Fifty years ago, when the United States was actively involved in the space race with the Soviet Union, NASA launched an ambitious project: developing a spacecraft to land on the moon. But the biggest problem is that no one knows what the moon's surface looks like. Is it soft and powdery? Is it covered with needle-like crystals? Or is it made of large glacial boulders? The answers to the above questions will affect the weight of the machine, the materials it is made from, and how the machine should move. Some might think it's impossible to design a vehicle under such vague conditions, but doesn't it look exactly like what we do when developing new products or features? Now, let's look at the process. 1. Data collection First, we need to define what data is. It doesn't always mean "number".
It's just information that can be presented in digital form, but qualitative insights are also considered data. Whether we're writing SQL queries or talking to customers, both are doing a "data collection and analysis" thing. In data-driven companies telemarketing list like Facebook, Toutiao, etc., we can already see that when data scientists and UX researchers work together and look at the same problem from different perspectives, some interesting discoveries can often be made. For example, a study by Facebook and the University of Milan showed that the distance between any two Facebook users is not as far as the legendary 6 degrees, but 4.74 degrees.
"Facebook brings the relationship between people closer by 1.26 degrees" Data is also not equal to "user data". Even if our product doesn't exist yet, we can still use various data sources. There are marketing intelligence tools like Analysys, Aladdin applet index, etc. When we do some research work - such as competitor analysis or case study, we can try to analyze the value of our products without building anything where. Let's go back to the moon landing program, what data can we collect about this program? Although no one has ever been to the moon before, we all know the cost of a failed lunar mission is enormous, so even less data is better than nothing.