One of the most powerful aspects of the Enhanced Campaign model is the ease with which advertisers can now adjust bids by geography at scale. Using location IDs passed through Google's click performance report, advertisers are able to tie sales and orders directly to geographic locations in order to calculate granular bid modifications for greater optimization. It's important, however, to take a look at how the average cost per click already varies for an advertiser depending on geography, and the reasons that these differences by geography exist.
Higher Income Locations Have a Higher Average Cost Per ClickIf you've tested and analyzed Google's zip code based average household income (HHI) demographic modifiers, you likely found that higher income areas converted more often than poorer areas, as RKG has found for most advertisers. Another trend that quickly reveals itself, however, is that cost per click rises with income level despite 0% bid adjustments, such that traffic from those in the wealthiest areas is the most expensive. Here's sample account-wide data from one advertiser: So what's the deal? Is this a product of differing search behaviors between geographies with different incomes? Are the dynamics of the auctions in wealthier areas different from those in poorer areas, such that it costs more to get on the page? Or is Google putting their thumb on the scale in higher income areas to get more out of advertisers for this better converting traffic? Let's take a look.
CPC Trend Exists Even for Single, Exact Match Keyword CampaignsWhen looking at the variance in CPC by HHI for a campaign with many different keywords, it seemed possible that keywords with lower bids could be more heavily dominated by areas with lower incomes and vice versa. To test, we added HHI demographic modifiers with 0% bid adjustments to a campaign that contained just a single high traffic, exact match keyword. This campaign exhibited the same trend in CPC: Thus, while search habits certainly vary between areas with different average economic station, this is not what's causing the increase in CPC in higher HHI areas.
Higher CTR in Wealthier AreasTaking a look at the same single exact match keyword test, we also find CTR is highest in the highest income areas and lowest in the lowest income areas. As quality score is determined on an auction by auction basis, this variance in CTR could theoretically lead to fluctuations in the amount of competition involved in a single auction. However, we haven't seen any evidence of significant fluctuation in average position by HHI, indicating that this isn't the result of an influx of competition for auctions in higher income areas.
Google Dialing Up CPCsGoogle does provide Enhanced CPCs (ECPC) which allow advertisers to give Google the control to bid up to 30% more than the keyword max CPC for auctions in which Google believes the traffic is more valuable to the advertiser. This feature, however, was not enabled for the advertisers whose data has been presented in this post. The aforementioned lack of significant fluctuation in average position by average income seems to indicate that the higher CPC we've been seeing in higher income areas is not the result of other advertisers activating ECPCs or utilizing other geographic bidding adjustments. Thus, it seems possible that Google is creating this increase in CPC in higher HHI areas. One possibility for how they would go about this is to increase the CPC required to be listed in a top placement, which as RKG's Mark Ballard has written about in the past, increases the CPC paid as if Google itself has entered the auction.
Impact on Geo Modifier BiddingIt's important to understand that the cost per click paid is already varying by the average income of the area, as the difference in value of those clicks may already be accounted for in the variance in CPC seen across HHIs. It's also possible that there are other performance impacting geographic attributes for an advertiser that already exhibit differences in CPC without modifiers in place. As we move into adjusting bids based on more and more attributes, both geographic and not, it will only become more important to understand how Google's auction is already impacting the cost per click paid for different audiences. Big thanks to Braxton Puryear and Seth McClellan for their assistance.
Join the Discussion