 sites. Comparing yesterday (10/31) to the previous Monday, 27 of the sites had over 100% increase in ”(not provided)” traffic. We looked at the data in a second way, as well. Looking at yesterday’s visits compared to the average daily traffic driven by “(not provided)” last week, 30 of the sites had an increase over 100%.

The full breakdown across 37 sites is as follows:

The biggest percent increase observed was actually our own site. seerinteractive.com had a 625% increase in ”(not provided)” traffic, comparing 10/31 to last Monday. We also saw one site that received 2300 “(not provided)” visits in all of last week, but 1100 yesterday alone.

Key Takeaway – Google is rolling out the new secure search to more users, and we all need to be prepared for the outcome. 

For anyone who wants to run through the same process for your sites, here’s the methodology that was used across over 75 initial sites:

1. Gathered total “(not provided)” traffic from Monday through Friday last week
• Used Monday through Friday so lower weekend traffic wouldn’t affect the data
2. Pulled the ”(not provided)” traffic from last Monday and yesterday
3. Calculated average daily ”(not provided)” traffic
• Any sites that did not exceed an average of five ”(not provided)” visits per day in this time period were omitted from the remainder of the analysis.
• Although there were very interesting observations to be found in this data, the initial numbers were too low to provide a significant enough value, and we already had strong enough data from the remaining sites to complete the analysis.
• The remaining 37 sites span a wide variety of industries.
4. Compared yesterday’s (10/31) traffic to last Monday (10/24) and compared yesterday’s traffic to the average daily ”(not provided)” visits last week
• Calculated and graphed percent change for each

11/2 Update

This update was partially sparked by AJ Kohn and from wanting to share additional metrics to add context, especially with what we saw this morning.

Yesterday’s post jumped the gun. The big spike we observed was either the result of Google rolling out the changes midday Monday or releasing more users to secure search yesterday. Looking at yesterday’s data, saying the “(not provided)” numbers jumped is an understatement. Since we’ve already seen evidence that this is rolling out, let’s evaluate the actual impact rather than looking at data the same way we did yesterday.

The following chart breaks down the same 37 sites listed above, looking at the percent of yesterday’s Google organic traffic that came in with “(not provided)” in lieu of the actual keywords:

Out of the 37 sites, 16 had over 10% of the Google organic traffic come from “(not provided)” and one of the sites was at 21.05%.  Far more than the single digit impact that was estimated early on.

Did you see a big increase in ”(not provided)” traffic yesterday, as well? Share your thoughts in the comments below or via Twitter – @rachaelgerson.

Many thanks to @BrettASnyder for his help both pulling the data and going through ideas for this post, and for convincing me this should be a short blog post, not a quick Tweet. 

Posted in analytics, SEO | 22 Comments »

Go Beyond the Rankings: Talk Revenue with your Clients

Tuesday, November 1st, 2011

We’ve all been there before. You have a great linkbuilding strategy. You have every step of the process mapped out. You’re projecting 20 links and feel fully confident this strategy is enough to make an important keyword jump from 7 to 3 on Google.  Now you just need your client to make the investment for the strategy.

But there is one problem. They don’t see the value of the return on their investment.

Too often we get caught up in winning the rankings race, but not explaining to a client how it will directly impact their business.  We are all on the same page that increased rankings means increased visits means increased conversions means increased revenue.  But how much money are we talking here?

That’s why providing data to your clients that speaks in dollars and cents is critical.  In this particular example, I’m going to show how to get approval for that linkbuilding budget by showing wins in the revenue column.  Note: this strategy works particularly well with ecommerce clients that have transactions set up in Google Analytics.

Identify Keywords with High Average Values & Strong Conversion Rates

For larger e-commerce sites, it’s virtually impossible to track every keyword that builds revenue.  As SEO’s we often talk about “low hanging fruit” as part of a long-tailed keyword strategy to identify opportunities that we’re not always tracking/pushing to optimize.  Looking at a site’s past performance based on average values is just another way to attack the quick wins.

In this analysis I looked at one of our clients in the motorcycle apparel space.  For this particular industry, six months of data made sense because “hot products” tend to change relatively quickly.  In most cases the more data, the better.   But if your client sells products that have a quick turn over or are seasonal, set your data accordingly.

