Brand damage:
How Huawei was banned by (most of) the Five Eyes alliance
Tracking the change in preception of a corporate brand
Tracking the change in preception of a corporate brand
This research uses Twitter data to track how public perceptions of Huawei have changed from 2019-2023.
Huawei is a Chinese technology company that has developed a robust international brand – it is currently ranked 92nd in the Interbrand rankings. Huawei makes phones and laptops for consumers and is also a leading provider of ultra-fast 5G telecommunications equipment. Since 2018, Huawei has been under siege by Western governments. Critics complain that Huawei subsidizes its equipment prices to promote international adoption, that the company stole its critical IP from Western firms, and that its equipment is a national security threat. This research uses Twitter data to track how public perceptions of Huawei have changed from 2019-2023. Twitter was popular among politicians, journalists, and the broader public as a place to discuss current events, making it an ideal data source.
In 2018, Huawei was accused by the Australian government of selling compromised 5G telecommunications equipment that threatened national security. Australia banned Huawei equipment nationwide and informed their American counterparts about their concerns. US President Donald Trump followed with an American ban on Huawei and ZTE equipment in 2019. The White House encouraged the UK, Canada, and New Zealand (NZ) to follow suit.
The American mandate to ban Huawei left the UK, Canada, and NZ in a quandary. The leaders of these countries were not convinced that Huawei equipment posed a threat, nor did they want to risk their domestic economies by delaying the adoption of 5G technology. It also seems likely that none of these national leaders wanted to be seen as lapdogs to an American President who was deeply unpopular in their respective countries.
I used Twitter conversations to help me understand:
(a) What are the major topics regarding Huawei under discussion in each of the Five Eyes* countries?
(b) How do the events between the US and the allied countries eventually affect the decision to ban Huawei from the domestic market?
(c) How do domestic events affect an ally's decision to ban Huawei?
(d) How does the spread of pro-Huawei PR spread to people compared to information that is critical about the company?
For simplicity, this page focuses only on Canadian data, but the complete analysis also includes data from the UK and New Zealand. The following section contains a full background to the topic.
* The Five Eyes is short-hand for the United States, United Kingdom, Australia, Canada, and New Zealand – five nations with strong historical and cultural bonds that led to a national security information-sharing agreement.
My first step was identifying Twitter accounts in one of the Five Eyes countries. I used the location field that users can choose to fill out to determine their country. Users that did not have the location field filled out were excluded from the analysis. I wanted to be sure that a person was from a given country before I included the Tweet in that country’s collection.
I used natural language processing techniques to find major topics in Twitter discussions over time. This is detailed below.
I created a Bidirectional Encoder Representations from Transformers (BERT) topic model for each country. I used the BERTopic Python package developed by Maarten Grootendorst (https://maartengr.github.io/BERTopic/). This package uses c-TF-IDF to create topic clusters from the data.
You can see the Jupyter notebook of this early analysis below. For simplicity, I will focus on Canadian events.
Additional visualizations:
Bar chart of keywords from clusters
Hierarchical clustering results
Dynamic visualization of clusters
The initial analysis showed almost 20 topic clusters in the Canadian Tweets. I saved these clusters and summarized the ten most retweeted tweets in each topic cluster to an Excel file. I extracted the URLs that the Tweet linked to.
In almost all cases, a link to a newspaper article was being commented upon in the tweet. In each case, I read the source article(s) being discussed. This process gave me a good idea of the substantive topic being discussed in each cluster.
After I summarized each of the topic clusters, I verified my assessments by using the OpenAI API to get summaries of each topic. I found that our summaries of each topic cluster were similar.
Here are examples of the resulting Excel file I generated along with the OpenAI analysis of the topics discussed.
When I completed the initial analysis of tweets for the USA, UK, Canada, and New Zealand, several themes emerged. I found that I could describe the clusters of topics as part of a few broad themes. These were discussions related to:
1. Business concerns,
2. Political topics,
3. Security topics,
4. Technology topics,
5. Discussions about other Allies, and
6. More general discussions about Huawei PR (press releases or activities by the company.)
I developed a set of keywords to define each of these themes.
Figure 1- Simplified representation of words as embeddings. From https://docs.cohere.com/docs/text-embeddings
One of the most fascinating aspects of language models is word embeddings. Language models don’t deal with text but instead represent words as a vector of numbers. You can think of these numbers as representing a coordinate in some n-dimensional space. Language models will assign words a vector of numbers based on the context in which a word appears relative to other words. The numeric values aren’t interpretable by humans directly. The model will determine these values using its hidden layer weights.
But to give you the gist of the idea, consider Figure 1 below. Where would we put a cow in the two-dimensional embedding space – would we choose placement (a), (b), or (c)? Well, a cow is similar to a calf, so it makes sense to be close to that word, but it’s also an older version of the same animal, so it is analogically similar to what a dog is relative to a puppy. In this case, point (c) makes sense.
