Dr. Marco Maier

Stats of being on the Hacker News front page

published: 2024-01-03

In 2023, I launched one of my side projects on Hacker News (HN). I have been on HN almost since its beginning, but have always been mostly following the discussions passively. For this side project, I chose to try an actual launch on HN.

The project I'm talking about is hacker-jobs.com. It started as a personal project to explore the capabilities of Large Language Models such as OpenAI's GPT. The site uses GPT to parse the job postings of the monthly "Who is hiring?" threads on Hacker News, in order to create a structured representation of the job postings. I have also written an article about the project and how to extract information from unstructured texts with the help of GPT.

The functionality of the project turned out to be quite useful, so I decided to keep it running and to try to promote it on HN. And luckily, I succeeded in hitting the front page of HN, which brought many visitors to the site.

My submission sitting at place 17 on the Hacker News front page.

In this article, I present the traffic stats of that day. The stats are especially interesting because at the time of hitting the front page, there was no other source of traffic for the site (the site was completely new). So all the traffic can be attributed to the HN post. I guess it is a rare example of launching a new project on HN isolated from any other sort of traffic or marketing activities.

How does Hacker News work?

For those not aware of the mechanics of HN, here's a brief summary. HN's main outlet is the front page. Members of the community can submit stories (i.e., links to other websites), and other users can discuss and/or upvote the submission. When first posted, a story starts in the "new" section of HN. In case the story gets enough activity in a certain amount of time - the actual algorithm is not disclosed publicly - it can make the jump onto the front page. Again, depending on the activity, the story will then rise or fall through the ranks. After some time, the story will be pushed back to the second, third, and so on pages.

The convention when posting one's own project on HN is to prepend the title with the "Show HN:" marker. Posting a "Show HN" story on HN has the effect that it additionally gets into the "show" section of HN. Or rather, it starts in the "shownew" section and can then advance to the "show" section from where it has the chance to get on the front page.

In general, I think that most submitted stories do not get anywhere near the front page. But getting on the front page is kind of like a little jackpot because it will drive traffic to the submitted link. If your submission qualifies for being a "Show HN" story, this seems to increase the chances of reaching the front page because I guess that the HN audience has a slight preference for things built by other members. (But this is just a personal assumption without any data to back it up).

How do I get on the Hacker News front page?

HN is one of the rare places on the internet that was able to keep the quality of its community and contents relatively high. The best way to get noticed on HN and reach the front page is to contribute content that is genuinely interesting for the HN audience. There aren't any real tricks for reaching the front page. It comes down to making something people on HN want to read and then avoiding stupid mistakes when posting. For the latter, I have collected all the information you can find online. You can save yourself some time and just read my Hacker News Front Page Guide.

HackerJobs on the HN front page

I posted a story titled "Show HN: A structured list of jobs from 'Who is hiring?', parsed with GPT" to HN on Wednesday, March 22, 2023 at around 11am UTC. This is 7am in New York, so around the time the day in the US starts. For me locally it was 1pm and I had planned to be available to react to comments on the submission. It is also important to note that I just posted the story and did not do any "marketing" for it. Although it is not recommended to try to get upvotes from friends, etc., I guess that many submissions at least get a handful of upvotes from acquainted users. In this case, there were no such friendly upvotes. Everything developed organically (which is good to know because it shows that you really do not need to have an existing network to push submissions).

The post started in the "shownew" section and quickly jumped over to the actual "show" section. There it stayed for some time until it really happened: The post reached the front page! It then almost immediately fell down to the second page and I was already filing this as a mostly failed attempt. But to my delight, the post recovered and got back to the front page where it stayed for almost one hour.

I made a few screenshots of different positions on the page, but I hadn't really set up any kind of tracking. Luckily, I later found the site https://hnrankings.info which seems to crawl the front page every 5 minutes and records the rankings. You can see the journey of my post in the following graph:

My story's HN ranking over time: Source: hnrankings.info/35259897/

In total, my post received 68 points/upvotes and got 28 comments. This is far from what the most popular submissions would achieve, which often stay on the front page for over one day and get hundreds of comments. Still, even this limited exposure provides us with some interesting numbers to quantify what being on the front page can do for a site or blog post.

Website stats with Plausible Analytics

I am using Plausible Analytics to analyze the traffic on my websites. Plausible is an Estonian company "dedicated to making web analytics more privacy-friendly". I am not affiliated with them, just a happy user.

A screenshot of the Plausible Analytics dashboard for hacker-jobs.com on the day of hitting the Hacker News front page.

One important thing to note: Plausible's tracking is based on JavaScript. I expect the HN audience to include a higher number of users with JavaScript turned off or using some kind of ad/tracking blockers. So the actual number of visitors might be a bit higher than what is recorded (and shown in the following).

Plausible allows to export the data as CSV files. This is what I did for the launch day and the subsequent days. I then analyzed the data myself.

How many visitors can I expect from being on the Hacker News front page?

Probably the most important question is: How many visitors do I get from having a successful post on Hacker News? The following graph shows the raw numbers for unique visitors per minute:

The number of unique visitors per minute while being on the Hacker News front page.

One can clearly see the moment the post reached the front page and also when it fell down to page two after about an hour. To make the number a bit easier to see, I plotted another graph that shows a rolling mean of a 10-minute window:

The number of unique visitors per minute (10-minute rolling mean) while being on the Hacker News front page.

