Testing Methodology

We collect hundreds of messages every second. We take a sample and send the sample to Mailshell and competitive products simultaneously in random order. We measure accuracy, latency, and timeouts. For the accuracy errors, we retest every 5 minutes to all products that failed until the errors are fixed. Here is a sample breakdown of the messages: SpamIQ.

Testing Methodology

We collect hundreds of messages every second. We take a sample and send the sample to Mailshell and competitive products simultaneously in random order. We measure accuracy, latency, and timeouts. For the accuracy errors, we retest every 5 minutes to all products that failed until the errors are fixed. Here is a sample breakdown of the messages: SpamIQ.

* VBSpam benchmarks are based on VBSpam November 2011 test results assuming that what Mailshell errored on was correctly classified by others and vice versa.

Testing Methodology

We collect hundreds of messages every second. We take a sample and send the sample to Mailshell and competitive products simultaneously in random order. We measure accuracy, latency, and timeouts. For the accuracy errors, we retest every 5 minutes to all products that failed until the errors are fixed. Here is a sample breakdown of the messages: SpamIQ.

* VBSpam benchmarks are based on VBSpam November 2011 test results assuming that what Mailshell errored on was correctly classified by others and vice versa.

Testing Methodology

We collect hundreds of messages every second. We take a sample and send the sample to Mailshell and competitive products simultaneously in random order. We measure accuracy, latency, and timeouts. For the accuracy errors, we retest every 5 minutes to all products that failed until the errors are fixed. Here is a sample breakdown of the messages: SpamIQ.

* VBSpam benchmarks are based on VBSpam November 2011 test results assuming that what Mailshell errored on was correctly classified by others and vice versa.

Testing Methodology

We collect hundreds of messages every second. We take a sample and send the sample to Mailshell and competitive products simultaneously in random order. We measure accuracy, latency, and timeouts. For the accuracy errors, we retest every 5 minutes to all products that failed until the errors are fixed. Here is a sample breakdown of the messages: SpamIQ.

Testing Methodology

We collect hundreds of messages every second. We take a sample and send the sample to Mailshell and competitive products simultaneously in random order. We measure accuracy, latency, and timeouts. For the accuracy errors, we retest every 5 minutes to all products that failed until the errors are fixed. Here is a sample breakdown of the messages: SpamIQ.

Testing Methodology

We collect hundreds of messages every second. We take a sample and send the sample to Mailshell and competitive products simultaneously in random order. We measure accuracy, latency, and timeouts. For the accuracy errors, we retest every 5 minutes to all products that failed until the errors are fixed. Here is a sample breakdown of the messages: SpamIQ.

Testing Methodology

We collect hundreds of messages every second. We take a sample and send the sample to Mailshell and competitive products simultaneously in random order. We measure accuracy, latency, and timeouts. For the accuracy errors, we retest every 5 minutes to all products that failed until the errors are fixed. Here is a sample breakdown of the messages: SpamIQ.

Testing Methodology

* VBSpam benchmarks are based on VBSpam November 2011 test results assuming that what Mailshell errored on was correctly classified by others and vice versa.

** We collect hundreds of messages every second. We take a sample and send the sample to Mailshell and competitive products simultaneously in random order. We measure accuracy, latency, and timeouts. For the accuracy errors, we retest every 5 minutes to all products that failed until the errors are fixed. Here is a sample breakdown of the messages: SpamIQ.