The dying of coupon apps and websites

This picture has been in my review list for a while. Coupon sites and apps were having a tough time, and none of them seem to be able to find a viable business model.

This particular one got bought by an Indonesian e-commerce group, and trying to pivot to an online market place.

Must be tough to compete with sites like banggood, ebay, aliababa and TaoBao.

My take for new e-commerce players? Find a niche that is profitable. You do not need to sell a million dollars a month to live lavishly.

創業兩年後的轉變 #1

兩年前開始自立門戶，自家 Mobile 及 Web App 公司正式開業。

香港開公司同日本差唔多，就是老豆老母同打份工的朋友都覺得你冇錢開飯。客觀的事實就是，冇錢供樓交租，都很難在這裡生存。10 間有 9 間的公司，5 年內都是結業收場。

當時為了極速撲水救忙，就開始一邊狂接 projects，一邊自己做 product。當然日日都幻想同個客收完錢，就會有時間做 product 掛。實情係佢會逢星期一至五都會 call 爆你個電話。到星期六就當然會有工作要趕，好似 bugs 呀 (我就當然美化咗個名叫 "issue" ) ，另外有 project specifications 入面冇既 item 又因為係 grey area 入面，唔做又怕收唔到尾數順手做埋佢，總之就個 product plan 完全不記得。

冇法，吃飯要緊。

一開咗公司，就多咗好多人，想同你做下 partner，又話者要做 advisor 之類。

故事 #1，佢好有 marketing skill 又好似識好多人，總之你有個好既 product，佢就會揾 investors 或者 leads 俾你。你就係要揾地方 R 到錢或給公司股份回報佢。

故事 #2，佢好有 idea，總之就你間公司出時間人力物力，佢就負責 "business development"，去開下會呀 make 下 connections 之類，但係永遠係 50% 50% 股份，呢類「公就你贏，字你輸」既 startup 行騙手法，真係又見唔少。

要生存，就一定要有定位。你間公司要有技術，要有一定既 network 同 marketing 技巧。呢度要好強調「自己公司」呢 4 個字，core business 一定唔可以發夢 outsourece 俾人做。真係 "not in sight, not in mind"。

市場上，兩年前做 startup 最 hot 的就當然是一模一樣既 iOS 及 Android Apps 了。現在就多了好幾個出路:

- Hardware
5 年前，呢一條路，如果你唔係「廠佬」就想都不用想。自從有咗淘寶後，去大陸訂 electronic parts 可以今日下午 4 點落單，明天早上就收到。

可以做硬件產品，可以玩 Bluetooth，可以玩係曉飛既 drones。

- B2B / B2C
又係要多謝淘寶，等人知道網購都有得做。(無可避免，突然有好多新客話要「做一個淘寶」出來。) 有同化咗為同胞既朋友，會成日話「唉，香港网购唔够国内好」，只可以講，國内的政策，香港人都不要再發夢可以有份玩，你控制不了。

但是還可以 go global 的。手機購物又真是天方夜談? 香港網購真是沒有得做? HKTV 都是在玩的?

耐不耐就有 IT 人抱怨自己去到 35 歲都係揾兩萬幾，其實你有大把機會，係你未把握到。

[iOS] Fascinating stuff: NSDataDetector and CFStringTransform

I am working on a new iOS app that needs to parse address and phone number nicely.

Well, turns out Apple has this NSDataDetector thing. Nicely covered by NSHipster.

Another great stuff is this CFStringTransform. You pass in some Chinese text like "中文" and it will spit out romanized version of "zhōng wén".

How-to guide from "Day Job" to "Consulting" to full "Product Company"

If you are a developer/programmer, this guide is for you.

There are tons of blogs that tells you it rocks to be a developer or how special you are. I am going to spare you from that.

I just want to chime in on my experience on how to transition from working for "teh man" to have your own startup.

Stage #1: Save for the stormy days ahead

Rule number 1 is always have enough money in the piggy bank, to pay for rent, food and other expenses. Basically, the more you can save the more runway for you to launch your company. Ideally, you should save enough money to live off your savings for 2 years. There are always some kind of emergencies, so be prepared.

Entrepreneur classes in college teach you that the 3F: Friends, Family and Fools are you best bets to raise money. Unless you are a real genius, you will fail on your first startup, so obviously try to avoid them if possible.

Some places like Hong Kong and Malaysia have some funny government grants and loans for "high tech" startups. I personally think these grants are curses, as I have never seen these companies actually gained tractions. Seriously, if you cannot bootstrap a startup yourself, you would not be able to make it with US$40,000  from some funny grant programmes anyway. But prove me wrong. 

