Dimensionator is Dead, but DMA lives in the New GA
Geographic analysis in Google Analytics is one of the most useful features. However, reporting based on “city” can be very misleading because it is too narrow. Likewise, reporting by “region” (State, in the US) is too broad. Google Analytics lacks Zip Code resolution, so what’s left? It would be great if Google Analytics reported by DMA’s (designated marketing area) – the geographic definition of media markets in the United States that have been used almost universally by marketers for decades.
A few years ago now I wrote a post revealing how you could access DMA data in Google Analytics. Yes, that’s right, Google Analytics has “market” area data for GA, and has for many years. The issue is, they just don’t expose it in 99% of the GA reports where it would be so useful to have!
I even went so far as to have one of our engineers build a nifty little tool we called “dimensionator” to make accessing DMA and many other “hidden” dimensions in Google Analytics quick and simple. The only problem is… Dimensionator doesn’t work in the new version of Google Analytics, and it’s looking like quite a challenge to make it work (sidenote here: if you’d like to see dimensionator live on, let us know on Twitter – enough interest and we might just ba able to bring it back).
So, what to do? There’s an awesome treasure-trove of Market area Geographic data locked away in Google Analytics so how can you unlock it? I’ll show you… and this trick works in the NEW GA too!
The Problem with City vs. Markets in Google Analytics
If you rely on reporting by “city” in Google Analytics, you’re probably severely under-counting traffic, or you’re going through great pains to try and get the right data. Here’s an example of the vast difference when reporting the “city” of Los Angeles vs. the Market Area of Los Angeles.
“City” Report for Los Angeles in Google Analytics finds just 45 visits from L.A.
“Market” Report for Los Angeles in Google Analytics finds 120 visits from the L.A. market area in the same time period.
How Geo-Resolution Works
The reason why there’s such a difference is that the geographic resolution engine in Google Analytics relies on a method of reverse-lookup against the IP address of visitors in the back-end (note: IP’s aren’t made available anywhere in Google Analytics – they’re solely used in the back-end on a sort of throw-away basis as near as I can tell). That lookup determines the Point of Presence (a key part of how the Internet works) location for where that IP address is registered. This could be a business address if it’s an IP assigned to a dedicated customer, or a consumer ISP’s network POP. In the case of consumer ISP’s like Cable and DSL providers, those POP’s (points of presence) can be anywhere from mildly related or wildly unrelated to where the visitor actually is.
For example: I’m sitting in a Starbucks in downtown San Francisco as I write this, and whatismyip.com says I’m located in two different places that are nowhere close to here (two databases report two different locations). One says I’m in Glendale, CA and the other in Arcadia, CA. This is a pretty extreme example that puts me hundreds of miles from my actual location!
Most consumer ISP’s will be quite accurate within the market area, and only loosely accurate to the city level.
Take a market like Los Angeles where there are many suberbs in the larger metro area. The actual city limits of Los Angeles are pretty narrow. If you’re using an ISP that has a network POP even just a couple miles away, you’re likely to be reported as in a different City. Visits from the literal “Los Angeles” mean network points of presence that are literally in the city limits of LA, which is a relatively small area compared to the market area (see this map – the highlighted area is the city of Los Angeles – there’s lots of other populated ground around that which won’t be considered in the city, compared to this map, showing the LA “region”).
Accessing DMA Data in Google Analytcs
So, how do you go about accessing DMA in the “new” version of Google Analytics? It’s actually not that hard. Here’s my primer, in a quick Screencast video. Enjoy! Questions? Fire away on Twitter.