I wonder if that might fit the average pattern of a casual stats user a bit better (only actually checking the data when asked by the manager) to keep them hooked for a much longer 'wall clock' time and get them to eventually convert (I've been depending on this for half a year now)
In practice, because it's self-hosted, "cracking" might be an issue. Customers might edit the files that affect the retention, for example. Maybe most customers won't do it, but I don't know. This also feels a bit like I would have to implement some "DRM", which I really don't want.
Now that you mentioned it, maybe a better trial would be a freemium model, where I can serve a different version for free that only has some features. The problem with this, is that the customer won't get all the benefits of using the product, so they might not like it enough to upgrade to the full version.
It's an analytics platform, so I could offer just basic stats for free and for premium all the other features (segments, heatmaps, recordings, A/B tests, AI integration, etc.). This could work as a good marketing technique for the top of the funnel, but then customers would still probably want to trial the pro features, so I am stuck with the same problem as before.
After reading more of those comments, I am still unsure what's the way to go. Maybe simply keep the trial but make it 30 days instead of 14 days.
The problem with premium features, is that I would then need a free version of the product. But I don't want to provide a free version without support, as it's a self-hosted product.
A lot of companies are firmly on the evil track, looking for any dark pattern that makes the numbers go up, sadly. Good on you for not being there.
Also, the app is self-hosted, so part of the trial benefit is that users can test the installation process, which is usually the biggest push-back against self-hosting.