IP-Based Blocking

Let us begin the new year with one of the most common questions we encounter whenever clients start digging into our service: Can you simply block IP addresses?

With bots setting-up ranges of IP addresses for an attack, clients often wonder if we can simply block based on IP lists.

The answer would be no. We could not provide a list or range for our customers to block by themselves.

For example, let's take Amazon. Traffic that comes from Amazon AWS can have the best, if not the most, converted users. Very large companies use proxies on Amazon AWS in order to filter the that comes out of the organization. Blocking certain IP addresses from Amazon or ranges of IPs would risk degrading your advertising performance.

Let us give another example—say, university websites or websites of large organizations. We do not block them because if we do, we risk giving a lot of false-positives and if we whitelist them, a lot of false-negatives.

Our technology

Our technology works at a very granular level. On occasions of a virus installation or bot transaction, this is what we will do:

Say we have one mobile device that is behind an NAT IP address—which has 10,000 mobile and desktop devices associated to it—the moment the installation/transaction occurs, we will mark the impression as a bot. However, if we identify a real user on the same device on the same NAT IP, who sends one transaction, we will mark it as valid impression. 

We differentiate, between bots and humans at the transaction level.