当一个拥有较少粉丝数的直播账号开播时，很多人会觉得会没人看，但现实是在线观看人数很多。

When a live streaming account with fewer followers starts broadcasting, many people may think that no one will watch, but the reality is that there are many online viewers.

出现这个现象的原因就是大家混淆了直播权重和短视频账号权重，实际上二者为独立运行的两个系统，互不影响。

The reason for this phenomenon is that people confuse the weight of live streaming and short video accounts, but in fact, the two are independent systems that do not affect each other.

1、抖音直播间权重

1. Weight of Tiktok studio

这些数值综合起来就产生一个综合数值，也就是我们的直播间权重。而这个综合数值作为电商直播运营人员大家都是无法看到的，如果这个数值的具体情况被公布，作为电商直播运营人员完全可以靠技术手段把任何产品都打造成爆款直播间，这对于平台的良性运营是不利的，电商平台只有把流量的价值更高的转化才能体现出平台的高价值，如果把平台的计算数值的方式公布就会导致流量被手动操控在不良产品和直播间手中，而不是以产品优劣为导向分配流量，就像是考试前就得知了正确答案，这自然能获得好的成绩，但这对于其他考生是不公平的，自然这个考试的结果也就没什么意义了，而这个权重数值计算方式也就相当于是考试答察。

These values are combined to produce a composite value, which is our live broadcast room weight. However, as e-commerce live streaming operators, everyone cannot see this comprehensive value. If the specific situation of this value is announced, as e-commerce live streaming operators, they can rely on technical means to turn any product into a popular live streaming room. This is not conducive to the healthy operation of the platform. E-commerce platforms can only reflect their high value by converting the value of traffic to a higher value. If the calculation value of the platform is announced, it will lead to traffic being manually controlled in the hands of bad products and live streaming rooms, rather than being allocated traffic based on product quality. It is like knowing the correct answer before an exam, which can naturally achieve good results. However, this is unfair to other candidates, and the result of this exam is meaningless. The, And the calculation method of this weight value is also equivalent to exam answering.

2、相似标签和赛马机制

2. Similar tags and horse racing mechanisms

相似标签主要包括用户基础信息（内容偏好、购买偏好）、用户行为数据（人气数据和成交数据）和直播间标签（浅层标签是人气模型，深层标签是成交模型）。

Similar tags mainly include user basic information (content preferences, purchase preferences), user behavior data (popularity data and transaction data), and live room tags (shallow tags are popularity models, deep tags are transaction models).

赛马机制主要是看人气数据和成交数据，这两个数据代表直播间权重。

The horse racing mechanism mainly relies on popularity data and transaction data, which represent the weight of the live broadcast room.

这时的赛马机制是这样的：初始流量→赛马排名→优先推流。开播后的初始流量是由上一场直播的数据决定的，根据上周的直播数据和上一场直播的数据，决定了初始流量是多少的，这样基本是开播15分钟之内，至于第2波、3波、4波流量主要是根据人气数据和成交数据决定的，数据好就会继续推流，数据表现不好就不会推流。

The horse racing mechanism at this time is as follows: initial traffic → horse racing ranking → priority promotion of traffic. The initial traffic after broadcasting is determined by the data from the previous live broadcast. Based on last week's live broadcast data and the data from the previous live broadcast, the initial traffic is determined, which is basically within 15 minutes of broadcasting. As for the second, third, and fourth waves of traffic, they are mainly determined by popularity data and transaction data. If the data is good, the traffic will continue to be pushed, and if the data performance is poor, the traffic will not be pushed.

三、直播间自然流推流的决定因素1、直播间流速的背后逻辑。

3、 The determining factor of natural streaming in live streaming rooms is 1. The logic behind the flow rate in live streaming rooms.

停留时长定义内容质量:停留时长是决定直播间推流的一个重要因素(5分钟推流一次)但停留时长没有一个标准值，比如你新号开播的时候，同时间段跟你一起开播的竞争对手，他的停留时长只有19秒，而你有20秒，那么系统一样会给你推流。如果你是老号，粉丝沉淀到了一定程度后，你得区分开粉丝的停留时长跟新用户停留时长，这个数据在电商罗盘里可以看到。

Definition of dwell time and content quality: dwell time is an important factor in determining the streaming quality of a live broadcast room (once every 5 minutes), but there is no standard value for dwell time. For example, when your new account starts broadcasting, if a competitor who broadcasts with you during the same time period only stays for 19 seconds, while you have 20 seconds, the system will still stream you. If you are an old account and your fans have accumulated to a certain extent, you need to distinguish between the length of time your fans stay and the length of time new users stay. This data can be seen in the e-commerce compass.

成交密度决定内容效率:在同等的流量和曝光情况下，你的访客效率最低，为平台产生的利益最小，那么系统就会把流量给到更优质的直播间，所以成交密度也会影响你的推流速度。控流速的基本逻辑就是把你单位时间内的转化单量跟单位时间内的转化金额做好如果能把这两点做好，你的流速就相对可

Transaction density determines content efficiency: under the same traffic and exposure conditions, your visitor efficiency is the lowest and the benefits generated for the platform are the smallest. Therefore, the system will allocate traffic to higher quality live streaming rooms, and transaction density will also affect your streaming speed. The basic logic of controlling flow rate is to balance the number of conversion orders per unit time with the amount of conversion per unit time. If these two points can be done well, your flow rate will be relatively stable