GA Demography Report Accuracy
I had a campaign targeting Female audience (ONLY), and the performance was good. However, GA Demographics report shows that more than 60% of the users who reached the landing page is Male audience. All the publishers have confirmed the targeting and they did target the right audience (Female).
So my question is, how accurate is the demographics data? I have been reading numerous articles and seems like this report is 80-85% accurate. In that case, could you help me understand the possible reasons behind this pattern in GA report?
Thanks so much!
Re: GA Demography Report Accuracy
1) Age is broken down into the following buckets (18-24, 25-34, 35-44, 45-54, 55-64, and 65+)
2) Gender (male or female)
3) Affinty Categories (basicly interests)
4) In-Market Segments (what the users online product purchases are or interests)
5) Other categories (more detailed than affinity categories)
This data comes from third party DoubeClick cookies for web traffic channels and from anonymous identifiers for those users who use mobile devices. GA looks at these cookies and/or identifiers it makes educated guesses as to the most likely type of person (for instance if it can determine that you looked up school books, and shower caps it might determine that you are a female between 18 - 24) (if it can see from the tracking cookies that you did a search on italian food it might determine that you have an affinty interest in cooking and and other category of Italian dinning) (if you went to ebay and searched on coins and bought an american eagle it might classify you as a 30-55 year old male who is in to numitist and an other category of US coins) this is basicly all speculation on GA's part, keep in mind that Google collects a lot of data on users so over time it can make a rather good eduacted guess. But here is where it gets complicated GA does not want you to be able to actually identify a user that would be a violation of Persanly Identifable Information or PII so they subject that data to Thresholds and Data Sampling. Thresholds are identified when the data can possibly lead to the users identity and you know when it happens because a yellow notice will show below the report title. Data Sampling is also limited to I believe 10% although I have seen different figures but in any case the data is sampled and a very small percentage of the data is even counted, again this is to protect a users PII. So when you take all of this into account it can give you an educated guess of the demographic that makes up your user base but it is not to be relied upon for any real hard decisions it is only an estimate that is based on a sample of your vistors data and some individuals vistors can configure their browser so that no data is even passed to GA.