So in this stage, working for "teh man" is inevitable, besides you can learn some valuable skills and toolsets, whether it could be Microsoft, Java or iOS stuff. So, be proficient as you are mostly on-your-own in a start-up.

There are a lot of romantic Silicon Valley stories about bootstrapping a company from credit card. Either these are mostly outlies or they are made-up by some PR companies.

If you do moonlight as a pirate, you should keep pillaging some more ships. Start networking with some people for business opportunities and fun purpopses.


Stage #2: Getting ready to launch your tech company

Don't quit yet
It is probably a myth that you really have to quit your job in order to "commit 100%". That is really not necessary. All you have to do, is to spend 15 to 20 hours after work to start building your product, or consulting for some clients.

Idea time
Start planning for your own product, whether it is a big giant AI robot, iOS app, Android game or some enterprise product. You definitely want to get your idea and products validated if possible. That is another post.

Consulting
Now is also great time to start consulting for some clients. Start asking for consulting work from your network. Build a nice company website with excellent CSS frameworks like blueprint css. Start shipping _some_ apps to be your portfolio. 


If you can finish an app for your client _alone_, you should probably be alright, otherwise, you should just probably forget the startup thing, and go back for paychecks until you have the chops. 

Incorporating the company
You may consider to incorporate the company when you start getting tractions on your clients. USA (at Delaware) and Hong Kong are great places for developers to register your LLC or Limited. These 2 places also have great credit card gateways support, such as authorize.net, Stripe and the lesser known AsiaPay.

China is definitely a mess, so unless you think you have the energy to spend 3 to 6 months on it, don't bother. Adjust your business plan to collect money elsewhere from China if possible.



Number of co-founders?
Paul Graham's Y Combinator would recommend you to have at least 2 co-founders. I would agree with it, but do not afraid to start one if you think you have a great product anyway.



Stage #3: Going solo


You can consider to quit your job and go solo if:

1. You are mentally prepared for a roller coaster ride
2. You really have 2 years of cash in bank
3. And have the complete support from your significant other(s). 
4. You have got your product prototyped

Congratulations, your sufferings have just begun.


Consulting vs. Product
To stay afloat, you just need decide how much you want to split between "Consulting" and "Product" while having cash-flow positive.

Ideal case: You spend 0% of your time on Consulting, and 100% on Product. Your app got listed on AppStore frontpage, then company either cashed-out, get bought out and live happily ever after.

Of course, this is not likely to happen. Keep looking out for local opportunities on developing apps, especially during these period of sky-rocket growth period. US, Europe and East-Asia developer should avoid work from outsourcing site like ODesk. Your app startup could obviously undercut bigger players initially, with lower operating cost.


Office
If you are an app or web developer, trying not to have an office will be a great money saver initially. Your clients simply will not go to your place 99% of the time. Offer to go to their office, or meet at a Starbucks. If your work is good, they would not care where you pound the code out anyway.


Of course, office is a must if you start to hire. Do your best to find a quiet and convenience place obviously.

Practice makes perfect
New app startups would definitely inevitable sucks at shipping app at the beginning stage. Therefore, ships as many as 10 to 20 apps a year so you can to perfect your skills shipping apps.



Here is the part 1 of "Day Job" to "Consulting" to full "Product Company".  Part 2 is all about product.


Reference:

1. We are entering the age of developers
2. You Are the Exception to the Rule

Introducing Commute.hk 香港交通相機 iOS App

I just released a new iOS app called the Commute.hk 香港交通相機. This app gives you quick accesses to all the traffic camera views in Hong Kong.

You can browse the cameras easily by using map, list and grid views, as well as adding your own favorite cameras for easy referencing.

All traffic news and camera imagery are provided by the Data.One service operated by the OGCIO of the HKSAR government.

If you drive a fast car, I am sure you will find this app very helpful. By the way, this app will push any special news traffic info right to your app as quick as it gets.

Here are some more screenshots:

HK eTransport vs. Google Map, and just open up the bloody GIS data!!

There is a new transport app called HK eTransport (iTunes AppStore) for those who gets lost all the time.

And here is what it looks like on the app:



And here is what it looks like on Google Map:



Apparently, the difference of Google vs. your garden varieties GIS app is, they understand the concept of "ranking" better.

Okay, as a developer, I can understand that "HK eTransport" have put a fine job, but please, just open up your bloody GIS data for the pros to do it!!! The only thing Google Map missing is those mini-bus routes, and I bet it is the bureaucrats who are getting into the way.

Check out the NYC Big Apps 2.0. Are you listening?!!


