Friday, April 7, 2006

What I learned - My Online Quiz (2001)

My Online Quiz (Age 14 - 2001)


It was 2001 and the "it" thing to do was to create a quiz about yourself (or anything really) and send it to your friends to see how they would do. 

There was one big site doing it, that site was called Quiz Your Friends and after having countless friends send me their quizzes to take, it gave in went to make one myself make one myself.   But, you see Quiz Your Friends site was just terrible, I do mean terrible.  It was full of tacky ads and completely Flash based.   So, I found myself in a dilemma,  I did want to make a quiz, but just didn't want to use Quiz Your Friends.  At this point, I had already seen playing around creating dynamic web applications for a little over a year, and this quiz site seemed like it would be fairly basic application and figured I could just hack out my own version of it.

24 hours later...  I had written a basic application that let users sign up, create an unlimited number of 10 question quizzes, and each quiz had a high score.   I registered the domain MyOnlineQuiz.com and made the application live.   

At this point, I made the first quiz and sent it to 10 friends and waited..  Not only did they all take it, they went in and all make quizzes themselves and sent their quizzes to their friends and so on.  Within hours of making my first quiz, I had 50 users, by the next day it was 200 and the following day it was at 500 users.  This trend continued over the next several years and reached a peak of 2.3 million users at the time I shut it down in 2008.

What did I learn from My Online Quiz?

Technical Side:

Database Design
My database server was a 550MHz / 512MB box running SQL Server 2000 in my parents basement. The Database itself was over 30GB (storing nothing but text).   and I spent tons of time refining queries and storing data in temporary caches that I might need later for that user.

Hacker Proofing
Never half ass code and leave something exposed, hackers will find it and exploit it.   A few months after launch, I became a victim of an SQL Injection Attack that renamed all the quizzes to  "{{You've Been Cracked}}".  

Back Up!
I was extremely lucky that I had setup a daily backup from day one and was quickly able to revert back after the attack after fixing the vulnerability.   If I had not set that up, I would have lost the thousands of quizzes created and all the user accounts.


Business Side:

Setting Trends
When AIM started allowing Limited HTML web-sites to load in peoples profiles.  I realized that I could leverage it do allow people to add a quiz to their AIM profile and their friends could take it right from there.  What still might be the most flattering moment of my life, within a month QuizYourFriends had completely copied it and added the same capability to their application.

The Network Effect / Going Viral
The only goal with this project was just to make a quiz that didn't use QuizYourFriends,   Thinking about a growth model wasn't exactly on my young teenage mind.  However, seeing what was happening organically caused me to start actively thinking and factoring in growth strategies when designing new features and in feature projects.

Create Something You Need
This one goes without saying.  The best products are created by people who see a need and not trying to make money.  Might sure you're solving a problem and not just creating something cool.

User Experience Matters / Analytics Driven
From day one I found myself playing around with how things to functioned and the design.  An accidental blessing was that I had enabled basic logging setup from the start.  

At the time I was just interested in traffic numbers, but I quickly realized the power of analytics.  I discovered things like only X% of people who went to the landing page actually went to the sign up page and created an account (required to make a quiz).

I redesigned the flow to allow the user to start making a quiz from the landing page without having to sign up for an account up front, and I did still required them to do it, but moved it to the end of the flow.  The result was about a 100% increase to quizzes being created.


Allowing users to start making a quiz from the landing page (without an account upfront).


Social Networking
I like to joke that I created Facebook before Facebook (I didn't).   But I did discover early on that quizzes were being created in clusters in friend groups.  This was extremely noticeable with users using their .edu email addresses.   Once I saw one user from a university there were many more to quickly follow from the same university.

This made me think to add a basic social profile to user accounts, including a profile photo and let them tie it either a high-school or college and let other users from that school quickly see what other students had created quizzes also and it was a big hit.