Mini Program
取消
Data
中文
EN
取消
  • Business tools

    • # Product Description

    # Turn on the function

    • Weixin Mini Program Administrators can go to The We Analytics website and click "Start We Analytics" to open the function. Once it is opened, they can go to the Experimental Tools to use.

    • Developers who use experimental tools need to have We Analytics membership privileges. We Analysis Administrators can modify and set permissions in Member Management.

    • Test Tools Portal: [We Analysis Platform ] - > [Test Tools] - > [Test Board].

    # What is an A / B experiment?

    A / B experimentation is a way to compare the effects of multiple versions of a product's features.

    The basic idea of an A / B experiment is:

    • The A / B experiment begins by dividing a portion of the users of the product -random- with a small traffic control risk - and dividing the users -__random]]__Show a trial version (control a single trial element), obtain user data for each trial group, and use____to find the best trial version for a given goal (digital quantification effect).

    Take a Weixin Mini Program experimental copy comparison as an example:

    • When a Weixin Mini Program wants to determine whether different copywriting has an impact on user conversion, and the product team is in doubt, it can use A / B experiments to make the decision.Start by randomly choosing a small percentage of the traffic from online and creating an A / B experiment in which group A users present copy 1 (subscribe to a message push) and group B users present copy 2 (recharge your energy and tell me). After a period on line, we use the statistical test of hypothesis (eliminating random noise) to determine which type of text to use.

    # Why do A / B experiments?

    • Many a little makes a mickle, scientific decision making

      • Growth Hacking: Internet products need to be data-driven, and experimental platforms are data-driven tools.

      • Less is more than more. Increase by 1 percent per day, and after 365 days the data effect will reach 37 + times the original.

    • People make mistakes, but math doesn't.

      • Google and Microsoft have similar statistics: Senior product managers are right about strategy effects 33% of the time.

    • Data trust

      • The improvement of strategy effect may be due to random fluctuations, and it is not possible to simply statistic sample data to get a conclusion, and it is necessary to get an accurate conclusion through rigorous A / B experiments.

    # Weixin Mini Program A / B experimental advantages and applicable scenarios

    • advantage

      • Provides the Weixin Mini Program official core metrics and supports custom metrics, as detailed in [Indicator Management - Indicator Creation - Measurement Overview].

      • Support population targeting and population analysis capabilities.

      • Support layered experimentation to improve traffic utilization.

      • Automate the production of experimental conclusions to facilitate decision-making.

      • The threshold for development is low. No SDK required, the base library (2.14.4 and above) supports experimental capabilities.

    • Applicable scenarios

      • Suitable for developers who need to operate in the Weixin Mini Program ecosystem, supporting Mini Program front-end UI, algorithm strategy and background function requirements.