UPDATE: For a more recent (and better) post on the topic below, see my November 2010 post Track SEO with Google Analytics.  It includes a video and a link to a companion article that demonstrate a much more thorough approach to using filters than what I describe below.

———

We’ve recently found the “filter” feature in Google Analytics to be a great way to measure effectiveness of our SEO and other internet marketing efforts for clients over time. In this post I’ll explain why and give an overview of how to use it.

The problem with unfiltered reports and totals

If you’re like most small businesses using Google Analytics, you spend a lot of time looking under Traffic Sources looking at the Referring Sites report and the Keywords (non-paid) report. These reports tell you much of what you need to know about where your traffic is coming from — and tracking their total numbers over time can give you a good sense of how well your marketing is working.

The problem is that your Referring Sites report will often be dominated by a small number of sites, and your Keywords report will often be dominated by a small number of keywords relating to your brand name. These sites and keywords are critical, but as you grow, your marketing and SEO efforts will focus more and more on traffic from new sites and new keywords.

Let’s say you’re an employment lawyer in DC named “Teresa Millbrook.” Six months ago, both your Referring Sites report and your Keyword report were very top-heavy:

  • Two sites – usalawyers.com and dclaw.com – were driving 95% of your referring site traffic
  • A handful of “brand” keywords (i.e. keywords related to “teresa millbrook”) were driving 95% of your organic search traffic

Since then, you’ve reached out to a bunch of relevant websites to boost your inbound links and optimized your website for non-brand keywords like “dc employment lawyer”. Now you want to know what impact your efforts have had.

If you’re like most people, you’ll scan through your reports to see what new referring sites and keywords pop up, maybe comparing back to your list of target lists. If you’re more advanced, you’ll compare the total visits from referring sites and search engines 6 months ago to today (e.g. October vs. March).

The problem with these approaches is they miss the long tail, which is the most likely place you’ve had success. Sure, we all want to see a new referrer or keyword sending hundreds of visits, but the far more likely scenario from SEO efforts is that a large number of new sites and keywords are each sending you a tiny number of clicks. In aggregate, these new sources of traffic may be sending you hundreds or thousands of clicks, but because many of them are only sending you 1-3 clicks a month, they’re easy to miss if you’re just scanning your reports.

You may notice the trend if you compare total referrer or search visits over time, but you’re doing an apples-to-oranges comparison. What if your traffic from usalawyers.com and dclaw.com dropped by 25% over the last 6 months? You might miss the increase from your new referrers. Or what if your traffic from brand keywords increased by 25% over the last 6 months? You might mistakenly attribute the jump in your total visits to your non-brand SEO.

How to use filters

The only way to get around the problems above — and gain a true measure of your SEO and marketing effectiveness — is to use filters. Filters allow you to do an apples-to-apples comparison of long tail sites / keywords before and after you implemented your marketing efforts.

If you’re Teresa the lawyer, here’s how you’d use filters in Google Analytics to measure your Referring Sites traffic growth:

  1. Under Traffic Sources, click the Referring Sites report.
  2. Select your “before” time period, let’s say it’s October 2009.
  3. Click the Advanced Filter link on the bottom of the page.
  4. Keep the filter on “Source”, change the drop-down to “Excluding”, and enter “usalawyers.com”.
  5. Click “Add new condition”, choose “Source”, “Excluding”, and enter “dclaw.com”.
  6. Click “Apply Filter”.
  7. Write down the Total Visits number.
  8. Select your “after” time period, let’s say it’s March 2010.
  9. Repeat steps 3 through 7.
  10. Compare the Total Visits between the two periods.

Then you’d use the same basic process for your Keyword report, filtering out keywords with “teresa”, then keywords with “millbrook”. If your brand name gets a lot of misspellings, include those too (e.g. “theresa”, “millbrooke”); in some cases, you can avoid having to include all of the variations by just shortening a keyword to a “base” version that includes many misspellings (e.g. “mill”). (Just check to make sure doing this doesn’t cause you to count traffic from non-brand keywords that share that same base.)

Final Thoughts

The next time you’re in Google Analytics, try using the 10-step process above. Hopefully when you’re done, you’ll see a jump in the Total Visits from “before” to “after” and you’ll gain new insights into what is causing the increase.

Keep in mind that the comparison this process enables isn’t perfect; there may be seasonality and other factors at work. But it is much more scientific than the more commonly used approaches: eyeballing and comparing totals of raw traffic reports. By using filters, you’ll gain a far more accurate view of what’s working and what’s not in your marketing efforts.