Corey from Custora here (http://www.custora.com) - if you're interested, we provide online retailers with an interface to manage lifecycle/drip messaging. We integrate with Mailgun, as well - and are very psyched for the new API features!
Hey, Corey from Custora here. We're huge fans of cohort analysis in general - it provides all kinds of insights into how customer behavior is changing. It's a powerful way of viewing historical data.
However, we need to be careful when using cohort analysis as a technique to predict lifetime value. There are many situations where cohort-based CLV predictions miss the mark by quite a bit.
We're working with a couple SaaS companies right now to get a feel for how to best apply our platform in that setting. If you're interested in chatting about your SaaS product, email us at info at custora - happy to chat!
CLV is a big part of our platform - we run a lot of analysis on the individual customer level. We use probability modeling to calculate accurate CLV numbers, and we do a lot individual-level analysis to detect when high-value customers start fading away.
In the example you cite, I agree - if someone just aggregate links generally, and had no intention of pushing people towards verified child porn, there was no intention there. Those link shouldn't be penalized.
On the other hand, if someone did verify the illegal content, and posts links with the clear, sole intention of driving others towards the illegal activity - then what?
I don't think you can decouple the link entirely from intention - links are sometimes expressive, as this article points out.
Instead of just giving ultimate, endless rights to link to anything for any reason, we say: base case, free speech, link to anything. But if someone can prove your intentions were illegal (no easy task), you're not necessarily protected.
Intention is a tricky gray area, but this would be in sync with other free speech protections - e.g. someone can press charges for libel on things you publish, but the burden of proof is on them to show the statement was false, harmful, and was made with malicious intent.
This is actually making a very strong argument to protect hyperlinks and the right to use them - but to do so under the lens that hyperlinks are free speech.
Much like newspapers have the right to say whatever they want. Of course, there are limitations. People can sue the NY Times, but it's not easy - the burden of proof is on the people bringing the suit. The 'strict scrutiny' standard, as I understand it, would make it very difficult for someone to attack hyperlinking. It does, however, enable that discussion in the case of extreme situations.
I'm curious as to why extreme examples are "weak" in the context of evaluating a framework to evaluate hyperlinks. Whatever policy the courts put in place, they have to handle both extreme and non-extreme situations.
Following your line of reasoning, we'd accept that building a website with the sole intention of linking to kiddie porn would be OK - because the author of those links has no control over what's on the other side.
As with most discussions like these, the trick is in how to draw the line, how to strike the balance.
Thanks for the feedback - it's helpful to hear reactions about what's most important to add next the site.
We're learning that there are large aspects of the customer retention challenge that are consistent across many types of businesses (e.g. for a retailer, how can I tell if my customer is gone for good or just idle?). Of course, some aspects are unique for each vertical. In those cases, we're exploring how we can supplement existing efforts - some of our beta clients have piped results from our analysis into other tools.
Thanks for the kind words - and for the suggestion for more info on the team. The co-founders are computer science guys, and we also studied probability modeling over at Wharton. Excited to bring these tools to market!