This is a talk for the WAUS forum (Workshop on Advanced and Usable Software) I've presented on Dec 14th, 2012 @ National Cheng Chi University, Taipei, Taiwan
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
A Startup Perspective to Service and Software Development
1. Danny Lin
Managing Director, Plurk
Workshop on Advanced and Usable Software (WAUS)
Dec 14th, 2012 @ Taipei
2. Danny Lin
• Software Developer, Cartoon Company
• Movie Screenplay Writer
• Product Manager @ various scale of software companies
• Columnist @ various magazines for over a decade
• Operate a BBS (FidoNet) for more than 8 years
• Head of Communication Products, Yahoo
• Lead Product Manager, Google
• COO, Online Game Company
• Managing Director, Plurk
3. Founded by 3 co-founders in Toronto, Canada in 2008.
Officially launched to public on June 2008.
Anonymous (virtual identity), horizontal timeline,
emoticons, karma/badges
We wanna make it funny and easy.
It can be your meme
It can be discussions
It can be chitchat
It’s social.
100% using open source software
Our team members are also open source contributors
We have open source projects: http://opensource.plurk.com
4. What is a Startup?
startup is a company or temporary organization
designed to search for a repeatable and scalable
business model.
my 2 cents:
Very limited resources
Everyone is multitasking.
Scalable. Flexible.
6. Almost real-time.
Tons of requests in
Earthquake!!! queue.
User experience.
7. All your metrics are growing in a positive way,
and you’re happy with it.
A exponential problem.
timeline data
size
Life is good,
But hard.
number of users
8. How many money do you have to support the
growth?
I don’t have so much money.
I need to feed my team for the next 12 months.
I don’t wanna kill myself or the service.
Do you still want your users to feel happy?
Off course. I truly do.
I hope they feel no big difference if a change
is needed.
9. Spend $$$$
Buy new servers
NO.
Don’t do that.
You’re not Google or
Yahoo.
10. Re-architecture?
Rewrite your code to get better
performance, handling more
requests?
GOOD.
Not perfect. And it
takes time.
New changes might
screw up everything.
11. Look at how users are using your service!
female (59%) , male (41%)
Heavy (10%), Medium (85%)
Low (5%)
85% users posted more than 8
new messages, 20 responses, read
approx. 30 messages per session.
CONCLUSION: They care more about timely
responses than getting a huge number of updates.
12. Don’t wanna spend too much $$$$.
Not buying new servers.
Make minimal code changes.
Users feel no big difference.
Limited size of updates list.
Satisfy 80% of users’ need.
13. Limited size of updates = 80+% of user need
A linear solution to an
exponential problem.
timeline data
size
Life is better,
And cool.
number of users
15. control your breath
I might quit. well-equipped
You need a race “strategy”
You need support
It takes time to practice and
It’s a long journey prepare
Photo courtesy of Nordea Riga Marathon, under Creative Commons License.
16. It’s a quick run.
Need to quick react, and be really solid.
just do it,
no perfect “plan”.
just less than 1 mins,
and it’s done.
Unload unnecessary concerns.
A short race at top speed.
Better cycle to fix things.
Photo courtesy of abby chicken photography, under Creative Commons License.
17. Grassroots effort -
Users are your customers At Plurk, every employee is
also a customer care staff.
Users are your friends
Directly and immediately
Users are your collaborator responsive to the needs and
Users are the best beta testers wishes of the people
involved.
What do they get in return?
In the end, it’s about sensation.
A badge, an official thank, being appeared on
your page, something they feel cool, or
honored, would be seen as a reward.
18. Translators
Content Moderators
You getlanguages translatedfrom the community.
36 the help from users by Plurk users worldwide.
TheyThey love to contribute.
understand what the community need.
They feel honored, in return, get a Plurk translator badge
They are real people, not robot.
shown on their profile.
20. Quick React.
Fix critical ones first. (what users cares the most)
Unload unnecessary features, learn from users and
data.
Release tiny features frequently, less than 3 at a time.
Do something funny and sometimes crazy instead of
following the rule.
Don’t follow every move your competitor did.
21. Think Out of the Box
No one define rules for you. You rule it yourself.
Explore new needs from users and data.
Trust your new idea. Do not hesitate to try.
Take it easy. Make it simple.
Enjoy the complaints.
Don’t try to release 100 features at a time.
Simplify development processes.
Simply your product designs.