Again, the numeric values of the language models assigned to words (or sentences) are not interpretable by humans. But these values are still beneficial! We can use embeddings to see how similar two words are (i.e., to see if they occurred in similar contexts in the training data). We can also apply operations to word embeddings to find what words are similar. I use this technique to determine which tweets are similar to my inferred themes.
I extracted embeddings based on keywords related to each theme and compared these theme embeddings to the embeddings of all Tweets. This comparison allowed me to categorize whether any given tweet belonged to a general theme.
I verified my choice of themes by seeing how well these themes aligned with the topic clusters. Table 1 shows Canadian tweets and how they align with the themes. (Note that in this early version of the table.) Table 1 serves as a sanity check on the themes I have outlined. I found the correspondence between the topic subjects and themes reassuring.
The topics show a broad alignment with the themes I described. For example, topics 0, 1, and 2 all deal with Jean Charest – a Conservative leadership hopeful who previously consulted for Huawei. These topics discuss his politics and his business ties to the company. Ultimately, Charest's relationship to Huawei was an important reason why he lost the leadership of the Canadian Conservative Party. Similarly, topic 7 details Canadian telecommunications companies asking the government for bailouts if a ban is enacted. (The topic labeled -1 can be ignored – this was unclustered tweets.)
Table 1 - Proportion of tweets in cluster found to be similar to a theme. Note: Huawei PR omitted from table and topic clusters that were thought to be related to Huawei PR.
IMG File
Figure 2 is a stacked area chart visualization of these themes over time in Canada. The vertical axis is the number of tweets, and the horizontal axis is the date. Overlayed at each point in time are major events that occurred in the world or to an allied nation that may be relevant.
Figure 2 shows some interesting patterns. Until the arrest of Huawei CFO Meng Wanzhou in January 2019, most discussion about Huawei on Twitter was related to Huawei PR or general business discussion. Just before the arrest and afterward, there was extensive discussion about other topics concerning Huawei: politics, security, technology, and allies. This event marks a broad shift in the types of topics discussed. In essence, the arrest of Meng Wanzhou and the subsequent arrest of two innocent Canadians by the Chinese changed the way people thought about Huawei. It may have been the beginning of the end for the company in terms of they way the public thought about the organization.
Figure 2 - Stacked area chart of themes mentioned in Canadian Tweets 2018-2022
PDF version
Figure 3 shows the diffusion of the broad topics between the network of Twitter users. In this plot, a person is counted only the first time they are exposed to a topic. One can think of a diffusion plot of the spread of information as similar to the spread of a virus in a crowd. Once someone catches the virus, subsequent exposures are not counted.
This plot intends to capture the rate at which information spreads to people across a network. Figure 3 shows the spread of the broad topics. Early on, the spread of Huawei PR dominates (i.e., discussions related to Huawei products and generally not critical to the company). However, after Canada placed Huawei's CFO under arrest at the behest of the Americans, there was a jump in conversations about Huawei that were critical of the company. It's clear from Figure 3 that the discussion of Huawei rapidly spread throughout the network, outpacing more benign pro-Huawei PR and product information.
Figure 3: Diffusion of topics in Cananda, 2019-2023
The data visualizations in Figures 2 and 3 show how perceptions about Huawei changed over time. Figure 2 shows that there are many Tweets in Canada concerning the company. The number of Tweets concerning Huawei is mainly dominated by discussions related to Huawei products and related events. The tweets sharply increased after Huawei’s CFO was arrested in Vancouver. The arrest led to China arresting two Canadians in retaliation, leading to regular media coverage. China’s bullying style of diplomacy subsequently caused a sharp increase in critical coverage of Huawei’s business, its relationship with the Chinese government, the security of its products, and its poor relationship with other allied nations.
Figure 3 shows how conversations more critical about Huawei diffused further in the social network. There was a tipping point when Huawei’s CFO was arrested in Canada, and China took two Canadians hostage. This event caused a spike in critical discussions about Huawei that soon overtook the spread of Huawei’s PR.
Yet, despite the concerns of ordinary Canadians and the country’s allies, the Canadian government refused to ban Hauwei’s equipment until late 2022. By that time, discussions about Huawei had become relatively quiet. The two Canadians held hostage were returned the previous year, and there were no international events of note. So, what drove the Canadian government to ban Huawei?
In late 2022, the Conservative Party of Canada was in the midst of its leadership race. The establishment candidate was Jean Charest, a former Conservative Party leader and Quebec Premier. His more populist opponent, Pierre Poliviere, attacked Charest based on his past work consulting for Huawei. He argued that Charest put Huawei’s interests above those of Canadians. After years of reports about the two Canadian hostages, Poliviere’s message struck a chord. Charest lost the leadership race. Liberal Prime Minister Justin Trudeau saw the weakness of his position and, shortly after that, banned Huawei’s 5G equipment in the country.