Along with date range, using the right filters will help with your analysis. Again, this all depends on the industry and your client’s product offerings.  You know their business, so set filters that are going to give you insight on real opportunities.

Next, sort by Average Value, determine the amount of rows you’d like to include and export your results to a CSV. Rather than explain how to quickly put this data together, I’ll refer you to Mark Lavoritano’s post on Striking Distance Keywords. It’s a great read and it shows you how to use RankChecker to quickly find where all of your potential opportunities are ranking on Google.

Let’s fast forward to the results.  Here I have a spreadsheet that shows keywords (sorry, had to block those out) with high average values, along with their respective ranking and conversion rate.  This additional data gives you even more insight on what will gain the quickest win. Off the bat I immediately identified three keywords that were long-tailed, had high average values with strong conversion rates and ranking within striking distance (Note: run a quick spot check on the keywords to make sure their positioning is accurate.  RankChecker is a great tool for proximity, but you may find slight discrepancies when doing a Google Chrome Incognito, Location USA search).

I Have My Data. I Have My Target Keywords. Now What?

Slingshot SEO’s recent Organic CTR Study has provided updated information that you should use to your advantage. While the AdWords Keyword Tool is also great measuring stick, using these CTR rates combined with actual data may provide a stronger analysis when pitching your client.

Let’s look at Keyword 1 and what we already know about it based on six months of data:

- It’s ranked 7 (let’s assume this as constant over the last 6 months)

- It has an average value of $490.32 (this can be skewed by outliers, so go back into GA to find out)

- It has an average 306 visits a month

- It has earned $2,043 a month

- Using Slingshot SEO’s CTR data, we can conclude that with a 1.88% CTR for keywords ranking 7^th, Keyword 1 has been searched 16,114 a month (Note: AdWords Keyword Tool estimates 14,800 exact match, global searches).

Now that we know the amount of searches a month, we can project what it’s worth to move up in the rankings.  If Keyword 1 ranked 3^rd, here is what we could expect:

- With a ranking of 3, Keyword 1 can assume a 7.22% CTR on 16,114 searches a month = 1,160 projected monthly visits

- Using Keyword 1’s conversion rate, ranking number 3 would bring 14.5 conversions a month

- Using Keyword 1’s Average Value, ranking number 3 would earn $7,122 a month

- Therefore, ranking number 3 for Keyword 1 would bring an additional $5K each month and an additional $60K a year

Drawbacks

Because everyone looks at data differently, it’s easy to see some of the drawbacks by using the data outlined above.

We’re looking at a fairly small sample size; what if one transaction skewed the data for the average value?  Go back and look to make sure it’s been relatively consistent. If not, find another keyword.

What if Slingshot SEO’s data isn’t accurate? Chances are, it’s not 100% and varies depending on the industry and the type of results. So play with the numbers and maybe use past data to form your own conclusions on CTR.

What if the keyword’s ranking fluctuated throughout the six months, further skewing the data? You’re probably right that it did have some movement.  But even if you were half off, you’d still see an additional $30K a year in revenue.

Keep in mind that SEO is not the only industry that uses data based on market research to make dollar-value projections. The underlying point is simple: Money talks. Speaking to your clients about rankings combined with revenue will help get you what you need to succeed.  Use keywords that have historically performed well, and show them a quick win where you can point out what it means to their bottom line by improving rankings.  Don’t just tell them you’re going to build 20 links.  Tell your client that building 20 links means moving a valuable keyword from 7 to 3 on Google, and increasing revenue by $60K a year.

The rest is up to you. You’ve shown the numbers, now you’ve got to show the win.

And, as the late great Al Davis once said, “Just win, baby.”

USE THE TOOL YOURSELF:

Special thanks to Mark Lavoritano for building an easy to use spreadsheet to share with everyone. As I was putting together this data, Mark was kind enough to put his excel chops to work and build a tool that involves very little work for the user and provides a quick snapshot of the impact on revenue based on keyword positioning. 

Download the tool here, add some data and let us know what you think.  Have a suggestion for improvement? Drop us a line in the comments and share your feedback!

