Data analysis is a reliable product research method. There are many misunderstandings and we cannot be superstitious about it. In the end, we should ask people to make decisions.
During the second war, the British Air Force decided to reinforce the fuselage of the aircraft to reduce aircraft losses. Due to the limitations at that time, only a few parts on the aircraft can be reinforced with armor. They carefully observe, analyze, and collect statistics on the planes that return from the bombing task. Most of the bullet holes are found to be concentrated on the airplane wing; only a few are located in the cockpit. From the data point of view, it is the most cost-effective to reinforce the wing. however, the actual lack is the opposite. The cockpit is the most reinforced place, because almost no plane has been hit by the cockpit. "Voice" data is the best way to get it. However, if you do not consider the silence data, this data analysis is unreliable. soIn addition to the data results, you have to try to interpret the data, and the data is completely dependent on people..
- Regard silent users as intermediate states of support and opposition
Both websites a and B operate similar businesses and have stable user groups. They have all undergone similar website page revisions. After the revision, website a was not praised by users, but was scolded by many users. Website B Neither praised nor scolded by users. From the data point of view, the revision of website B should be relatively more successful, because no user expresses dissatisfaction. But this is not the case. Although website a has been scolded by many users, it indicates that many users still care about it. For website B, users do not care about it anymore.Website a refers to Facebook, and website B is Microsoft's live space..
- Data is the only criterion for decision-making.
Data analysis is generally considered to be a cost-effective approach, and it is not easy to make mistakes. For the management on behalf of the capital, compared to relying on human decision-making, data-dependent decisions seem more stable.
This decision may be effective in improving products from 0.5 to 0.8. because of an existing product, data is placed there. 50 access times out for 100 users. If this problem is solved, the performance is improved by 50%.
However, for new products ranging from 0 to 0.1, it is difficult to obtain the data, so we need to make great efforts to obtain the simulation data. it usually takes one week to figure out what a product should do for two hours. the simulation results are far from the final reality.
A/B test or a prototype system, first come out and then verify. In some cases, it is more effective than getting data first..
- Data is absolutely objective
In order to reduce internal consumption, data is often used for decision-making.Data itself is subjective, and completely objective data is not. Data acquisition methods, data interpretation methods, and statistical methods are all human decisions. It is not impossible to draw two opposite conclusions from a piece of data. Even if there is no subjective bias, it is limited by methods and vision..
In decision making, people are not data. Although there are so many uncertainties, there may also be disputes, complaints, and dare not take responsibility.