Watch: Data’s Impact on London’s Lyst.com
Quality of data, over quantity of data, is key to accuracy in e-commerce
At our very first Data Saloon, we gathered a few of London’s brightest minds in data science, along with a data-hungry crowd dressed like cowboys. It was a night to let loose and learn how utilising even the smallest bits of data can make a big impact on your business.
Teo Ruiz, former Lead Engineer at Lyst.com, spoke about that startup’s data strategy. Lyst is a fashion e-commerce marketplace that aggregates data from hundreds of retailers and makes it available to users in an organised and understandable way. Although the company isn’t dealing with big data yet, the site receives over 2 million unique visitors per month.
For Lyst, gathering quality data means keeping the product up-to-date, especially with regards to pricing. If the data is wrong, the company can’t be successful.
Working with 400 different retailers - each with their own technology and engineers - can get crazy, especially when partner retailers alter their code without warning. To solve this problem, Lyst uses “spiders”, which are small modules retailers place on their website to scrape the most important data coming through.
What happens when data is reported incorrectly? Lyst still emphasises that human knowledge and intuition is a key component to interacting with their computers.
Watch Teo explain how spiders, moderators, classifiers, and proactors work, and what they can do for your startup: