Just the other week, my co-worker Todd Singleton and I competed in the TechCrunch Disrupt hackathon in New York City. Mashery was a sponsor of the event and we were there to spread the good word about the Mashery API Network. Prior to the event, Todd and I came up with a few ideas but never came up with any solid plan. So we winged it!

We recruited a couple of young graphic designers on-site because neither Todd or myself are “front-end” guys. The idea that we decided to pursue was building a replacement for “Googling” individuals by utilizing social media profiles and related meta-data.

In the spirit of tech buzzwords, we pivoted our project several times in a matter of hours. We tasked our young graphic designers, Kate and David, with building a UI to nameless product that had no clear value proposition. All we knew was that there was a multitude of data about people out in the wild, and we wanted to build something that would replace “Googling.” And then came Harold from Hunch.

Harold Cooper, an engineer from Hunch, was giving an API workshop talk about their API. Hunch is a product and technology company that is building a taste graph — a way to make and discover recommendations based on users’ tastes. Harold was discussing Hunch’s get-recommendations API call which provides an incredible amount of retail, travel and other recommendations based only on an individual’s Twitter handle or Facebook ID – and they don’t even have to be a Hunch user. Harold also mentioned that Hunch was awarding $1,000 to the best hack that utilized the Hunch API. Wow!

We had about 18 hours total to build something, and what we ended up submitting (with just minutes to spare) was Whodata, pronounced  like, “Who Dat? Ahh..”  A little Detroit flavor.  Over 100 teams presented their apps in front of a panel of judges including Jeff Clavier, Christopher Poole (moot) and Bradley Horowitz. With nearly a full days of hacking and no sleep, we were given only 60 seconds for our demo pitch. Despite the lack of sleep and turbo demo, it all worked out.

Whodata was built with PHP, Java, Perl and JavaScript. It uses the Qwerly API to get social media profile information based on Twitter name. Next, it uses Klout API to pull down social influence data. And lastly, it uses the Hunch API for taste graph/recommendation data.

Whodata was awarded with the “Best Data Integration” by Palentir Technologies who sponsored the hackathon. The prize was $1,000 of Amazon Web Services. We were excited and flattered to win. The hackathon was incredibly fun and very well organized. Big thank you to TechCrunch for holding an incredibly fun and well organized hackathon. And of course, a big thanks to Palentir for the AWS credit.

Shout out to team Whodata – Kate Proulx (@proulxsie) and Dave Caputo (@davecaputo), two young roller derbying graphic designers from New York.

As for the $1,000 AWS credit – I’ll be using that to spin up instances and provide environments for people participating at future Mashery sponsored hackathons coming near you!

About a year ago, I told a friend of mine that I’d like to take the work I’ve done for seller real-time SMS notifications on eBay, and bring it over to Etsy.  Etsy is a marketplace for handmade goods.  I tabled the idea until last month when Etsy announced their next generation API (v2).  I figured it would be a fun project to hack and give me an opportunity to pick up some new skills.  For those that know me well, I’m on a mission to learn new stuff, and my objectives of the project were to:

• get an actual artist to make the art
• layout the the Web app with nothing but <div> tags and CSS
• use more jQuery and less hacky JavaScript
• OAuth using a generic OAuth library (ended up using PHP)
• do it in less than three days

Mission accomplished (sans less hacky JS).  It’s EtsyText, version one and free.  Sell an item on Etsy and you’ll instantly receive a text message on your mobile phone.  There are hundreds of Etsy community members using it, and it’s just fantastic to hear their enthusiasm and feedback in both the discussion forums and convos (internal Etsy member messages).  For the Etsy sellers, this helps them stay informed about their sales immediately, without having to be near their computer.  There are many more EtsyText features to come as Etsy continues to enhance their API offerings.

Unless you’ve been living in a cave, you’ve probably heard about Java support in GAE.  Getting GAE/Java up and running couldn’t be easier because it’s packaged as an Eclipse plug-in.  When you’re setting it up, you also get GWT bundled right in.  And deploying your first Hello World application?  Easy as microwave popcorn.  Ola Bini (ThoughtWorks) of JRuby notoriety gave a session at Google IO on JRuby on GAE, which means that Rails fanatics should be happy to know that their apps should deploy on GAE with little modification.  Ola also built a demo of (probably) the first 3rd-party Google Wave robot extension (with early access to Wave sandbox)… written in Ioke, a folding programming language (designed by Ola) that runs on top of the JVM, deployed on GAE.  A demo without the fancy IDEs.. just Emacs and deployed at the CLI.

