Why web developers have more to learn from Wall Street than Steve Jobs

In 2006, the biggest risk-taker on Wall Street looked like John Paulson. This certainly wasn’t based on Paulson’s past behavior. Paulson managed a middle-of-the pack hedge fund. He made careful, boring deals. He rode the bus and liked to ride his bicycle. In other words, he was the anti-Gordon Gekko.

The risk Paulson was taking? He was betting against the mortgage market using credit default swaps. Insuring $1 million in high-risk mortgages was dirt cheap – around $10,000. If all of the homeowners made their payments, Paulson would be out $10,000. If all of the homeowners defaulted, he would make the entire value of the bundle – $1 million.

For Paulson to make money, these high-risk borrowers needed to default. At the time, most analysts thought that a perfect storm of rising unemployment rates, higher interest rates, and poor local economic health was needed to trigger widespread defaults. However, Paulson’s extensive modeling showed widespread defaults required just one trigger: flatlining home prices. With home appreciation rising at 5 times the rate seen from 1975-2000, there was plenty of room to fall back.

We know that Paulson’s trades paid off. In fact, Paulson’s personal share of the profits ($4 billion) was almost as much as Google’s net income in 2007.

Only from the outside did Paulson’s actions look high risk. From the inside, it was a data-driven decision. It was anything but a gamble.

On the surface, a software business seems dramatically different than credit default swaps. However, when you’re delivering software via the web like us, you’re not really in the business of writing software. It’s 30% software, 70% micro-economy. It’s a world John Paulson would dream of – mountains of data, collected in real-time, with ample opportunities for testing theories.

We’ve written less code and grown faster the more we’ve studied our micro-economy. There are lots of different data sources – video recordings of new customer initial experiences, recorded via Clicktale, help us identify frustrating areas for new users. Tracking signup referrals helps us identify the types of referrals that work best. Analyzing differences between customers that stay with us for a long time and those that don’t helps us recognize warning signs in new accounts. Last but not least, we collect feedback via UserVoice to aggregate suggestions.

When we think about Scout as an economy first, software second, our decision process changes. Before, we’d think about a cool new feature we’d like to add to Scout, then build it. Now, we look at Scout usage data first. We look at trends we’d like to change. We develop features to try and change those trends.

John Paulson showed there’s plenty of money to be made when you correctly model an economy. It’s no different for our tiny micro-economy at Scout.