Analysing 100,000 product launches
Matt Harbord
16 replies
I've pulled data on 100,000 product launches from the PH API and I'm now starting to analyse it.
I'm hoping to answer questions like what difference does having a video make; how easy/hard is it to get featured on different days; how many votes are usually required to get top 5; stuff like that.
So, what do you want to know?
If you have any burning questions I can answer from the data I'll do the digging and let you know!
Replies
Veit Josef Schneider@veit_josef_schneider
1. How often and with how long time spans between the launches have successful products launched on PH?
Share
Uh got some more questions :)
- Can you explain the criteria for being featured on different days?
- What are some common features among products that have been successful in the past?
- Have you noticed any trends in terms of product categories that tend to perform better than others?
- How does the number of comments or user engagement affect a product's success?
- How does the launch time and time zone affect the product's performance?
- What percentage of products that are launched actually get featured or gain a significant amount of votes?
Looking forward to some insights :)
@veit_josef_schneider I'll add those to my list!! There are certain limitations with the data PH share so there's a limit on what we can answer confidently; for example they only share "featured" products, you can't find out how many launched but didn't get featured (which is really annoying when trying to compare those two sets!)
@veit_josef_schneider To answer your Product Categories question - this is actually good data to look at, as PH doesn't seem to "do" much with the topics you've selected, so the voting isn't contaminated by them promoting certain topics.
There's some graphs in the link here:
https://twitter.com/MattHarbord/...
But basically - the community has a strong preference for products that help them personally build better products - so the top topics are "no-code", "growth-hacking" and "design-tools"; they don't care for consumer apps. "games" (and apps in general) have a net-negative contribution. i.e. the average score for anything that *isn't* a game is higher than the average score for things that are a game.
@matt_harbord thank you! so reading this I could assume that instead of "naming" my product an "app" but rather a "tool" could increase visibility/clickthrough?
@veit_josef_schneider It's hard to say, but I'd definitely make sure the positioning is all about how it solves the users problem - and giving them a "tool" is a good psychological nudge towards that goal.
Basically the only conclusion we can come to is that the stuff that helps builders build is most popular - which is only really useful to know if you're a) wanting to build a product specifically to do well on Product Hunt, or b) want to give yourself a better idea of what a successful launch looks like for you once you've considered *where* your product sits versus the audience and their preferences.
I've not done any analysis yet on the "name" or "description" of the product - but will let you know if I do come up with anything.
@mycelebs_elle 
So we can't see how votes are distributed throughout the day - only the total; so it's not possible to look for things like gaining momentum after a certain time.
But I looked at launch hour - and not being up early hits your chances of being number 1 pretty hard (though this may just be a case of the best products having their act together enough to schedule it properly and not rush it out late).
So we can't see how votes are distributed throughout the day - only the total; so it's not possible to look for things like gaining momentum after a certain time.
But I looked at launch hour - and not being up early hits your chances of being number 1 pretty hard (though this may just be a case of the best products having their act together enough to schedule it properly and not rush it out late).
Is the data available on the quality of upvotes? Since the number of upvotes vary a lot more between different days, what would you say would be a better metric to judge a successful launch?
@dipak_sisodiya no, only the raw number - but interestingly the numbers don't vary much on different days for the middle of the pack - they do for number 1 spot (which is what we all focus on!)
@matt_harbord What is the highest number of upvotes anyone has every gotten? What day was it?
@dipak_sisodiya https://www.producthunt.com/prod... - one thing we can't separate from the API is when votes were made; so I imagine a lot of the top performers have promoted their launch after the fact and continued to collect upvotes.