Does User Geographic Location Facilitate Floor Price Differentiation?

eCPM is a metric that can be crucial when comparing impression value, especially when there are no differences in floor prices (it is highly correlated with floor price, so it is not advised to use it while comparing differently priced parts of the inventory). And so, in order to check impression value variability between geos, let us first take a look at the differences in eCPM between different geos on the same, uniformly priced part of the inventory:

It is clear that in some cases the differences in eCPM can be significant – here the highest eCPM is more than eight times higher than the lowest one – this kind of difference illustrates that there is potential for geo-based floor price differentiation. It can also be observed that there are certain regions where ad impressions are much cheaper – like Central Asia, and ones where they are more expensive – for example the US and Canada. So if it doesn’t seem justified to give each geo individual pricing, it might still be a good idea to do this for groups of countries.

Differences can also be spotted while looking at bid distributions, shown in the figure below:

When it comes to buyers targeting traffic from Kazakhstan – low bids clearly dominate. There are hardly any bids above 0,60 EUR and the cumulative revenue curve flattens around 0,40 EUR; whereas more than 99% of all bids on this geo are lower than 0,40 EUR. Setting high floors here would be risky as this could affect the revenue considerably. Bidders targeting users from Russia and the United States, on the other hand, bid visibly higher – this means that impressions served to users from these regions are much more valuable to advertisers; and setting a higher floor in this instance could help protect the inventory against bid shading.

To sum up, both eCPM values and bid distributions show that there can definitely be a marked difference in the value of traffic from different regions and countries.

It is often the case that advertisers spend much more on users from a specific geographic location; and some might even bid on users from only one country. Moreover, this applies not only to companies running locally but also to global corporations. The figure below shows advertisers who spend the most across each of the three chosen geographic locations on the same inventory.

In this case, companies spend much more money on users from one geographic location of their choice than on others. There are companies that bid a lot on users from one country but spend nothing on others. This is actually quite a common occurrence – the figure below shows the number of advertisers who bid on impressions from a certain number of countries:

It turns out that more than a half of the 1508 unique advertisers who bid on this inventory were only interested in showing their advertisements to users from one country; and only about 10% of them bid on more than 4 geos.

These figures illustrate that the differentiation of floor prices based on geographic location often entails targeting different advertiser groups separately. This means that changing floor prices for impressions from one geographic location may only impact the bidding behaviour of a limited number of advertisers – which might turn out to be beneficial, especially if you want to price one geo lower than another. Changing floor prices should not make impressions from other locations subject to bid shading.