Follow Ryan Fontana on Twitter

Posted in analytics, ecommerce, SEO, tools | 7 Comments »

Is (not provided) Driving You Crazy? Get Some of Your Data Back

Tuesday, October 25th, 2011

Google’s big announcement left many organizations angry about the loss of data. Since traffic now comes into Google Analytics as (not provided), we lost our ability to conclusively separate branded traffic from non-branded. Not only would we run the risk of over reporting the results of SEO efforts by allowing (not provided) to remain in our reports, but all of the other metrics, especially conversion rate, would also be affected.

Although we can no longer see the exact keywords searched, we can still pair the (not provided) information with the landing page for each visit. We want to pair this information because the landing page topic will give us a good indication as to the keyword used, if the keyword wasn’t branded.

There are three options below for how to get this data, with the first option being the easiest and the third requiring administrator access to your Google Analytics account.

Note: If you already have a good way to get this data, skip down to the “Next Steps” portion below.

Option 1 – Create Filter to View Keyword + Landing Page

Setup:

Step 1 - Add Secondary Dimension Landing Page

Step 2 - Filter to Include Keyword Containing not provided

•  Pros – Easy to set up
• Cons – Manual, needs to be set up each time, Fast-Access Mode

Option 2 – Create Advanced Segment to Include Only not provided

Setup:

Once this segment is created, it can be used across multiple profiles, and can be used on almost all reports in Google Analytics. So to get to the landing page information, just get to the Landing Pages report.

• Pros – Easy to set up, can be used across most reports in GA
• Cons – Fast-Access Mode

Option 3 – Create a Brand New Profile to Include not provided

Create a profile that replicates all of your existing settings and goals.

Tip: See #3 in this post for information to automatically copy your profile settings into a new profile.

Create this filter to only include (not provided) traffic:

• Pros – No Fast-Access Mode issues, data is kept completely separate
• Cons – Cannot be done if user doesn’t have admin access, creates another profile to monitor

Next Step

Regardless of which of the three options you selected, you can use the results for the next step.

You need three key numbers from your analytics tool:

• Total Organic Visits
• Number of (not provided) Visits
• Number of Organic Visits that were branded

If you have these three numbers, you know what percentage of the site’s Organic Traffic is branded. If we apply the same percentages to our (not provided) numbers, we are left with a reasonable estimate of how those numbers break down. This can be done on the site level, or on an individual landing page level.

• For example, let’s say a site has 1000 total Organic Visits, 42 (not provided) Visits, and 764 of the Organic Visits were Branded. If we assume these percentages to be a relatively accurate representation of this breakdown for the site, we can extrapolate that 33 of the (not provided) visits may have been Branded and 9 Non-Branded.

Also keep in mind that sites receive branded traffic to different pages. For example, in looking at ten sites, we found the majority of the branded traffic lands on either the homepage or a Login page. This is why you want to break out Landing Pages separate from the Overall numbers. With our downloadable spreadsheet, you can break out the traffic to a site’s top Landing Pages to determine if any branded bias exists for your site, as well.

Want to run the same analysis yourself? Get the spreadsheet here.

Tip – You can use this RegEx for most sites to find the branded breakdown for homepage and login pages in GA: ^\/$|^\/index|^\/default|^\/home|login

Finally, create an alert an Alert to notify you when traffic from (not provided) increases by x%, with the percent varying based on the amount of this traffic coming into the site you’re evaluating. For example:

What is SEER Doing?

We’re in the process of editing all of our Search profiles to Exclude (not provided) traffic. We recognize this means we’re under reporting our efforts, as some of the traffic we’re excluding is non-branded, and that numbers will continue to decrease as Google rolls out the changes to more users. However, in the interest of keeping our reporting as honest and valuable as possible, we would rather under report than over.

What are you doing to combat the changes? 

Posted in analytics, SEO | 9 Comments »

10 Google Analytics Browser Extensions to Save Time

Friday, September 2nd, 2011

There’s only so much time in a day, which is why we always look for ways to automate repetitive processes.  When we could focus on developing tools to scale link research, writing macros in Excel to evaluate data, or hacking APIs, why waste time on basic, redundant tasks?  Browser extensions are a staple for time-saving, so check out these 10 extensions to save time while improving your use of Google Analytics.

