How product managers make data "speak" in products"

Last month, I took the plane at seven o'clock pm from Hangzhou to Beijing. As a result, at, the "flight standard" told me that the flight was delayed. One hour later, it again told me that the delay continued, until.

This app has been providing me with information, but this information does not give me more decisions. If the product manager of this app could imagine associating more information, could he tell me that there is a flight at next door, and he can immediately change the ticket to fly away? This is not a data product. It only makes the best use of information on the product, and immediately gives the product new value.

However, it was very poor. At that time, I didn't get such an information service. I waited from till midnight. The constant delay information provided by "Apsara stack standard" only made me more and more worried.

The difference between informatization and data is that informatization provides us with a reference, while data allows us to act directly. There is a huge gap between recommendations and direct actions.

Why cannot data value be implemented? Why are the data cases of a large number of enterprises difficult to highlight the value? It is important that the product manager does not understand the data. Many product managers are still at the previous stage of product production. They don't know how to use data to improve products by making products by feeling, and even less realize that data has become the core raw material for products. In the past it era, we used data simply and seldom extracted data to solve problems. Why should I emphasize the extracted data? Because if we want to make the data generate value and make the data analysis framework more closely fit to solve users' actual problems, we need to embed the data into the product or production process, in the last mile of data extraction, let the data "speak" in the product ".

How to make the data "speak "? In the past information age, we were best at guiding actions based on historical data statistics. For example, we will calculate the taxi tip for the past month at six o'clock P.M. on Friday, calculate the average bid, and then tell the user "We suggest tipping 5 yuan ". Most of the data we use is static data with a single angle.

Now, we prefer a panoramic and dynamic dataset. For example, we can obtain the Congestion Degree of different streets to calculate the driver's sensitivity to tip. we can collect nearby weather conditions, concert time data, and so on to predict a certain period of time, the amount of taxi tip that may be sold in a certain location. Such an algorithm uses more comprehensive big data to provide services through more dynamic environment data rather than historical statistics.

In the future, product managers need to know how to use data to add value. There are three key points: productization, digitization, and business vision. At present, many product managers are more concerned with productization, ignoring data.

How can we use data to add value? Lead the data.

Suppose I need to select a school for my daughter. If I have to wait for three months before the examination result is obtained, will the school be too bad? If I can determine whether the school is suitable for my daughter based on data computing, this is the data front-end. The key to many data values is data pre-creation, which allows more data to be embedded into the product to generate value.

Another case that is easier to understand is Google unmanned vehicles. Google's unmanned vehicles are using a data analysis framework to implement services. The premise of this service is that the quality, stability, and computing speed of data have been well-developed, so that the "Data guiding action" enters a completely automatic situation. Google engineers used thousands of models to support this data analysis framework to ensure that unmanned vehicles would not experience accidents while driving.

In contrast, many companies are still making decisions using statistical data. If we apply the data analysis framework to the company's business, we will find a new value.

Maybe you will ask, we have been saying "embedding data into the business", how should we embed data in the actual operation level?

In my team, I encountered the confusion that the solutions provided to me by the product team, data team, and Operation team are always difficult to connect together as a piece of sand. Simply put, it is difficult for the product team to have data concepts, and data teams rarely have product ideas. The operation team is not used to making decisions using data. But the difficulty is that, if there is no way to link these three teams together, where can we start with the value of data.

Whenever this happens, I will ask my team the following questions? Who's the problem? Do you need to solve it now? Can there be data to solve the problem? If all data is available, what is the solution?

Although these issues help to sort out ideas, it is still very difficult to cross between products, operations and data teams. My usual solution is to ask the team members about how many decisions they need to make every day in a specific production process (sometimes a decision-making process? Which decision points can be replaced by data? It is a very effective method to find the inspiration for "Data embedding" by combing decision points.

It may not take too long for product managers to discover the importance of data for products. Data must be integrated with products, otherwise the value of data is difficult to implement. Turning data into products is the biggest difficulty for product managers and the greatest opportunity and imagination for product managers.

How product managers make data "speak" in products"