Can someone explain the Product Hunt algorithm?
Sandra Djajic
10 replies
So yesterday, I launched Klu.so on Product Hunt. https://www.producthunt.com/posts/klu-3
We received the most upvotes, comments, and reviews from the community, from users who have been on Product Hunt for a while, and I'm trying to understand how we could have landed in third place.
I'm incredibly happy and grateful for all the visibility we received, but I want to better understand how a company with 700 upvotes and 200 comments can finish first, while a company with 800 upvotes and 400 comments doesn't.
Replies
Teo Deleanu@teodeleanu
Sandra, I followed your path and the thing is they implemented anti-organic algorithm.
Here is why you might get penalized:
1. New voters registered on PH (come especially for your page, like your digital grandmother π΅)
2. Voters with less than 5 followers (considered unpopular? π)
3. Voters with no recently or no voted products (PH ghosts π»)
4. Direct links to your PH page from social media - not proven to bring minus points
5. Automatic voting π€
6. Sudden spikes in votes at a given time (possible mass email campaign)
7. Voting without visiting the product website (said to check if the user pressed "Get it" before voting)
8. Voters from many different sources (shows you post heavily on multiple communities)
9. Requesting votes - big βοΈ - PH wants the product to be optimized through community support, not just to reach #1.
Also this is why some might get better traction(X multiplied):
1. Top Hunters upvotes and hunt (big reco to contact them 2 weeks before)
2. PH staff who helps you elevate your story (we heard they are Super β Duper nice, reach out to them)
3. Comments and active discussions on your product page (add everyone who worked on the app as makers and ask them to answer too)
4. Early votes during the day (early for PH means 00:01am PST β having early votes looks more natual as they think you canβt rise people from their bed with marketing so itβs true upvotes π₯± )
So probably the other 2 startups have more insights on how the algorithm works.
Some friends were talking with some CEO's at some point having a story about the developers that were getting "bribed" to help you uprank your app into apple store with algorithm insights. Any insights is helpful and might also help to get awau from penalties and go more into multipliers.
Share
Congrats on the launch!
We had a similar experience recently and one of the things I read about was that the new algorithms give weight to votes per hour rather than just total votes. So those launches with a steady and consistent slope of votes do better than those that have big spikes in traffic over the day. But I'm sure there's more to it than that.
Hunted Space
@brenna_donoghue Yes, this is my theory as well. I noticed that we had a steady growth from the beginning, but the other company had a different growth pattern. In the last few hours, they had more upvotes per hour. I think this is where the issue is because the algorithm predicts that the growth will continue like that. However, now we can actually check and see that this is not happening. So, we still have the most upvotes and comments since our growth is still stable. It's very interesting to look at.
StartupSoft's Teams Digest
besides what previous commenters also said, I'm convinced that they take into account geographical distribution too. We've had a few customers' team upvoting us and I think since they all were in proximity/shared IPs (if upvoted from offices), and they've subtracted a few hundreds votes from us, support later said that they counted them as "spam" activity.
I know that previously if the same thing occured, they didn't subtract those votes, just didn't count them into the ranking.
Also, we had different numbers showing up on the launch page vs notifications and makers pages.
Support told us that that was a glitch and while they subtracted those votes we shouldn't be seeing the real numbers.
To be 100% honest I think that algorithm is just a bit glitchy but I don't think there's any foul play involved from the Product Hunt itself.
In the end of the day, I've had about a dozen launches so far both for my businesses and clients', and my opinion that as long as you tank in top-5/top-7 you are going to get pretty much the same exposure/traffic/leads. And at the same time, the Product of The Day badge is not as impressive as it was a few years ago. So, don't worry! You did a great job and ranking in top-3 is amazing!
I also came to check on Klu ranking this morning, and the first thing I've seen there is no Klu at the top. They are changing the algorithm to help less visible products come up... The system is rigged! Another thing is that someone already mentioned here, the top products of the day are hunted by YC execs (big potatoes in the game) so... You know... Unlucky you launched on the same day. But still huge success for Klu team, great product, keep in mind you can always launch again and that's one of the good strategies here to launch multiple times.
Supatale
Went on Product Hunt today to check out your success and was amazed that you came in third. After researching those who beat you, here are my hypotheses:
#1 was run by the cool hunter @mwseibel. From what I understand, he does all the launches from YC alumni. And from what I understand, the votes of makers who have ever posted anything also carry more weight than just community members
#2 was also run by top hunter @kevin. I'm assuming it's also the makers who are subscribed to it.
In any case, I congratulate you! And I'm off to get ready for my launch.
CafeSocial.io Mixer
@mwseibel @kevin @kate_bovkunova
Same here and was a bit surprised to see you not ending up on #1.
I think the conclusion from Kate makes a lot of sense. Both Michael and Kevin have 30K/15K followers and the age of those followers is much older. You joined 66 (?) days ago and what you have done in that time is impressive. Saw some much from you here and on twitter.
Top 3 is amazing.
Alconost Localization
I think it also may depend on how old are the accounts commenting and voting. I.e. I believe that their 'maturity' may count, and at the same time, accounts created just 1-2 days ago with no past activity may not help you