Google Wave was announced at the keynote presentation (HQ YouTube video) on Thursday morning.  I walked in a few minutes late, and they were showing real-time chat between two different client computers.  Imagine onKeyPress event triggers, as opposed to waiting to hit Send.  That’s what I saw on the big screens.  There was a huge applause, which puzzled me, because I used to use this all of the time back in the days of ICQ chat.  But as the demo went on, it became far more impressive… far more compelling from a product/service development standpoint.  Google built a set of APIs that allow us to construct real-time collaborative apps — built entirely on GWT.  Wave keeps a history/timeline that tracks revisions and can be played back by the user to see the state of the collaborative work (whether it be a wiki, blog post, map, anything really) at any given point in time.  And because it’s based on GWT (AJAX), the same code can be deployed on mobile devices, such as Android and iPhone.

Google also made the conference fun.  Kicking it off by giving away a free HTC Magic… I felt like a kid (but from a movie) under the Christmas tree (in real life, I usually received an envelope with cash).  Wednesday (day 1) night, there was a nice “after hours” party (21 and up), with lots of drinks, food, authentic party-rousing DJ (Mike Relm), Maker Faire gadgeteers, and video games.  I played Missile Command and Centipede, and pwnd.  There was also a bar-code scavenger hunt with our new Android phones.  I haven’t heard what the first place winner received.  I shared a lunch table with the winner, Abraham, a few hours before the hunt ended, and he was guessing the prize was a t-shirt, whereas I imagined it was a Google-branded hybrid SUV that gets 130 MPG.

To sum it up, I thought it was a great conference.  I only wish I could have attended more sessions.  I look forward to playing with Google Wave, deploying actual production JRuby apps on GAE, tooling around with my new Android phone, and hacking my first Android app.  It was  also very affordable ($50 academic, $300/early-bird).   I’ll definitely be back next year.

The last part, startup metrics, is the topic that Dave covered in his talk.  He started giving this talk a couple of years ago.. and he calls it: Startup Metrics for Pirates: AARRR!! (Startup Metrics for Product Marketing & Product Management). You can see his SlideShare presentation here.  AARRR is an acronym that stands for:

Acquisition: users come to Web site from various channels
Activation: users enjoy 1st visit and have a “happy” experience
Retention: users come back , visit site multiple times
Referral: users like product enough to refer others
Revenue: users conduct some monetization behavior

I’m not going to regurgitate his entire presentation in this article.  Dave presents a lot of interesting ideas and tools that companies should use, but I’m here to get specific about Measuring What’s in the Middle.  As Dave puts it, that’s the good stuff, referring to what’s in the middle, Activation, Retention and Referral.

Caveman Metrics – Ugh.

If you’re an entrepreneur or have ever been involved in an online marketing campaign, one of the first questions you ask yourself (or get asked) is, How do drive millions of people to our site? That’s acquisition.  There are many ways to obtain traffic, from paid advertising campaigns, promotional offers, getting online press, as well as a host of sinister methods.  Recently, a strong emphasis has been placed on building traffic through social networks, such as Twitter, Facebook and MySpace.  So, lets just imagine you plan on doing all of these above — paid advertising, SEO for organic result boosts on search results, blogging, Twittering, promotional giveaways, the whole nine yards.

What do you do next?  Naturally, you log into Google Analytics, Web Trends, or look at your Webalyzer reports and say, “Wow, look! Last week we had 3,000 visits to our site, and 2,900 were uniques!” You take a look at your subscription/signup stats, and say, “Whoah, dang. In that same period we had only 12 account registrations and 3 inquiries.”

With out-of-the-box Web measurement tools, that’s pretty much what you get.  Search keyword referrals (how people found your site), ad campaign CTR (click-thru rate), visitors/uniques, day/time graphs, and so on.  They look really cool, but what are they really telling you that’s valuable?  This is what I call caveman metrics.  You take a cursory head-count of visitors coming in, and take a look at the bottom line of how many signed up.

Pirate Metrics – Aarrr!

Be very specific about what you’re measuring.  Within the activation, retention and referral categories, define goals for which you can measure conversion (success).  For instance, create a metric for a “happy first visit” — such as spending X time on the site, and performing Y actions.  Another metric could be the measure of sign-ups for a specific feature within the site that you’re trying to push the user to, such as signing up for a free widget, newsletter, etc.  Only with specific goals and measures can you tweak and re-measure with any level of confidence that you’re steering user behavior in the right direction.