Update: it looks Google Maps have all the HK minibus routes properly supported.

Smartone Vodafone Callguard 來電管家

I am pretty happy thatSmartone Vodafone has done a junk call filter, at the operator level.

While they charge you $15 HKD, I think it is a reasonable price for a simple peace-of-mind. I have yet to see what kind of filtering implementation they are doing.

Ideally, operators should do a "Google Voice" service, so you can report spam with Bayesian filtering.

Bravo, Smartone Vodafone.

Helo advertising

Interesting helo marketing campaign by ICBC.

I can understand a helo campaign at NYC or Santa Monica beach, but seriously, I don't think anyone can spot the banner without a telescope in HK.

Apple Store HK almost coming

The new Apple Store Hong Kong has the wall finished. Guess they will be opening soon?

網上創業小秘笈 2.0

(呢篇文在 2007 年喺 b.fleur.hk 首先發表)
要營運一間搵到錢既網上公司一啲都唔易。以下係小小貼士希望大家行少啲冤枉路。

0. 香港公司註冊
以前開公司都幾煩。無限公司去e-registry自作業。有限公司都係俾 d 錢搵 agent。Google 大把。

1. Marketing Marketing Marketing!

宣傳係好重要既。 因為無論你個 Website 做得幾好, 無人去你個 website, 有咩嘢意思?

做 online marketing 你可以試下買 Google Adwords。Yahoo! 就唔係好得，成日有 SEO 公司同您 sell 埋簽一年包 X 個 keywords 就唔好上錯賊船。Facebook 就一時時 OK。

Google Place 記住要做。唔洗錢的。

又不妨自己可以試下做 search engine optimization, 同埋設計下的宣傳品同單張派下咁。但係唔該唔好搵 SEO 公司幫你增加流量，因為你 quantify 唔到佢既工作其一，佢地既所謂 report 係超級搵笨。有錢要找個天下無敵既 developer 同你做好個有 business model 既 website!

學人地做一個 iPhone/iPad/Android app 做 brochure 未必有用。用 Onswipe加 Wordpress 咪得喇。如果你既 App 有 Gimmicks 的話，就儘管去把！

2. 善用 Web 2.0 年代的新技術

Webhosting:
做一個好網站，page loading 一定要快。最好就要喺 4 秒之內要完成。所以一間快既 webhost 係好重要。要再改善 page loading 時間，不妨用 Linode 或者 Amazon S3 去做分流。例如 HTML 源碼可以放在 webhost，Image 同埋 Flash 檔案放在 Amazon S3。Linode 呢 D 公司有Load balancer同埋Backup，至少夜晚訓得落。

DNS:
大家千奇唔好忽略 DNS，佢對你 website 既速度有少少影響。如果你係用 Godaddy 呢類 domain registry，佢地係會送 DNS 服務的。用 .hk 域名係冇的，所以要自己 setup。好多 VPS 好似 Linode 會俾你任開 DNS zone，要想再快同 professional 的話就用 DNS Made Easy。

Blog:
你可以用 Google Blogger 去 host 你個 blog。現在你重可以用自己既 domain 同埋 subdomain 去 用 Blogger。 Wordpress都 OK，但係記住要 update﹐唔好俾人 hack。

流量統計:
用 Google analytics 可以知道有幾多人入去你既地的網站。以前無 Google Analytics，做流量統計又要行 Analog, 一係用 AWstats。 現在有 Google Analytics, 看 trend report 真簡便！

3. 電腦軟硬件
要善用 open source 軟件。呢家唔使樣樣都俾錢做既! 好似用 Linux 去行 server，用下 Django 同埋 Java 去做下 shopping cart。Database 用下 MySQL 同埋 PostgreSQL。唔洗樣樣都 Microso~1 前 Microso~1 後。 最好自己寫 Shopping Cart，話哂你係做 online business 喇? 真係想懶的話可以參考 Magneto。

E-mail 系統就一定用 Google Apps Standard喇。費鬼事自己 setup。

記住用 Exchange 去 setup 你公司既 iPhone/Android 手機，呢家唔興 Blackberry 喇，so 2007。

電腦一定要 Mac 機﹐唔洗煩 Viruses (暫時...) 同埋可以用 Time machine。今時唔同往日，Mac 要用 Windows 的話可以行 VMWare Fusion。

番大陸幹活就一定要有VPN，唔係就去唔到 GMail 同 Facebook。冇朋友者的話就自己 setup: 去 Linode 伸請一個 slice，再 install pptpd。

4. 電話系統
如果你想要個 "2/3" 字頭的電話number, 可以用 NWT 隨身固網電話。 幾蚊就有個 "2/3 number forward 去任何香港電話 number。 固網假裝做手機。但係轉電話號碼要 call 佢地 hotline，有小小麻煩。

又可以試下 HKBN 寬頻電話。可以周圍用, 屋企又得，office 又得。依家唔使一定用固網電話既! (記住唔係 '2b' 用 USB 嗰種。係有個電話盒嗰種。)

有冇想過電話唔洗下下一定用 Land line 先得？介紹大家用下 VoIP。用 Asterisk 去 setup 你地公司既 dial plan, 打入同出可以飛黎飛去。再勁 D 的話可以寫 program 去控制 (Tropo, Twilio)!

