The individual sample sizes are small, yes. The variance between the paid vs organic conversions was still stark to me, e.g. 54 paid visitors had a 0% conversion rate to trial vs ~20% with 5 of 26 organic users converting to trial.
FWIW, the reddit ad campaign is at nearly a hundred views now, and still no trials.
My experience in ads is that most digital display ads for technical people are a waste of money. You will have zero conversions, period. Our demographic does not click and convert. My data include million-impression campaigns for other salesforce products on channels like Reddit, AOL, Yahoo, and Google...
For perspective, 100 views is extremely small in advertising. Hobby-small. You'll need a sample size of a few thousand views before you'll get even 100 interactions to compare. The sample sizes of conversion for this experiment are too small to make any conclusions: https://news.ycombinator.com/item?id=4685928
I suspect you will see better conversions from this content ad (i.e. blog post) posted to HN. Which, coincidentally, I fear blog ads are a large percentage of the content posted to HN today.... still nice to see others' findings, so thank you for posting.
Anyway, good luck with the product, but I can't emphasize enough that you should be skeptical with any of the results you have. You simply have too few data points to deduce any trends.
FWIW, the reddit ad campaign is at nearly a hundred views now, and still no trials.