#  rules engine

**Published by:** [fixed star](https://paragraph.com/@fixed-star/)
**Published on:** 2022-10-28
**URL:** https://paragraph.com/@fixed-star/rules-engine

## Content

After the incentive voucher marketing system has been set up and operated for a period of time, a rules engine can be added to the system to achieve automated strategy growth. Generally, in the early stage, user groups and their corresponding incentive coupon rules can be delineated by increasing BI to analyze user data. After determining the user group and coupon rules, the coupons are distributed to target users through manual configuration of the system. Then, the difference of subsequent value contribution between the coupon issuing user group and the control user group is analyzed, and the incremental value brought by the coupon is calculated. After A few tests and iterations, we can settle on some clear rules, such as giving a group of A users an incentive voucher of a certain value when they perform a certain behavior. In this case, we can develop the rule to solidify, let the system automatically execute. Whenever the user's behavior meets certain conditions, the rule engine will be triggered to execute the corresponding coupon action. If the rules engine is a little more powerful, you can even let growth people add the coupon generation rules directly, without having to solidify each rule through development. But instead of creating an infinite glove, it's better to do it gradually, iterating to optimize the incentive marketing system. It is important to remember that you do not directly develop curing rules at the beginning. You must first manually analyze the data by growing BI and then solidify the rules after verifying them with multiple tests. Automate certain rules and manually explore uncertain ones.Only in this way can both efficiency and effectiveness be achieved. LTV computation is a step by step iterative optimization process. In the growth stage, a dedicated team is required to continuously track and optimize the calculation of LTV. Mature products have a relatively large number of users, and the data accumulation time is also relatively long, so the calculation of ELTV will be more accurate. For new products, because the number of users is relatively small, the impact of statistical bias will be greater when calculating LTV. Also, because of the relatively short period of time that ELTV has accumulated user data, its estimation accuracy will be relatively low. However, there are two key points to pay special attention to when calculating LTV/ELTV for both new and mature products. First, with the improvement of the product and operation system, the contribution of users in the product is gradually increased. However, when we calculate LTV/ELTV, we rely on historical data and assume that the future will be the same as the past, so the value of LTV/ELTV will be relatively underestimated. Therefore, when calculating LTV/ELTV, we need to predict the future trend and add an appropriate growth factor to make LTV/ELTV more representative of users' future contributions. The second is that many products have network effects. As network effects increase, users will spend more time in the ecosystem, generating more transactions. The typical one is the e-commerce platform. As more sellers or platforms offer more goods, users will spend more on the e-commerce platform. For the e-commerce platform with bilateral network effect, the increase of sellers and buyers and the positive feedback loop of mutual promotion will greatly improve the LTV and ALTV of users. Therefore, for products with network effects, we must have a judgment on the critical point of the formation of network effects. After crossing this critical point, the subsequent growth is often geometric.

## Publication Information

- [fixed star](https://paragraph.com/@fixed-star/): Publication homepage
- [All Posts](https://paragraph.com/@fixed-star/): More posts from this publication
- [RSS Feed](https://api.paragraph.com/blogs/rss/@fixed-star): Subscribe to updates