• Tweet

As you already know, Google no longer provides organic keywords used to find your site when someone is logged into their Google account. Since Google keeps users logged in by default, the number of “not provided” keywords is just going to increase. Mine is currently at 26% of Google traffic and is bound to get worse. Especially with 600,000 new users joining Google Plus every day giving with more of a reason to stay logged into their account.

Having keyword data is extremely important to your site’s growth, especially if you are measuring conversion data. After all, know what keywords are actually converting can have a huge impact on your bottom line especially if you run an ecommerce site. We can talk more about conversions later, but that is not the point of this post.

The fact is that Google is a business, and Google is going to do what is best for them even if it hurts your business. So, as my grandma would say, “Don’t put all your eggs in one basket”.

If the recent change demonstrates anything, it demonstrates that Google cannot be trusted and you should have multiple streams of traffic and revenue. Its time to stop being one dimensional when it comes to traffic. So if Google really messes your site up or stops sending traffic you can live without them.

Again, not the point of this post, but worth saying.

Data Sampling

Data sampling is used in many cases to create snapshots of data. It stands to reason that if 25% of your data is missing, it may resemble the other 75% of your data. Brick and mortar retailers as well as the old fashioned catalog companies have had to rely on this method to determine what advertising is working and what is not working. How?

• Focus groups
• Coupons
• Questionnaires
• Tested Advertising Methods is an excellent book that explains this process in more detail for traditional retailers that is easily translated to an online company

I don’t want to give you a lesson in statistics, mainly because it will show just how little I know of the subject But, basically you can segment your traffic into different categories and make assumptions on the 25% of traffic you can’t see.

For our example you could simply say that you want to divide your keywords into three categories:

1. Head keywords: more than 100 visits per month
2. Fat middle keywords: 11 – 99 visits per month
3. Long tail keywords: 1 – 10 visits per month

You can define your keywords anyway you please, but the point is that you can still make an educated guess as to your hidden keywords by segmenting this traffic.

1. Download all keywords into a csv file for the chosen time period. You may have to make multiple downloads. Thanks again Google.
2. Take out the traffic numbers for the “Not Provided” traffic
3. Use Excel’s table options to segment your traffic totals by category
4. Determine each segments percentage vs. the total traffic (without the “Not Provided” traffic)

So, if head keywords make up 40% of your traffic for the visible keywords, they probably make up 40% of the traffic for the “Not Provided” keywords. Get it?

You may have this make up:

• Head: 17%
• Middle: 33%
• Long tail: 50%

You can assume your “Not Provided” keywords have a similar make up. That’s fine and dandy, but I see a real issue here. The fact that long tail makes up most of the traffic and by definition may only be sending a few visitors each month. So, there may be long tail traffic hidden in your “Not Provided” keywords.

NOT PROVIDED + LONG TAIL = MISSED OPPORTUNITIES

So, how can you make sure you are not missing out on these opportunities? Your history may help…

Your historical data is extremely valuable

Since you can’t get some of your new data, your historical data becomes a ton more valuable. Why? Because you can create a snapshot of where your site stands compared to past history and improve on that position. This is something you should be doing anyway, but just in case your not, here you go.

We are going to focus on low hanging fruit and new opportunities. The first thing you want to do is pull all of the long tail data out of Google for an extended period of time. For example, all searches that have sent 120 visits or less for an entire year. Once you have that data in a spreadsheet, tedious work begins. Here is how I do it, ready for another <ol>?

1. Pull all keywords out of Google Analytics for the past 12 months containing all keywords. Have to combine spreadsheets typically to get them all into one.
2. Filter out all keywords that do not meet the criteria (sent more than 120 visitors in the entire year)
3. Create a separate category in the table for search rankings. Run your favorite rank tracking software and determine the current ranking of all keywords.

The keywords that currently appear on the first page but do not appear in your last months keyword data probably means its a part of the elite “Not Provided” keywords. You can really use this data to pick off some low hanging fruit if you like by the way

Landing pages

Even though Google hides the keywords, they don’t hide the landing pages. So, if you click on the “Not Provided” keyword and select Landing Pages  as the secondary section you can see where your “Not Provided” traffic is going.

You can do a little backwards investigating and look at the keywords for each landing page that are provided and create a profile for each page of keywords that are sending traffic.

I like to look at all of those keywords that are sending traffic and run them through a rank tracking script. Basically, I am looking for keywords that are on the bottom of the first page, or the top of the second one, since those keywords are going to provide spotty traffic and can easily be improved.

Rank Tracking

I actually don’t didn’t spend a ton of time running rank tracker reports. Actually, I hardly ever did for my own sites and would only occasionally do them for my clients. Since we can’t see all of our traffic the only real option is to see our ranking for as many keyword variation’s as possible.

What about Google Webmaster Tools? Do you really trust that? If so, more power to you. I don’t, I prefer my own ranking reports that depend on the good favor of Google.

I will write a post in the near future explaining how to scrape Google’s search results as many times as you want. Well, as long as you have access to multiple IP addresses.

In conjunction with wordtracker, wordstream, or SEMrush you could get great data just from ranking reports. I actually have started seeing some really great opportunities when I was just relying on Google Analytics.