Watching the Web Discover goingtorain.com
The buzz around and traffic to goingtorain.com is slowing down now and I'm amazed how many people responded positively and thought it was actually useful.
While talking to Dave about it yesterday, he remarked something along the lines of, "of all the awesome, useful shit you've ever made, the thing that finally became famous was this stupid little site."
I tried to follow the traffic in real-time when I was at home, tailing the
site's log file and seeing where people were coming from, both in terms of their
city/country and the referring website they clicked through from.
As a new referring URL would frequently pop up, I'd visit it and see how people
were reacting to the site.
The initial version of the site just displayed the answer and the forecast in
Fahrenheit, and as mention of the site was quickly making its way outside of the
US, a number of people commented that they wanted to see the forecast in
Celsius.
Amazingly two people actually e-mailed superblock.net
sales@ and support@ addresses with suggestions about the site even though
Superblock wasn't mentioned anywhere on the site except for the HTML source.
After changing the site to show Celsius for all countries outside of the US, I
also (shamelessly) added a small Superblock link at the bottom which has
generated over 2000 referral hits.
The influx of traffic to goingtorain.com started on Thanksgiving evening from StumbleUpon which has since generated about 17,000 hits. I think the first few people that commented on it assumed it was a static site like this one and quickly dismissed it.
From there it was posted to Hacker News where it was quickly moved to the top of the front page and generated about 5000 hits. It now has 102 points but, due to its age, is moving off of the front page. The discussion there was mostly positive and gave me some ideas for the Celsius change.
At around the same time, the site started making its way across Twitter where it has now been mentioned 270 times. The thing about Twitter i find most interesting is that it's not just people clicking an up arrow on a link aggregation site or replying with a short comment, but that they actually have to make the effort to copy and paste the URL, write something about it, and think it's worthy of sending to the 15 or 15,000 people following them. Some notable twitter updates which made me smile:
"The web's simplest weather forecast"
"Ever wondered why it's so hard to find your weather on most weather sites? No more"
"The simplest way of figuring out if you need an umbrella or not. I love it!"
"simplest weather site there is. Implications for health care, taking lots of data and clarifying"
"This is honestly the best site ever."
"I really, really, really wish I had thought of this first"
One Twitter user replied to Om Malik that the site was a "just a copy of umbrellatoday.com", but since I made goingtorain.com in May and the umbrellatoday.com domain wasn't even registered until august, I couldn't really have copied them. Of course, that's not to suggest that their site is a copy of mine or that such a basic idea of summarizing weather information is even worth arguing over. umbrellatoday.com does look nicer and offers some fancy SMS capabilities, but it also requires users to enter a zipcode and doesn't seem to work for non-US requests.
Eventually the site was posted to Reddit, but it never moved to the front page with only 89 points and seemed to have a lot of negative comments. Strangely someone posted the link again to the worldnews sub-reddit where it received another 83 points. I think if all of the up-votes had been applied to the first link it might have made it to the front page.
People also bookmarked the site 586 times on Delicious and it made it to the top of the popular page. There are some nice comments made by some users as well.
The entire site remains one single 430-line PHP/HTML/CSS file. I'm somewhat surprised (and thankful) that Google hasn't blocked access to their weather API from my server since it's made over 160,000 queries in the past two days. In the event that happened, I was going to change the page to make the browser generate an Ajax request to Google's server, request the weather itself based on the city I passed it from the IP geolocation lookup, and do the XML parsing to show yes/no and the forecast. All of the hits to Google's servers would be distributed among each user's IP, but it would be slower and require Javascript.
IP-to-city geolocation is being done through maxmind.com's web API service which didn't seem to have any speed issues with the amount of queries I was doing. I did notice a few users complaining about it not being very accurate in some countries, but I can't really do much about that.