We can see that for this post, we got about 35 to 40 unique visitors per minute while being on the Hacker News front page. After falling back to page two, the number of visitors also dropped to about 10 unique visitors per minute while being on the second page.

Now assuming that a page that is a bit more popular and that stays on the front page longer would also attract a bit more visitors, I think a good rule of thumb might be that being on the Hacker News front page gives you about 50 unique visitors per minute. That is 3,000 visitors per hour or 72,000 visitors per day.

For my post, since it stayed on the front page for only about one hour, the total numbers are not as exciting but it is still a huge peak in visibility for a site that freshly launched and had zero traffic otherwise. Looking at results over several days, the following plot shows the total (i.e., cumulative) number of visitors to the site:

Cumulative number of unique visitors after being on the Hacker News front page.

After the initial about 3,500 to 4,000 visitors from being on the front page, we got to almost 8,000 unique visitors in total over the next four days. I guess these visitors often come from other sources that aggregate or re-publish the posts of the HN front page. For example, there is an RSS feed for the front page. Everyone subscribed to this feed will see the story whenever they use their feed reader.

I guess that the long tail of the visits scales sub-linearly with the time on the front page. As far as I know, many aggregators include a post as soon as it hits the front page, no matter how long it was there. So I think that almost doubling the visitor count in the days after the front page exposure is a result of being there for a relatively short amount of time.

Where do visitors see a Hacker News post?

People read HN in different ways. Looking at the referrers of the visitors can give us some insights into where people found our story. The next plot shows the main referring websites of the submission day.

Referrers of the visitors coming from Hacker News.

Most of the visitors came directly from the HN website. The second largest cluster of visitors did not include a referrer in the request. I assume this cluster consists of a mix between browsers not sending the referring site and people coming from other sources, e.g., RSS readers. The only other individually notable source of traffic was hckrnews.com, an unofficial alternative interface for HN, albeit with a large margin to the first two clusters.

The main insight from this data is that most people actually read HN on the Hacker News website, although the HN audience probably would be more than capable to use other interfaces.

Which countries and regions do visitors from Hacker News come from?

When posting your own content or product to HN, you probably hope to get exposure to a somewhat relevant audience. Consequently, it might be important where the visitors are located. Hacker News is a US website but probably has been able to attract a global audience for many years now. Identifying a website visitor's current location is usually done by mapping IP address ranges to certain regions. I am not sure how Plausible Analytics does it exactly but I guess it does something similar. Consequently, the estimated locations have quite some margin of error but nevertheless help to get a coarse picture of where visitors come from.

The first plot shows the main countries the visitors to my site came from.

Countries of visitors from Hacker News.

The US-based traffic is clearly the largest cluster, almost larger than all the other countries combined. Next up are European countries such as the UK and Germany. Notable sources of traffic also include Canada and India. Most of the traffic comes from Western countries, which probably is true for Hacker News in general, but also might have been affected a bit by posting during the US daytime and thus during night for most of the Asian countries.

Plausible Analytics allows us to also look at regions and cities. The next plot shows the main regions from which visitors came from.

Regions of visitors from Hacker News.

The number one region where the HN audience seems to be located is - unsurprisingly - California. However, as we can infer from the huge size of the "Other" cluster, the regional origins are relatively diverse (within the previously seen bounds of the main countries).

Although the estimates might not be very accurate, we can also look at results at the city-level.

Cities of visitors from Hacker News.

It is no surprise that we mostly see the largest cities from the previously mentioned regions.

The analysis of the visitors geo locations shows the general Western tech world bias of the HN audience, but this also confirms that for tech and entrepreneurship-related content and products, it can be a very relevant source of visitors or consumers.

What devices do people use to read Hacker News?

Let's say you want to promote your new product on HN, then it might be important to know whether people use their mobile or desktop device when arriving at your site. The next plots shows the distribution of devices while my site was on the HN frontpage.

Devices used by visitors from Hacker News.

In this case, I got almost a 50:50 split between desktop and mobile devices. I assume that the proportion of desktop users is higher than what you would get on a typical website, but I guess this makes sense considering the more techy audience of HN. This also is reflected in the used operating systems as we can see in the next plot.

Operating systems used by visitors of Hacker News.

The most used OS of the visitors from HN is MacOS on desktop and iOS on mobile. Especially on desktop, Windows has a considerably larger market share, so this is a somewhat special situation with the HN audience. Additionally, we see a significant number of Linux users.

Finally, we can have a look at the used browsers. The stats can be seen in the next plot.

Browser used by visitors from Hacker News.

The clear leader is Chrome with Safari holding the second place. Interestingly, there are almost as many Firefox users as Safari users.

Is it worth it to try to get on the Hacker News front page?

As I have shown in this article, getting to the front page of HN leads to a considerable amount of traffic for the submitted website. Whether the traffic is worth a lot or not depends on your target audience. If it matches the typical HN audience, it might give your blog or project a boost which results in some users or regular readers. The good thing about Hacker News is that it still has a relatively high bar for quality and relevance of the content. This also means that you should really think about whether your topic resonates with the HN audience. If the answer is yes and if you create some good quality content, there's a good chance that your submission will get some traction on HN. Don't forget to read my Hacker News Front Page Guide in order to avoid stupid errors!

Impressum & Datenschutz