
About 10 months ago, an idea was roughed out on a piece of scratch paper, and a product/business venture was born. So we jumped off the proverbial start-up company cliff...
The business idea was brought to me by my close friend Dan Caplin. Dan spent the past 3 three years starting a fashion apparel tradeshow called Trafik, launched first in Atlanta then moved to Miami. He observed an opportunity to improve the ways boutique retailers order goods from their suppliers. We brainstormed about the idea through 2005, but didn't pursue the opportunity in earnest until early 2006 when Molham Aref, a friend / colleague/ and veteran of the retail software world, listened to our idea, saw the vision, and got on board.
Dan would continue to build the Trafik brand, and facilitate introductions to the apparel industry; Molham would serve as a sounding board on technology choices, design, execution and go to market strategy. I jumped at the opportunity to start with a blank slate and design a web based application from the ground up.
So we went to work.
I don't have the time or inclination to give a week by week accounting of how we transformed our unnamed idea on scrap paper into ExchangeFrame, the sexiest web based retail business tool around. (I suppose that’s a lot like saying that my child is the smartest kid around -- but you what? -- I don't care, so go suck it.) I do however, feel like sharing a few of the lessons I learned. That's not to say that what I learned was original, or hasn't been written about hundreds of time before, but it's my point of view on the subject.
People, People, People.
If I were wearing a Century 21 blazer, I'd be saying location, location, location. But I'm not. I'm saying that with the right team of people, things go right. As much as qualifying skills, we qualified understanding and enthusiasm in the vision of what we were building. You can't pay for enthusiasm (for that matter, you can't pay someone to fake it either). Smarts + enthusiasm = good things. A good developer can show code samples and answer your questions well. GREAT developers can write code with paper and pencil, on the fly.
Nimble and Lean.
Since the beginning, we work in teams of three. A Designer, a GUI engineer, and a server side engineer. Even now, as our product team has grown to fill our lab, teams are never bigger than 3 individuals, often having engineers swap places on teams. Occasionally, we'll work in teams of two, but three seems to be the magic number for speed and efficiency.
Open Source Software Is Your Friend.
Our entire application is built using open source software: HTML, CSS, JavaScript, Python, TurboGears, Linux, PostGresSQL, Lighttpd. ExchangeFrame is a heavy duty AJAX application... "AJAJ" actually, since we decided early on to use JSON in lieu of XML. We've also leveraged Prototype and Mochikit. We had a sophisticated prototype 45 days after our first hire that we used in user focus groups. At $0 license cost.
Engage Users Early On and Often.
We took a working prototype of ExchangeFrame to an apparel show early in development. We held invite only demonstrations and shared our vision with the very users we hoped would adopt ExchangeFrame down the road. This experience was invaluable. I don't know if there can be anything healthier for software development than to start an attitude of accountability with your user base as early in development as possible. Sure, sometimes a user doesn't know what they need until it's available, but more often than not; uncensored feedback helped us see weakness, foolish design, or gratuitous feature creep. Other times, something we thought was 'ok' turned out to be the feature that sparkled most. But you won't know if you don't ask.
You can't cut corners on equipment.
Spending cash is tough. It makes you pucker in places you don't talk about in polite conversation. It makes me sweat a little, on the bridge of my nose. However, spending money on the right equipment (even if expensive) will pay off. As evidence, I offer you two unsolicited endorsements and one pimp slap:
Strix Wireless Gear: We didn't want to rely on exhibition halls software, so we decided to put together a mobile network that we could deploy, through our own mobile server, anywhere, anytime. We spent a CHUNK of change on mobile setup, but you know what? Strix gear CRUSHED any WAN we came across. The Miami Beach Convention Center offered a 'wireless solution' at ridiculous prices. Thank goodness we didn't take them up on it. Their idea of a 'wireless solution' was two consumer grade Linksys wireless routers DUCT TAPED to two different columns at opposite ends of a monstrous hall housing 200 exhibitors and 500 buyers. Our ExchangeFrame_Local WAN signal was Hulk Hogan in a room full of whiny emo kids.
Rackspace Dedicated Hosting: Before endorsing Rackspace, let me give you a little background. We had been using a MediaTemple DV (Dedicated Virtual) offering to get started. Ug. They couldn't scale quickly, but worse was the total attitude I got from their customer service team. I called them in a fit of panic because I needed a faster, more powerful offering, and I was willing to pay for it -- they waited 48 hours to call back. And when they did call back, I was greeted with a snarky stoner voice who wanted to schedule a consultation in another few days. WTF?! I had such high hopes for MediaTemple, but in the end, our experience indicated that MediaTemple is what photogs like to call "Prosumer" grade. Not quite pro grade, not quite consumer grade.
On the other hand, when I started calling around for other options, Rackspace answered my IM to sales within 20 seconds and got on the phone with me within another 20 minutes. Misty Crayton, my sales rep, asked about our business, didn't get discouraged when we beat her up on price, and stayed off and on the phone with me from the early morning till 11:30PM. She was taking my calls on her cell phone. And when we reached a deal we could stomach (barely - our costs were going up by 6x) she had a proposal waiting in my inbox when I woke up the next morning. They under-promised and over-delivered, having both our app server and db server configured and ready for business 48 hours ahead of their estimate
ExchangeFrame went from being frustratingly slow and hard to optimize on MediaTemple, to snappy as hell and fun to use on Rackspace, virtually overnight. Yep, you get what you pay for kiddies. No doubt about it.
Rackspace = 1. MediaTemple = Pimp Slapped.
Anyway, enough for now. I've wanted to write about ExchangeFrame for months. Call it superstition, or just nerves, but I didn't want to write about it until we went live -- for real. So I'm sitting here, like a proud, nervous, sleep deprived web-app daddy -- watching users sign up, designing and debugging feature enhancements, scheduling demos, analyzing forecasts, managing budget, and looking for ways to get our message out to the right people.
ExchangeFrame is evolving the framework for retail trade. Yep, the future is now.