Making Google Analytics Useful for Big Sites

Managing a Google Analytics account for a large site with a long tail of content can be frustrating. In my experience, the biggest reasons are sampling and the dreaded 50,000 limit where GA only reports on your first 50k URLs hit each day, lumping everything else into an “(other)” bucket. While you can make both problems go away if you drop $150k on a Premium Account, I believe you can still get a lot of value from the free version regardless of your site’s size

Specifically, I wanted to share a few techniques Redfin uses to get meaningful roll-ups of our traffic that is spread across tens of millions of different properties, each with their own URL. We like to roll traffic up by market, price band, and whether that area is serviced by our own agents or our partner agents, but the technique should be equally useful for any site with a huge number of pages.

 

Step 1: Parameterizing Your URLs With _trackPageview

Let’s pretend we’re an ecommerce site with product pages that look like this:

www.example.com/toaster-oven-model-a100/123456

That’s a great URL for SEO purposes since it’s short and includes the product name. However if you’ve got a million products, it’s difficult to look across all of your product traffic and describe trends if all you’re sending to Google Analytics is the actual URL. Thankfully, GA’s _trackPageview() has an optional page URL parameter. Using that, let’s instead pass GA an URL with a few parameters we care about that looks like this:

/productPage?category=home&subcategory=appliances&manufacturer=GE&priceBand=50-100&product=toaster-oven-model-a100&productID=123456

We haven’t lost any data; our unfiltered profile still tracks a unique URL for each product. Now you can create custom filtered profiles to get useful views on your traffic

 

Step 2: Create Filtered GA Profiles

Now that we’ve parameterized our product pages, we can create additional profiles in GA that filter to specific views of those pages.  A few profiles we might create are:

  • Views: Products by Category/Subcategory
  • Views: Products by Manufacturer
  • Views: Products by Price Band

For each profile, you’ll want to create an “Include” filter that only includes traffic to products pages. That way, traffic to your homepage or forums won’t muck with your reports.

 

Once you’ve got your filter set up, under the “Profile Settings” tab you can specify which URL parameters you want to exclude. Here’s where the magic happens.  Exclude everything you don’t want to roll-up traffic by and what you’re left with is a neatly rolled up traffic report that falls well under GA’s 50k limit for unique URLs.

In the screenshot below, I’m excluding everything but the category and subcategory parameters. If I browse to Content > Site Content > All Pages in Google Analytics, I can quickly look at my top product categories and even see how traffic to a particular category has changed over time.

Note that GA filters don’t work retroactively, so you’ll have to think ahead a bit to decide which parameters you want to pass to GA and which views of the data you’ll want going forward.

Discussion