LunarLuau

The Dollar-a-Day Advertising Strategy

A while back I heard about Eric Reis’ five dollar a day strategy for driving traffic to his site.  He was talking about it when he visited us at SEG.  The gist of it being that you shouldn’t really have a “launch” for a webapp, rather, just let it grow and adapt to your users in a more organic fashion by spending $5 a day on google ads to drive a small number of people to the site every day.

It’s actually a smart idea, because it drives a steady stream of people to the site and allows you to see how they use the site and see what bugs they hit along the way.  This is in stark contrast to simply getting a massive spike of traffic like we did with phonebooking back in 2004.  In that case, anything that was wrong with the site was seen by the thousands of people who visited it from the CollegeHumour link.  The bulk of the people saw the site as it currently was and didn’t come back.  It was very much a flash in the pan (not that I expected that joke to have any staying power, there are only so many humourous ways to assault a friend with a large book).

Applying this to whichishotter, I’ve been using a dollar-a-day advertising strategy on Facebook.  The idea being that I pay Facebook a dollar a day to show an ad for the site and this gets be about 5 or 6 random visitors a day.  The ads are targeted at Canadian women under 21.  It used to be 20-somethings in Vancouver, but Ray Lai had an interesting theory about how social media expands that we’re trying out.  This gets a small trickle of new users who vote on things, click on stuff, and break a few things.  I have a few users who are really good beta testers, but they are people I know.  The average user will not be someone I know and I know really technically competent people.

This allows some of the sites shame to be more private.  The first random user I had trying to signup actually hit an error when doing so; something quite embarassing.  Fortunately, any landmines hit are seen by a small number of people.  I can also track how many comparisons people vote on, which ones they hit, and there isn’t so much data to absorb that it’s like drinking from a firehose.

All-in-all, I’m going for the slow growth.  I’d rather respond to how people use the site rather than force them into certain behaviours.  Like, right now, I’m learning that people are hesitant to post comparisons of themselves.  Perhaps because they haven’t thought of anything interest at that second, or perhaps it’s not easy for the user.  These are the little things I can tweak as we go along.  It makes for slow growth, but I’d prefer that to flash in the pan.  Although, I don’t mind the occasional spike to move things along like with what happened when I posted the UW logo comparison.

The slow strategy also means I have more free time to enjoy the warm Vancouver weather   Site overhauls can happen when it rains.

Dreaming in Code

I had been eying the copy of “Dreaming in Code” by Scott Rosenberg on the bookshelf of our sister company for a while and finally picked up a copy for one of my monthly programming book reads.  This book is actually pretty high level and really aimed more at non-programmers who want to run a software project, or at least understand a little better how we odd people work.

It chronicles the early days of the Chandler Project, an open source software project that was meant to change the way we think about personal information management by knocking down the silos between contacts, events, and the like.  I had never heard of the project, but it seemed to have some pretty ambitious goals back in 2003 when it was started.  It seems to have been sidelined by gmail and google calendar though.

The title comes from a quote from Jaron Lanier:

To be effective at any large software project, you have to become so committed to it.  You have to incorporate so much of it into your brain.  I used to dream in code at night when I was in the middle of some big project.

I don’t think I learned all that much from this.  It’s very well written (Scott is the co-founder of salon.com), but I think I would have gotten more out of it had I read it in undergrad or just before I started reading software engineering books. It actually reminded me a lot of the movie startup.com, which also followed a Web 1.0 era startup as it went from inception to bankruptcy in the bubble burst in 2000.  Although, the book would go on some useful tangents about software methodologies.  It was a nice overview of the field .

If you are a non-techie with a website idea (and judging by the number of people who pitch ideas like that to me, I know there are a lot of you!), this is really good place to start understanding how software is created.  The big theme seems to be about how creating software is a lot different than creating a bridge or a structure.  Some will claim that software is always late and buggy because of a lack of discipline amoung programmers, but I like the nice retort to this in the final chapter of the book: “Nobody ever tried to change the design of a bridge after it was already half-built!”.

Screening Co-op Resumes

I’ve officially gone through the co-op process as many times as an employer as I did as a student.  I have to say, in both cases it took me a while to get it right.  We’re finally figuring out a process that allows us to properly screen applicants without pissing them off.  (Something we definitely failed at last time)

Along the way, I’ve definitely started to believe that whole “employers only take 30 seconds to look at a resume bit”.  At least, I’ve come to believe that I can tell within 30 seconds if I want to bother spending a half hour interviewing someone.  After filtering for all the superstars, I always pick one to “take a chance” on and see if they have some hidden talent that hasn’t shown on the resume; I’ve yet to be impressed under such a situation.  Still, for anyone clever enough to google for me before they apply to one of the SEG jobs I post, let me give you a few pointers:

