This post aims to open up the gates as to what to expect if you're fortunate to get an article featured on Hacker News. I've been a daily reader of HackerNews for months now and it's been a great source of interesting tech news. My general workflow is to scan articles in Reeder and push them to Instapaper to read fully later. If seeing these traffic patterns helps you in anyway, don't be afraid to drop me a tweet :) @jimplush
Yesterday I was surfing through a new site called Forrst thanks to an invite from @ka. I came across a resume mockup from Jamie Kite that I thought had some great ideas behind it. I crafted up a quick article and posted it to my blog with my own thoughts on it. I had just started the process of migrating my old blog to a new code base based on the Code Igniter framework. I wanted to have complete control of my blog, but it's been a challenge finding the time to actually finish it (For example I'm still putting my articles directly into my DB at this point). It was just a quick post to get some content going again.
I posted it around Noon Saturday and it has had a 24 hour+ run on the front page of Hacker News so there are quite a few things I've taken away from yesterday and today. My article "This may be the best resume I have ever seen" shot to #1 on Hacker News yesterday and remained #1 for over 5 hours straight I believe. It was in the top 10 for over 20 hours. I hope I didn't cause Jamie any unwarranted attention but I do believe she's onto a great idea and wish her the best in her future. So here is an inside look at how it came about and the aftermath of stats 24 hours later.
Before I submitted the article I wanted to install a more real time analytics package. After briefly surfing around I came across GetClicky which offered a free package if you're under 3,000 pageviews a day. Since I knew I'd mostly likely never hit that limit I signed up and installed my tracking code.
How I submitted:
This was the most shocking part to me. I just hit the submit page, put in the title and url and fired and forgot. I didn't ask anyone to click it, no gaming of the system. I had just happened to log into my getclicky account little while later to see if I got any hits from my submission, I usually see 10-20 referrals for just submitting links. To my surprise I had hundreds of online visitors. I figured then I had made the front page somehow so I started keeping a closer eye.
The Hacker News Effect
Visitors kept climbing, up and up into the thousands. I had realized I had not implemented caching which cause me to log into my Linode server and got a lay of the land. Apache was buzzing for sure and MySQL was getting hit pretty good. I quickly installed a Redis server and implemented caching on the articles on the article detail page to take the DB load off. That made all the difference in the world as Code Igniter is known to be a super fast PHP framework. Load remained low after that and I just loosely monitored.
By midpoint the next day I had roughly 30,000 visitors according to getclicky and 42,000+ pageviews. The benefit of real time analytics cannot be overstated and here's why.
- Server Failure - Had I not known right away my traffic was blowing up I would have not implemented caching in time and my server probably would have kicked over, ruining the viral run.
- Seeing inbound links coming in enabled me to interact with those re-posted the link on their blog and it also allowed me to push out links from Content publishers who were picking up the story as well.
- Using BackType and Bit.ly's stats page I was able to follow the twitter action and the virality of the link as it made its way through twitter and facebook and @reply people to further help gain more traction.
- Google Analytics is the old man in the room - Had I solely relied on Google analytics and had to wait until the next day I would have missed a lot of opportunities to interact with people and make some quick decisions when the article took off. Google Analytics is no longer your friend.
Interesting Stats and Info
So what are some interesting stats and things to take away from the rush of the Hacker News clan?
- Make sure you have caching enabled - it's just lazy not to. Even if you just cache some of the most used DB calls, you're ahead of the game.
- Have the ability to let people tweet or like your article from the article detail page. I failed on this and regret not adding it now for another small boost of traction
- Give users somewhere to go after they're done reading. Watching my stats in real time I was able to see the mass bounce after people read the article because I had no where to send them. That was a huge takeaway from yesterday.
- I started the day with 110 twitter followers and wound up with 136 for a net gain of 26 of a possible 21K people. (sadness)
- I received 14 @ mentions on twitter directly
- I had one email from my about me contact page
- My Hacker News Karma score went from 110 to 413
- I got to see the long tail of who redistributes Hacker News content -> 1,000+ small sites out there driving traffic from Hacker News links
- With realtime analytics and mobile apps for it you can make realtime decisions that can help keep sustained momentum.
Show me the money!
So how much cold hard Adsense cash did I get yesterday? I must have cleaned up for that giant traffic spike! Alas, 4 clicks out of 30,000 page views for a total of $2.76 :) So don't expect a goldmine from Hacker News traffic.
Below are screen shots taken through the day and aftermath, hopefully this info can better prepare you for your Hacker News effect.