1) Analytics Helper

Good for:  All Google Analytics Users
GA version: All
Browser: Chrome
Download: here
Disclaimer: Tool has occasionally given false negatives, particularly about issues with the code, but never gives a false positive. 

What it does: Notifies users immediately to whether the Google Analytics code is detected on a page while browsing. Includes the code type (asynchronous, traditional), any UAs present on the page, and a quick blurb on the code position.

2) Better Google Analytics

Good for:  All Google Analytics Users
GA version:  Old version
Browser: Firefox
Download: here
Disclaimer: Original userscripts don’t all function. 

What it does: A compilation of Greasemonkey userscripts by Erik Vold, Tyson Kirksey, and other talented developers, including some favorite functions:

• Add Social Media Metrics
  
• Export to Google Docs
  
• Cleaner profile switching – Keeps the date and report intact when you switch reports.
  
• Access “Profile Settings” from any screen (also seen above).

3) Copy Profiles in Google Analytics

Good for: Google Analytics Power Users
GA version:  Old version
Browser: Firefox
Download: here
Disclaimer: Filter order is not maintained when profiles are copied. Doesn’t copy Event goals that were set up in GA v5. When working on a profile with Event goals already set up, Paste functions times out when it gets to that goal.

What it does: This is a Greasemonkey script that makes it easy to replicate Google Analytics profiles. When goals, filters, users, and settings are 90% the same across 10 profiles, why waste time creating the same initial profile 10 times? This script literally lets users hit “Copy” and “Paste” to have the initial work automated.  Plus it’s really fun to watch while pasting!

4) Goal Analytics Goal Copy

Good for: Google Analytics Power Users
GA version:  Old version
Browser: Firefox
Download: here
Disclaimer: Works on up to 5 goals. Doesn’t always save the “Required step” checkbox. Doesn’t copy Event goals that were set up in GA v5. 

What it does: When profiles are already in place, #2 above is not a good solution, but neither is manually creating the same goal a dozen times.  Solution? GoalCopy Firefox extension! Copy individual goals from one profile to another.

5) GA Bookmarks

Good for:  Google Analytics Power Users
GA version:  Old version
Browser: Chrome and Firefox
Download: here

What it does:This tool also copies goals and filters, as well as profiles, by creating easy bookmarks. (Prior to the creation of #6 below, I always used this for Chrome. Now I use this tool primarily to set the filter order across multiple profiles.)

6) GA Copy & Paste (aka My New Favorite Plugin)

Good for:  Google Analytics Power Users
GA version:  New version
Browser: Chrome
Download: here
Disclaimer: Not sure what the limits are yet. At one point, I had 11 goals and six filters saved. This tool is still being worked on.

What it does: Similar to #4 above, GA Copy & Paste is a brand new Chrome plugin. Props to Educardo Cereto of Cardinal Path for this tool! Goals and filters are ridiculously easy to copy and paste, even easier to delete. The appearance is clean and perfectly matches GA.

7) gaSwitchr

Good for:  All Google Analytics Users
GA version:  Old version
Browser: Chrome
Download: here

What it does: Switch between profiles, but keep the report and date settings.

8) Straight to Google Analytics

Good for: All Google Analytics Users
GA version:  All
Browser: Chrome
Download: here

What it does: Jump right into Google Analytics without clicking the “Access Analytics” button.

9) RegEx Checker

Good for:  Google Analytics Power Users
GA version:  All
Browser: Chrome
Download: here

What it does: Handy little browser add-on Regular Expression checker to use when creating filters and goals.

10) Google Analytics Debugger

Good for:  Google Analytics Power Users
GA version:  All
Browser: Chrome
Download: here

What it does: Check your analytics code for errors without waiting 24 hours.  Detailed explanation from Chris Le can be found here.  (Firecookie can be used similarly for Firefox, but I never use it.)

Bonus – Add-on Compatibility Reporter

Good for:  Everyone
GA version:  N/A
Browser: Firefox
Download: here

What it does: This last tool isn’t GA-specific, but is probably one of the biggest lifesavers on the list.  Looking at the Firefox extensions, they aren’t all compatible with the latest version of Firefox. However, I’m not willing to lose my favorite extensions when Firefox updates. The Add-On Compatibility Reporter extension prevents Firefox from checking compatibility. Almost every extension keeps working, no matter which version of Firefox is being used.