• Your work experience is the biggest thing I care about.  Never put your education up front.  Everyone in the system will have more/less the same education levels and EVERYONE did well on the high school math contests.
• Have a section for your side-projects and if you don’t have any side-projects, make some.  I really wish I had known this when I was in co-op.  CS is one of the few fields where you can make your own experience along the lines of “I think I’ll make a simple search engine today”.  This gives you something interesting to write about in whatever language you choose and shows interviewers that you are a hobbyist, and thus would probably be a better employee.  I can respect that you might not have time to do something in your spare time, but if you have no work experience, you need to generate something to differentiate yourself from the herd.
• I want to see something other than work, but I really don’t care what it is.  There are only a handful of campus organizations for which I remember the complicated org structure, so chances are that I won’t know the difference between “Cheese Club Attendee” and “President of the Cheese Club” and the difference of responsibilities.  Don’t spend too much time going into the details, since they really won’t be of interest given that the job won’t focus on it.
• I usually post a 140 character job post, so don’t think I’ll fault you for having a short resume.  A resume should make me want to spend 30 minutes screening you, not an itemized list of every insignificant thing you’ve done.  Heck, if your resume contents were also only 140 characters, I’d probably interview you just because you seem interesting.  Although, I’d probably only do that for the first person who did it, after that it would just be a gimmic.
• In order to cut back on the length of the resume, chances are you can get rid of most of the acronyms on your resume.  “EJB, J2EE, JVM, grails, Java, Spring, Struts 2, Hibernate, JSP, Java Servlets, jUnit” can more succinctly an less gratuitously be written as “Enterprise Java”.
• I’m probably going to Google you before I spend 30 minutes talking to you.  Don’t get too worried about this.  I abhore the idea that professionalism is another name for being stoic and boring.  But keep in mind that I want to work with people who are fun and intriguing, not those who lack situational awareness.

Keep in mind, I’m just one opinion out there, and mine are somewhat derivative of someone else’s.

Whichishotter.com Beta

whichishotter.com; take a look.  If it’s slow from too many people using at once, just try again later.  It’s not exactly hosted on a supercomputer.  Oh…we don’t take kindly to ie6 users.

I’ve been alluding to “side project” that I’ve been working on to learn the Ruby on Rails programming stack.   I know the best way to learn a programming language is to do a pet project in it, so I decided to make a rails-website.  This started back in January, but rugby cut into a lot of my non-work time, so I’m just wrapping it up now.

Anyway, I wanted to make something that would “solve a problem”.  Something that too many programming projects don’t really do.  I mean, Twitter doesn’t really solve a problem that previously existed.  The problem I wanted to solve was that of “which of these should I wear/buy/display”.  How many of us have been in a situation where we were decided between 2 things to wear and wanted an objective opinion?  My solution is whichishotter.com, which is now in beta(ie. bugs might still exist and there are still some features that need to be put in).

The idea is that a user can upload 2 photos and have people vote and make comments about which is photo is a “hotter” representation.  This is very different from “hot or not”, in that people aren’t assigning an arbitrary number to you, but rather, are helping you become more hot in your own right.

You don’t get an unlimited number of comparisons, but if you signup now, you get a freebie.  This is mostly to ensure people contribute as well as get advice.  People get more comparisons the more they vote and comment on other comparisons.  For now, it won’t write anything to your Facebook newsfeed, but I intend to have it write when a new comparison is posted once there is enough content on the page for this to start to grow.  Expect that after July 1.

I would really appreciate it if you signed up for it (again, it won’t put anything to your FB feed just yet), voted for a few things and posted something interesting of your own.  A good amount of content is already up, but I’d be interested in seeing how some of my non-geek friends would use this as well.  If you find a bug, all the better.

Open Vancouver: I Want to Believe

Vancouver Mulls Making Itself an “Open City”

I want this to happen.  I really buy into the whole idea of “Google for Government”.  I just doubt it will actually happen to any significant level without massive pressure.  Realistically, this sort of thing would have to happen at the local level before everything becomes open.  I suspect Vancouver would be one of the first places to implement this, but I’m cautious:

The motion calls for the city to endorse the principles of:

• Making data open and accessible by sharing as much of it as possible while respecting privacy and security concerns.
• Adopting open standards for data, documents, maps and other media.
• Placing open source software on equal footing with “commercial systems.”

The third one about “open source software on equal footing” doesn’t really concern me.  OSS tends to be really crumby for anything that requires an interface, so I don’t really care if municipal employees are using Open Office; I doubt that would change much.

However, the idea of making all city-collected data is something that would probably cause change at a very fundamental level if done properly.  I had first heard of the idea in Wikinomics, but I think the Obama campaign popularized the idea.  Making all publicly gather information publicly available would allow people like me to do all kinds of funky analysis on how the government works (or doesn’t).

I sincerely doubt that any government would actually have the stones to implement this in a way that would bring about the fundamental change.  Few recognize that something isn’t “open” until it’s freely available on the internet.  Making the bus-schedules available is a start, but until I can comb through every cent of the budget allocation down to the department level it won’t lead to better governance.

I think at a fundamental level, no politician really wants transparency while they’re in power.  Case in point, the expense scandal in the UK; all parties were had some guilty members.  It doesn’t server their interests to have every minute detail scrutinized by geeks with Python scripts, so they’ll avoid it as best they can or implement it in some toothless way.

I really hope Vancouver city council takes measures to do this.  Although, I think at higher levels of government it’s something that would require constutional-level enforcement.  As much as I’d love to see the words “RESTful API” in the Charter, that’s not likely.