A few weeks ago, Google introduced the ability to target in search via household income. As a quick refresher, Google now contains the ability to layer household income targeting that is based off census data. The tiers are based off average household income data (publicly available data from the IRS). The top 50% are broken out into 10% buckets, with the bottom 50% all in one. Adding the targets into your account with 0% modifiers allows you to collect data on how groups behave without spending any extra money.  We decided to add all the targets with a 0% modifier into one of our e-commerce accounts (see how to here) to see if there was a difference in performance between the different incomes, and here’s what we found! HIT

1. ROAS and AOV performed pretty much exactly the way you’d expect: We saw that the ROAS and AOV increased as average HHI increased, with a big difference between the top tier and the next-highest. This makes sense as people who have a higher income will be more likely to buy more expensive things. Though our top tiers have less volume than our lower tiers, they are much more profitable and are worth bidding up.

2. CPCs increased as average HHI increased: Looking at the increase in average HHI, we can also see a corresponding increase in CPC. In this account in particular, we have enhanced CPC enabled (a Google feature that will increase your bid by up to 30% if it thinks you’ll have a better chance at converting), which indicates that Google’s bidding algorithm will bid these targets more aggressively. This means that if you have enhanced CPC or conversion optimizer enabled, chances are the algorithm is already increasing your bid based on this data.

Using the data, you can then add in an additional bid modifier to bid up profitable targets and bid down inefficient targets. In summary, everyone should at least add the targeting with a 0% modifier to collect data (especially in e-commerce accounts!) and add some bid modifiers if the opportunity presents itself. Good luck!


  1. Steve Cameron May 6th, 2014

    Interesting data – and, as you mention, not completely unexpected – but there are likely to be verticals where the trend is reversed. An obvious one, for example, might be payday loans – I guess as with most things in ppc there is no one size fits all solution.

  2. Susan Waldes May 6th, 2014

    Great findings! I have a theory also that “the rest” of the geo-target that doesn’t track to one of the HHI tiers is mostly search partner traffic. So, say you are targeting the US.If you have the 5 tiers for the US and then also the regular US target…are you seeing that what falls in into “the rest” tracks pretty closely to search partner activity? It’s not exact but within about 15% on most campaigns I look at. My theory is that you can create a “sort of, kind of” search partner bid modifier by using the HHI targets to represent Google Search and “the rest” to represent search partners. Does your data indicate this also?

  3. Anonymous May 8th, 2014

    Hey Susan! Unfortunately, we can’t break out HHI data by network, so we can’t get an exact answer. I pulled an account look at the data on the rest of the US and our Search Partner data. At a macro level, the numbers between Search Partners and the rest of the US target aren’t super close. How close are the numbers you’re seeing in your campaigns?


  4. Ashley July 2nd, 2014

    Interesting find! I’ve been applying this to one of my campaigns and want to view my performance similar to the table you posted. However, how do you pull revenue into Adwords in order to compare which HHI tier obtained a certain revenue?

Leave a Comment

Bailey Lau
Bailey joined 3Q Digital in September of 2012 after interning at Internet Marketing Inc. in San Diego. Bailey graduated from UC San Diego in 2012 with a bachelor's degree in Sociology. She loves finding new places to eat, watching sports, and playing softball and football. She currently works for Macys.com.