Were any of your favorite tools missed? Share them in the comments below!

Posted in analytics | 15 Comments »

3 Algorithm Changes & 3 GA Changes in 30 Days

Wednesday, August 24th, 2011

From Panda to 12-Pack Site Links and GA changes in session tracking and image search, can you trust any GA numbers?

Well, yes. Just because you read about a change, doesn’t mean you’re affected. It is always best to check your own numbers when you consider any of these type of changes. And as Brett covered in his post last week, GA can give you deceptive results. The nice thing is, you can often find these anomalies within the GA data itself with a little digging.

Your Actual Image Impact May be Minimal

At SEER, we started to see some increases in July around the time of Panda 2.3 & the image referral change. We didn’t want to dismiss these increases as image referrals without checking our own numbers first. After reviewing about a dozen clients who experienced visit increases, I noticed that the image referral impact for most of our clients was very minor! Despite the dramatic look of the referral drop graphs, many clients had a very small amount of traffic in this area. Consider this:

Yes, an 80% drop is good confirmation of that dramatic drop graph, but wait a minute, that’s only 47 visits that potentially shifted to Search. This particular site had about 48,500 search visits those weeks. That is a 0.10% impact. Most of our clients had impacts like this or even less.

Looking at Trends to Measure Real Impact

Of course, we did have some clients with a larger impact within the images that we didn’t want to dismiss. To look at these, I stacked the Organic Google Search Visit numbers with the Image Visit Numbers in Excel. Notice that the net effect is still up.

Using Unique Visitors to Assess Panda & 12-Pack Impact

For the session tracking issue, SEOMoz already covered looking at Pageviews and Unique Visitors to see through the GA session tracking issue.

Applying a similar concept, to look at the impact of the 12-pack I created a very simple custom report that looks at Unique Visitors first by Source, then drills to Keyword. (If you don’t want to make it yourself you can use mine.)

With just one week of data, we see the site brand is getting a slight boost, and brands they carry are taking a slight hit. Our overall Google impact is about 1% so far, but to really draw any conclusions we’ll need a few more weeks living with the 12-pack.

Check Your Numbers

As I said at the beginning, this post started because we were checking our own numbers against the issues brought up by other bloggers. Every client we looked at had a different impact for each of these 6 changes in the last 30 days. Once you’ve checked your numbers, leave a comment below and let us know what you’re seeing.

Posted in analytics, google, SEO | No Comments »

Beware: Google Analytics May Be Lying to You

Wednesday, August 17th, 2011

Many people have noticed and commented about the recent changes to how Google Analytics handles how sessions are reported or how image referrals are tracked.  As part of reviewing the effects these changes may have had on our clients, I wanted to dig into our Analytics to make sure that our data is as accurate as possible.  One of the great things about working at an agency is you have tons of data sources in all sorts of industries and verticals, so identifying systemic issues becomes much easier.

However, when digging into this data I found some startling things results that date back well before the above-mentioned changes went into effect.  To be clear, the below examples are not (in my humble opinion) a result of or even correlated to the recent changes to Google Analytics, but are certainly something you should review when validating your past and future numbers.

Referring Keywords

Now, the first place I turn when validating data is always Google Analytics, the source of data (imagine that!).  When I was taking a look at some data for a client that saw increases as of late, at first things looked great; our targeted keywords were moving in the right direction and there were no abnormal spikes across one or two keywords that would indicate spam or faulty reporting.

But just as I was cinching up my party hat and lacing on my dancing shoes, I noticed some strange long-tail terms driving traffic.  Take this example:

Not only is this an 8-word phrase that is only vaguely related to the client in question, the grammar is all out of whack.  In fact, Google even tried to correct me when I search for this:

So now that this has caught my attention, I wanted to dig a little deeper.

Please Note: I recognize that 25 visits is not something that will necessarily make-or-break our reporting.  However, this is just an example of how this phenomenon was observed over a bunch of keywords across multiple clients. 

Now, we know that we recorded 25 visits over the past month, but how were these split up?

