Amazon is one of the few businesses that can justifiably say its use of data is world class. In the latest Management Today intelligence report, A Leader’s Guide to Data Science, Amazon Web Services enterprise strategist Ishit Vachhrajani showed us how the tech giant does it.
1) Work backwards
At Amazon, new ideas are fleshed out in memos that work backwards from a specific customer need, examining what assumptions would need to be proven correct for that idea to work. Ishit stresses the importance of how this process is conducted. “We don’t use qualitative objectives like ‘improve page loading times’. We say ‘we want to reduce load time from three seconds to one second’,” Vachhrajani says. That’s a subtle change, but with a quantifiable outcome you can look for quantifiable data to support or contradict your assumptions.
2) Speak a common language
You’ll only build data capability if your people understand what it involves, and what it can and can’t do. To build data literacy, Vachhrajani suggests running hackathons with mixed teams of functional business executives and data scientists. “We’ll give them publicly available datasets, like parking tickets or traffic violations, and give them a fun, everyday problem like ‘find me a restaurant in Manhattan on Friday night where I can have dinner without getting a parking ticket’. That makes it much more relatable, and it helps people understand how to find the datasets they need, how to visualise them and how to combine them to answer these questions,” he explains.