How to Hack an Election: Internet-Style
The best argument against democracy is a five-minute conversation with the average voter— Winston Churchill
We like to think that the Internet is a free and fair democracy. That upvotes on Reddit, HackerNews, ProductHunt, and so on have some sort of universal meaning — that the posts on the front page represent the best posts of the day.
That’s about as true as thinking that politicians holding office represent the best candidates of the time.
The fact of the matter is that Internet points represent attention. And attention is increasingly hard to come by these days. It’s valuable. Companies will spend a lot of money to get your attention. So it’s not surprising that upvotes, like political votes, are largely a function of money. Upvotes and downvotes are for sale to the highest bidder on several different marketplaces. A small number of upvotes can get you to the front-page. And once you get enough votes to get to the front-page, the masses take over. It’s cracking the front-page that’s hard. And it’s only hard because it’s for sale and you’re not buying.
Maybe you should buy it. Is it even unfair if you do? Other companies are. And you need to keep up with the Joneses. The only difference between you and the Joneses is: the Joneses are better connected /s
Luckily, the Internet isn’t (yet) as corrupt as political campaigns. If you post something great — especially on HackerNews or ProductHunt — you’ve got a decent shot at success.
A few days ago, I posted a company I’ve been working on — GigaDiff — to HackerNews. I didn’t buy any upvotes. I didn’t tell any friends to upvote it. I suspected because I didn’t, the post would meet radio silence: no upvotes. But I was fine with that. I wanted to know if I was really onto something or not. Turns out I might be. GigaDiff made it to the front page for almost the entire day. 700 people visited the website. 14 of them signed up. That’s pretty damn good for Enterprise Software.
But I wondered to myself — not what would’ve happened if I cheated. No, I wondered to myself, what would’ve happened if I posted on a different day or time.
I posted GigaDiff on January 1st. Maybe it was a bad day. But my post got 33 upvotes, which was enough to keep it on the frontpage from the early morning til late at night. Still, I wondered what might’ve happened if I posted on the 2nd, or the 3rd, or even, daresay I, at the perfect time.
That got my nerdy brain worked up. I wondered: Statistically, is there even such a thing as the perfect time? To answer that, I gathered some data — all the “Show HN” posts from December — and made some charts to get an idea. It turns out that posting at certain times and days definitely (and not surprisingly) can help your cause.
If you’re curious, here’s a chart of the posts per hour, how many of those posts “successful” (got 30+ votes), and what percentage of the posts made at that time were “successful”.
As you can see, posting from about 6:00 p.m. to 5:00 a.m. isn’t very smart. Posting from 10:00 a.m. to 1:00 isn’t very smart either. The best times seem to be 8:00 and 9:00 a.m. PST and 1:00 p.m to 5:00 p.m. PST.
That’s what I figured. It’s intuitive. People are bored at work with nothing to do but read HackerNews around those times. Notice that people aren’t reading HackerNews during their lunch break — when you might have some time to post your company.
Similarly, they aren’t upvoting on Saturday either. Here’s a chart of which days the masses are upvoting on:
Statistically, Wednesday and Friday are the only days where your chance at success is a little higher. Sunday only seems to have a decent amount of successful posts because of the total volume of posts being the highest on that day. And as my late Great Grandmother once said, “Thursday is a day best avoided”.
Finally, here’s a couple of scatter plots so you can get a better idea of the posts:
I made charts for the dates as well. But obviously with only one month of data, there’s nothing interesting there. Except that the week of Christmas — when a lot of people are enjoying PTO — is pretty dead. My hypothesis is that even with several months the date will be insignificant. But I’ll test it anyway. I might find something interesting.
Until then, good luck with your businesses and with your postings. For your convenience, I created a site: ActiveTrack.us which will show you the voting frequency of HackerNews and different subreddits in real time. To quote Effie Trinket: “May the odds be ever in your favour”.
