You'll have a low H index if you have lots of low quality submissions, and you'll have a low H index if you have very few ultra successful submissions.
From wikipedia: The definition of the index is that a scholar with an index of h has published h papers each of which has been cited in other papers at least h times.[4] Thus, the h-index reflects both the number of publications and the number of citations per publication. The index is designed to improve upon simpler measures such as the total number of citations or publications.
That's why a lot of people like to look at more than one number when looking at a set summary, but you can't compare and classify these as easily.
An example: https://github.com/antontarasenko/smq/blob/master/reports/ha...
Its discussion: https://news.ycombinator.com/item?id=11499120
An explanation: https://news.ycombinator.com/item?id=11499158
1. Deduct a small value per post or as a function of total posts. Theory I apply is that all content has an assessment cost. More bang in fewer posts is better.
2. Evaluate posts seperately based on some value metric. E.g. thorny technical problems vs. "How I lifehacked my gym socks" pablum.
One way or another, the system must reflect value judgements.
I would suggest for certain kinds of submissions/topics to add logarithmic instead of linear influence on the h-index. Or to add some form of handicap.
After all SF housing is something that the community is obviously interested and passionate about - you can easily get 200 comments/upvotes on any article on the topic. So you check if this article contributes to the h-rank by removing the baseline upvotes/comments an article about SF housing gets.
TL:DR - for some topics your contributions should count towards the h-rank only if they engage the community above average.
On the other hand perhaps a simple comparison of keyword frequencies in title or all comment bodies would work? Still, doing this properly is nontrivial.