Hmm, ok.  Already I’m scratching my head but at the same time, our judicial system requires us to prove beyond a shadow of a doubt, so maybe a news story or a algorithm shift got us some temporary rankings for this phrase.  Skeptical as I was, if these visits were legitimate, you would think there would be some variation in the source, right?  Well, according to the same Google Analytics that tells us we got a spike in visits for this keyword, all of these visits came from the same City:

Note: When changing our location to Castro Valley, the website did not appear until the #11 result on Google Chrome Incognito (to prevent personalization bias).

The visits also all came from the same browser:

And even the same Screen Resolution:

So now that I’m pretty confident there are some issues here, I wanted to give 2 more examples of where we’re seeing this issue arise under different circumstances.

Do I Have a Stalker?

In a client call earlier this week, we discussed the curious fact that one of the top traffic-driving keywords for the entire month of June was the name of one of the Sales Executives.  Now, if this were the CEO or Directory of Marketing that just spoke at a conference or was featured on the news, that would be one thing.  No, this was a Sales Executive who herself was a little confused as to why her name would drive so many visits to the corporate website.  When looking into the data, we saw a lot of the same consistencies we referenced above: same browser, operating system, city, screen resolution, etc.  However, what was different about this example was that the visits were actually dispersed over several days:

Let’s Put it to the Test!

To figure this out, the client performed a test; they copied a random block of text from their homepage and clicked through to their listing in Google.  What happened next gives us the greatest cause for concern: over the course of the next week, the client did not close the tab by which he had accessed the SERP listing.  To be clear, he did not revisit this tab, refresh the tab, or re-search for the query in Google.  However, over the course of a week the 1 visit that he made as a test was actually being recorded as 27 visits in Google Analytics:

You will notice that there were no visits recorded on Saturday and Sunday, when he was out of the office and not actively using his computer.  While we’re not certain what caused this activity, there are some speculation that it may have been an auto-refresh in the browser or that the visits were recorded as the computer recovered from “Sleep Mode,” but we have no confirmation to this point (We are, however, currently testing if this might be the case)

Once the tab was closed and the cache was cleared on Friday, July 22, these visits were no longer being recorded and the traffic for the keyword returned to zero.

They Just Can’t Get Enough!

The last example we have shows how we can look to recognize potential faulty reporting by looking for abnormalities in your metrics.  The term in question is one that we’ve been tracking for quite some time, so we know for a fact that rankings on this term have not moved over the course of the past 2 weeks.  However, this week we saw visits spike with no discernible explanation.  The Sherlock Holmes in me was, once again, intrigued.  How did a term that experienced no rankings movement, no notable new references, and no changes in search activity that I could distinguish spike so much?

Rather than rehash the above points, I suggest you also take a look at any abnormalities in pages per visit or average time on site, as these will show you places where there was unusual (and potentially incorrect) reporting of the activity taken on your site.  Normally, we wouldn’t place too much stock in these metrics but for the sake of this argument they do provide a valuable insight into shortcomings of Google Analytics.  You will notice from the below that we saw absurd increases in these metrics:

The biggest thing to look at here is the comparisons to the site average:  if we averaged 55 pages per visit, this wouldn’t be remarkable.  However, a 1300% increase in pages per visit and almost 1600% increase in average time on site is even more of an indicator that there is something wrong with how this data is likely skewing your results to appear more favorable.

So what does this mean for you?

Before you celebrate monumental wins (or losses for that matter), you should always validate the data.  Just as we should always consider “what strategy may have caused this?” we should also consider what reporting deficiencies may have caused it.

Another suggestion is to look at Unique or Absolute Visitors, which from our review seem to provide more realistic numbers and are not affected by (for example) the changes to how sessions are being recorded.

Finally, always make sure you’re looking beyond your top 3, 5, or even 10 keywords.  While a reporting error of 25 visits (as with our first example) may not be a huge deal, a 25-visit error across 50 or 100 keywords can, in the aggregate, seriously hurt the validity of your data.

Are you seeing similar results in your data?  We encourage you to share your thoughts or feedback in the comments section below.  If you have specific questions you can also feel free to reach out to Brett on Twitter.

Posted in analytics, SEO | 8 Comments »