These are excellent questions.
What I would explain is that we apply something of an.... It's intelligence-driven, for sure, but it's a model that is iterative in nature. Intelligence will inform this, and seizures will inform this, and it is also informed by targeting. The national targeting centre is a big part of this, but so are the discretion, expertise and awareness of our border service officers at the front line, who are also teaching the targeting centre how to be more efficient.
Starting with that approach, we use advanced data analytics to identify anomalous patterns and trends, things that would surface as irregular within the commercial stream. When we are looking at two million marine containers per year, it is important that we are very efficient in applying targeting referrals for commercial officers to action, so that we are ensuring that we're achieving national security and public safety interests, but also ensuring the fluidity of trade at the same time.
As we set targets or containers for examination, we learn from those. We learn from our resultant rates—where we've opened a container, where we've done a tail-end examination, whether we've sent something to a container examination facility or used the LSI, a scanner. Whether or not we are successful is really important for us because when we're looking at thousands of potential examinations and thousands of potential resultant examinations, every one of those will teach more of them. If we've identified something based on certain indicators—for example, we've received advance commercial information—how we identify that will inform the next five seizures, and those five will, in turn, inform more.
I'll go back to the point that I mentioned about identifying—
