Email is one of the only industries where vendors “terminate” customers that pay their bills. It’s the unfortunate reality of a tool – email – so effective that many try to abuse it.
Things are no different for us at MailUp. Every day companies that don’t play by the rules try to take advantage of our system to send unsolicited commercial email (the fancy name for SPAM).
In the 2013, as we started growing internationally at a more rapid pace, we had a particularly busy time fighting spam. We canceled numerous customers, and started working hard to improve the filters that we had in place to prevent bad actors from doing any damage.
We had already set up several alerts and automated processes to detect senders that were exceeding certain thresholds, such as higher-than-allowed bounce and complaint rates. We also had introduced several “pre-flight” checks at the time lists were uploaded and messages readied to be sent. Still, it wasn’t enough.
We also noticed that we were spending a rather insane amount of time doing hands-on customer vetting, made more difficult by a growing number of users coming from all over the world (when you don’t speak the same language, things get even more interesting as you try to evaluate a potential customer).
An internal assessment showed that – on one side – too much time and too many human resources were being taken up by the customer vetting portion of the client onboarding process. And yet – on the other side – the rate of false negatives was still too high for our standards.
Something had to be done!
So we took a long breath, went back to the drawing board, and started thinking about how to redesign our customer vetting system. We had three objectives in mind:
- A less time-consuming process for the good guys
- A more efficient and effective way to catch the bad guys
- A more automated system that would allow us to assess the riskiness of the lead quickly, with minimal or no human intervention, before the very first email would exit our system
A more automated customer vetting process
We started tapping on expert advice from some of the gurus in the industry, expanded our Abuse & Deliverability Team (and we’re still hiring), and looked for new tools that would help us get the job done. During this process, we ran into a couple of companies that had exactly what we were looking for and provided us a way to fill some of the gaps.
E-Hawk is one of those good finds. It helps us more rapidly and automatically assess the risk associated with a sender. It makes it easy to obtain information that would have required us a lot of engineering to collect on our own.
In addition, and even more importantly, E-Hawk introduced a community aspect to the vetting process: it allows us to share with the anti-spam community feeds and information, and benefit from the knowledge submitted by other contributors. At MailUp, we agree with E-Hawk that this approach will help the email industry get the vetting process to the next level: bad actors change tactics very often, and the help of the entire community is needed to keep up.
A good vetting mechanism at the start of the onboarding process is key, but it’s just one piece of the puzzle. Another area that we needed to improve was the ability to detect issues with a customer’s recipient list.
Better list quality assessment tools
We had been using internal and external resources to identify problematic senders for quite some time. However, as our international growth was accelerating, we needed more and more reliable sources to feed our quality assessment algorithms.
One problem was that we could not afford to merge our data – which we trust as it has been collected directly by us – with other data feeds without undergoing a thorough analysis to determine their quality.
This fairly complex process has some constraints, and lead us to rule out data providers that only expose APIs, as we needed to bulk load records into our data warehouse.
After some research, we selected Indiemark’s Blackbox service as a solution. Blackbox provides large volumes of data, good quality, and frequent updates. Another piece of the puzzle fitting into place!
Being a rather pragmatic person, I perfectly know that we are nowhere near being done: there is still a lot of work ahead of us. And yet, it feels good to know that we are on the right path, and that we’ve made a lot of progress toward the objective of minimizing any misuse of the MailUp.
Thanks for reading. If you have any thoughts on the above, please let me know in the comments. I more than happy to share my thoughts and resources with the whole anti-spam community!
Head of Deliverability and Compliance, MailUp