Some stock market analysts agreed that Elon Musk’s attempt to delay the deal to buy Twitter is just an excuse to negotiate a lower price. From a statistical point of view, the method he proposed for analyzing accounts on a sample of 100 is simply not capable of providing a representative result.
As you know, Elon Musk justified his decision to work with a sample of 100 accounts by saying that Twitter itself uses such a number during its internal audit. As explained ReutersThis has already caused displeasure among Twitter lawyers, since Musk’s statements about the counting method violated the non-disclosure agreement between him and the company.
Remember that the ultimate goal of the analysis is to determine the share of counterfeit and duplicate accounts as well as those that are used for distribution of advertising material. Initially gave Twitter that its share did not exceed 5% of the active part of the daily audience. After the resource CNBCin this context a number of problems are brewing.
First, the company said on Twitter, the last quarter of the event reported that it could get the wrong idea about the number of active users in the last three years. The fact that the operation in March 2019 added that allows a single user to switch between multiple accounts, and user activity statistics could all be considered unique. As a result, the accumulated error reaches 1.4 to 1.9 million users, which is not the case in reality.
Second, the expert community cautions that the sample of 100 accounts may not be representative for determining the proportion of fake accounts. Among BotSentinel, which analyzes the activity of more than 2.5 million English-speaking Twitter users, it is argued that the proportion of dubious accounts on the social network is actually between 10 and 15%. It is important to understand that this percentage can be increased in certain industries if we are talking about controversial issues such as politics, cryptocurrency, pandemics and climate change. At the same time, the Discuss, domestic animals or harmless hobbies of users much less saturated with bots.
Statisticians suggest that to get an objective picture of the sample must be much more than the 100 accounts that Team Elon Musk will use to identify bots and spam on the pages of Twitter. This whole approach seems reckless to experts, so it may not really be related to the reasons that forced Musk to put the deal on hold. Some sources now believe, based on the stock market reaction, that there is as much as a 60% chance of the deal failing under the current conditions.