Show HN: Privacy-focussed end-to-end encrypted forms | Hacker News Reader

Show HN: Privacy-focussed end-to-end encrypted forms

palform.app

2 pointspalk3 comments

Palform addresses my personal concerns with commercial form builders; it places a core focus on privacy by end-to-end encrypting all responses. It doesn't compromise on functionality, containing basically all the features you'd expect from competing tools. It's priced very reasonably with a fairly generous free tier. I'd love some feedback on what you think of the idea/execution!

Open articleView on HN

What kind of analytics can you do with the data? I see a dashboard mentioned, but I wonder if there are more views.

Hi! At the moment, we show a range of simple digestible graphs, with a relevant graphic customised to each question type so it makes the most sense (e.g. a heat map for the address input). We also do correlation analysis, showing which pairs of questions have a strong relationship in their responses; again, in a plain-English simple view. You can even see the regression graphs for the most correlated question pairs.

Analytics is a struggle due to the E2EE, but we've done some benchmarks (which we're hoping to formalise and publish soon) and it's surprisingly lag-free even with tens of thousands of responses! So we've got the basic groundwork, and we're hoping to add more analysis in the future (as long as it doesn't require a stats degree to understand!)

Hope that helps, and thanks for your interest!

Hi! Analytics has been a struggle to implement due to the E2EE meaning we have to download _all_ responses client-side. However, we're very happy with where the implementation is right now.

To start with, you get pretty graphs to summarise each question, with the exact graphic depending on the question type (e.g. a heatmap for the "address" input). This is a quick, simple summary and anyone can understand these. They're also easy to download and share.

Then there's correlation analysis, showing which pairs of questions have the most strongly related answers. We summarise this using plain English to ensure people without stats degrees can understand it! You can also view the regression graph for each question pair, which can help determine other patterns.

Despite the complexity, we've done some initial benchmarking (which we're hoping to formalise and publish soon) that shows these analyses perform very well at scale, hardly exhibiting lag even with tens of thousands of responses. Overall, it's a great groundwork that'll be nice to build upon in the future as we add more sophisticated analysis without compromising on the simplicity aspect.

Hope that helps, and thanks for your question :)

Reply on news.ycombinator.com