Revolving Door or Turn-Style

Depending on how well you tune your internal Web strategy, your site can resemble a revolving door or a turn-style.  A revolving door meaning that people walk into the roating glass cylinder and exit before they even step into your establishment.  Or your Web site can resemble a series of turn-styles where you strategically lead them into actions for which they pass through (and are counted) with the expectation that they’re moving forward and getting some value/benefit out of the experience.

When you visit a Web site (even one that you traverse often), take notice of how this site leads you to actions.  That’s not by accident.  It was designed to lead you there.  When you visit a new site that is a complete turn off, or just doesn’t engage you into action, also take notice.  Those are every day examples of turn-style vs. revolving door experiences.

Conclusion: The Fundamentals Apply

There really is no short cut, no silver bullet.  Tools exist that can help you measure more effectively and efficiently, but these tools do not know your business.. they do not know your goals.  Only you know what those goals are.   The point of this article: go beyond basic metrics (i.e. WebTrends, Google Analyitics, Webalyzer, etc.) because alone, they do not tell you anything about your Web site strategy.  Get specific about the user segments that are visiting your site, and what you want each of these segments to do so that they’re happy, they come back for more, and they refer their peers to your Web site.  Keep tuning your site strategy so that you continually increase those results, and reap the plunder… aarrr!

Of course, I’m not doing Dave’s presentation/content justice.  He speaks at meetups/conferences often, and I highly recommend seeing him because his lessons are easy to digest and immediately applicable.  Here’s a video from a couple years back of Dave McClure giving a turbo presentation.  The content of his presentations today is very similar, which in my opinion, speaks volumes.

I found a total of five design firms that I liked — all small houses, nothing at the caliber of Fry or Organic.  Those with phone numbers available, I called.  No humans answered, so I left messages.  And all five I wrote e-mails to (or filled out their contact forms), explaining my interest in having them do some simple design and UI work for a “small but professional Web-facing application.”  On the technical side, I told them it was “basic HTML/CSS and static graphics work — no RIA (i.e. Flash, Flex, etc.) or heavy AJAX/JavaScripting”.

So, you’re well aware of the macro economic downturn we are experiencing.. right?  Here it is, ripe, juicy, low-hanging fruit for the taking.  And how many of the five hands are grabbing for it?  Zero.  Absolutely not a single e-mail or phone call back.  I was shocked, and really disappointed.   I actually pass by one of these firms a couple days a week, and I see people there, I assume, working.  Maybe someone just dropped the ball inadvertantly, in all five cases, right?

Several days later, I find contact e-mail addresses either by domain name record, or somewhere else on their Web sites (where possible, one only had a Web contact form — no e-mail, no phone number).  At this point, it’s purely for research: How many of these people/companies will respond with anything? Mind you, we received no bounce-backs from dead e-mail addresses.  The second wave of inquiries go out..

Absolutely nothing.  I am, to this day, floored.  I have been tempted to knock on the door of the firm that I called, left a message, and wrote two e-mails to, and speak directly to the first person at the door: Hey, any chance you’d like some business? But, no way.  As much as I would love to help out a local small business, if you can’t even return a phone call or e-mail, just imagine the teeth pulling required to get work done.  A molar for a first draft creative?  A root canal for a delivery on the UI templates?  I think not, especially when I would be the one feeling the pain of the extractions.

The moral (or molar) of the story is this:  He or she who “shows up” wins, because the competition sure isn’t. When an opportunity presents itself,  find out everything about it, and if it fits, own it. But step one is reply to your inquiries, and do it fast.

Davey Waterson from Apatana (of Ruby/Rails IDE fame) presented on ActiveRecord.js. Very cool and easy way to hook into databases in a cross platform manner in JS. Getting and setting persistent data across JS to a client or server is quick and dirty with ActiveRecord.js. It’s cross platform, cross browser. Easily connect to SQLite or even MySQL using Jaxer, Apatana’a Server-Side JavaScript engine. I use Aptana IDE (an Eclipse derivative) for Rails development in Linux, as a side note, and it’s quite sharp. Man alive, that’s three versions of Eclipse installed on my laptop.

Last talk was given by Tom Robinson of 280 North. He spoke about his port of Rack (minimal interface between webservers supporting Ruby and Ruby frameworks) into JavaScript, called Jack (on Github). I’ll let you check out the Github page for a better explanation of what Jack does for you, and spare you the cut-and-paste.

I liked having the meetup @ Google because there’s already plenty of people on-site to attend, and that made out of some quality Q&A sessions. Okay, I like the smoothies in the lobby, and snacks dispersed around the office too.