I launched QBike in 1998 as a project database site, based on the data aggregation models that were in vogue at the time. I wrote spiders for many of the top online bike shops, including Bike Nashbar, Performance Bike, Jenson USA, and created a search portal for bikes and bike parts. Within months, word spread, and traffic started rolling in. QBike was one of the first ever CSE (comparison shopping engines), and it certainly made more sense of biking and triathlon product listings that Yahoo! could (remember, this was pre-Google). QBike was a project site, a diversion from work, but by 2000 I was being contacted by bike retailers who wanted to be added to our database and by advertisers who wanted to place banner ads. In 2001, I turned on the afterburners and really notched up QBike functionality and page listings to turn it into a full-on CSE, complete with category pages, brand pages, email alerts and more.
Fast forward to today, and QBike has many different page types and lots of content, including:
I started at CNET as a Senior Database Engineer, and on day one I dove right in optimizing SQL and tuning stored procedures in Sybase. Shortly after being hired, CNET turned its focus on a huge website project: designing and launching www.computers.com. I was the logical database designer, using ERwin to design 5 databases and working with the entire engineering team to coordinate front-end designs, back-end tools, and integration issues. The www.computers.com project spanned 9 months, during which time the database group grew from three to eight engineers, and I was named Lead Logical Database Engineer.
In 1999, I was promoted to Principal Database Engineer and oversaw all database design and integration issues, including the CNET San Francisco and New Jersey offices. In 2000, it was decided that the engineering department would divide into 4 units, corresponding to the main 4 branches of the CNET tree at that time. I was promoted to Director of Technology and led the www.news.com and investor.news.com engineers in San Francisco and Boulder, CO. By 2001, the Silicon Valley tech industry was clearly upside down, and in September of that year the engineering department was restructured due to downsizing. Since the Boulder office was shut down, my group of crack engineers got merged into other teams, and my position was eliminated. Although I was invited to apply for other positions at CNET, I already had my sight set on other career opportunities. And so concluded an amazing 4 years for me at CNET.
I started as a data analyst at APT in Pompano Beach, FL in 1993. APT was the first company in the U.S. to install POS hardware with video display at supermarkets across the U.S. At its peak, APT was partnered with 20+ supermarket chains and had deployments in 600+ stores. My job was to make sense of the transaction data from card swipes executed by our 6,000,000 card-based members at the time. I was fortunate to have an excellent mentor in Ken Godskind, who got me started in UNIX programming.
I quickly got well-versed with Sybase database technology and attended training in California for T-SQL and SQR (SQL Reports). Although SQR was powerful in its own right, I found myself turning to customized reporting that I dreamed up using C and shell scripts. I was truly interested in the data and grokking the meaning behind the millions of rows of transactions at my fingertips, and to do so, I wanted to turn to statistical concepts not available through SQL, stored procedures, or SQR.
At the time, I wasn't programming in perl, so I didn't have access to perl statistical libraries. SAS was not an option, and R and S weren't readily available at the time. So I developed my own small library of functions in C to deal with the data I so eagerly wanted to dissect. While dealing with these technical challenges, I was also forging relationships with the APT marketing team to make the best use of the results I was finding.
The CEO of APT, Robert Wientzen, took me under his wing and pressed me to attend DMA events, including the national conference. I was heavily influenced by the marketing paper released by the Coca Cola Corporation about customer lifetime value and began to formulate models based on the concepts I learned.
I would accompany marketing leads on trips to Proctor & Gamble in Ohio and to Phillip Morris in New York City to present research results and to discuss business partnerships. Presenting some of the eye-opening results I had discovered in studying and analyzing the APT consumer data was a highlight of my work with the company. One promotion in particular stands out in my mind: the cigarettes and dog food coupon promotion that we launched with Phillip Morris. I had found an abnormally high number of occurrences in which our card members were purchasing these two products in tandem, and a bundling coupon seemed to be a natural extension. We launched that coupon campaign a few months later and tracked its success amongst the hundreds of other database-driven coupon offers we formulated.
After 1 year with the company, I was promoted to Data Manager and took over the team of 2 data analysts. I hired a 3rd team member and continued working closely with the marketing team as well as the engineering department, whose main focus was on the hardware challenges. Though things were fantastic at APT, I had my sights set on consulting, so I left the company in mid-2007. APT is where I really built up my data chops and got my first taste of the power of analyzing data.
I started as a Senior Engineer with CTP in 1997. My first project was a choice between Burger King in Miami or Microsoft in Seattle, WA. I had joined CTP specifically to take on bigger challenges and to travel the country, so naturally, I opted for the Microsoft position. My role at Microsoft was to analyze nightly batch data to find discrepancies and to find any patterns which would give rise to alerts of some sort. It was a short-lived project lasting only 4 months.
I landed next at Pacbell Telesys in San Ramon, CA, as part of a CTP team of 15 engineers to build ORGIS, an order entry system for ISDN service. If there is one thing I to this day embrace from the CTP philosophy, it is their rapid prototyping and deployment philosophy. While other consulting companies bid on projects and propose many-year timelines for development, CTP advocates rapid development and launch so that the product the client wants becomes a reality - before the underlying product becomes obsolete. And this certainly was the case with IDSN, a precursor to DSL - we launched ORGIS in about 4 months in a basic version, then continued building it out for the next 1 year, by which time - you guessed it - IDSN was yesterday's technology. My role on the Pacbell ORGIS team was database engineer, and I was also heavily involved with the logical database design. I really wanted to take it over completely, but the timing wasn't right. I worked closely with the GUI developers to write Oracle stored procedures in PL-SQL, and I wrote all the backend reporting for management staff with regards to order data.
As the project was winding down, my next assignment was to be Omaha, Nebraska or London, England. I chose London - but the management team really needed me in Omaha. About this time, internet companies were taking off in the Bay area, and living in San Francisco, I wanted to stay. So I opted to switch gears and take my Sybase and Oracle training and jump into the internet world, landing at CNET.com.