Fax 機可以試下 HKBN Fax-to-emai 呢一類 service。但係，千其唔好用 "e-fax". 佢係黑店之中的黑店!! 唔信就去 sippey.typepad.com 度睇下。

Plans to revive HK Junkcalls

I have to admit that HK Junkcalls is absolutely a product by accident. So far it has been downloaded over 100,000 times for the past 10 months.

The application does its work by importing numbers marked as junk call to your iPhone address book. Initially, the program only needs to import 2000 numbers. However, the list is slowly growing all the way to 8000 numbers and counting.



The addressbook in your iPhone is actually powered by an sqlite database. The bigger the list, the slower the your iPhone can properly lookup a phone number, which affects the performance of receiving a phone call. Above all, syncing the huge list of address book with your Mac is definitely not something very funny.

The good news is, iOS 4.2 and above let you put address book numbers "on your iPhone" only (first tipped off by a kind poster at mac.hk).

Therefore, I will restart developing the HK Junkcalls iPhone app more actively.

電訊業將轉變收費模式

iPhone Adhoc Over-the-air install

It is always a tedious task to build an ad hoc IPA app for beta tester to try out your iPhone app. First you need to ask them the UDID, which they may helpfully give you a screenshot of iTunes... then you have to build an IPA, and prayed that they know how to do anything with the IPA.

Now, there is this "Hockey" framework (Announcing Hockey, an iOS developer framework) that can automate the process.

For integration sample, you might want to check out this PSiOSAppTemplate.

Find my iPhone "API"

Apple rolled out free "Find my iPhone" for those on iOS 4.2 with an iPhone 4.

The nicest part is, you can programmatically track your iPhone 4 now "with an unofficial API".... for free!!!!

Plenty of examples in Ruby and PHP.

Alpine on Mac OS X

I have been using "pine" since the old unix shell account time. Remember those free stuff like http://sdf.lonestar.org/?

Turns out using "alpine" (the modern version "pine") is pretty straight forward with GMail.





1. First, compile "alpine" using MacPorts by "port install alpine".

2. Then, run "alpine" for the first time, which should generate a nice .pinerc at your home directory.

3. Okay, just change these few lines accordingly:

# Over-rides your full name from Unix password file. Required for PC-Alpine.
personal-name=YOUR_NAME

# Sets domain part of From: and local addresses in outgoing mail.
user-domain=gmail.com

# List of SMTP servers for sending mail. If blank: Unix Alpine uses sendmail.
smtp-server=smtp.gmail.com:587/tls/novalidate-cert/user=YOUR_GMAIL_USERNAME@gmail.com

# Path of (local or remote) INBOX, e.g. ={mail.somewhere.edu}inbox
# Normal Unix default is the local INBOX (usually /usr/spool/mail/$USER).
inbox-path={imap.gmail.com/ssl/novalidate-cert/user=YOUR_GMAIL_USERNAME@gmail.com}inbox

That's it!!! Welcome to 2010.

Sounds Right 1.0 is now available on AppStore

Sounds Right 1.0 (iTunes AppStore link) / (British Council hosted page) has been available on the AppStore. I am glad that the reception is fairly good given this is one of my first iPad apps shipped.

ABC Nightline - IDEO Shopping

iPhone OS 3.1.3 enables Tethering

rdiff-backup with SSHFS

Very useful guide to setup remote backup with rdiff-backup from Linode:

My approach is to use a Snow Leopard machine with an External USB Harddrive to backup my servers. On my MacBook Pro:
1. Install MacFUSE
2. Install MacFusion
3. Then use macports to install rdiff-backup

Then, run MacFusion to mount the remote directory and run the rdiff-backup.

However, you need to remove the sshnodelay.so file in the MacFusion ssh plugin directory in order for sshfs-static to work.

HK Junkcalls #2 on iTunes AppStore Hong Kong

I'm very pleased that "HK Junkcalls" is now #2 on iTunes AppStore Hong Kong. (Link to iTunes AppStore)