Profiling the Google Revolving Door

Google and Policy Makers - a Big Happy Family

Contents

Team Members

Huda Alsahi, Matteo Azzi, Prem Borle, Valentina Dopona, Nick Forrester, Simone Griesser, Benjamin Hervit, Michele Invernizzi, Lorenzo Piazzoli, Ana Pop Stefanija, Jelly Luise Schuhmacher

Facilitators: Christo & Fieke (tacticaltech.org)

Introduction

The Google Transparency Project (GTP, http://googletransparencyproject.org/) is a research initiative that records movement of people between governmental organizations and Google, thereby “tracking” people’s employment transitions e.g. from governments to Google and from Google to governments. Mapping out the mobility of people involved in both enterprises supports the investigation of power relations. As the GTP facilitators put it:

Google has been among the most vocal advocates for transparency and openness in government, corporate America [and Europe] and society at large, but does not subject itself to the same level of transparency. For all its calls for others to open up, the company is highly opaque about its own operations and dealings with government. From its relationships with elected and appointed officials to its lobbying and public policy operations, Americans [and Europeans] know surprisingly little about how Google gets what it wants from their government. (http://googletransparencyproject.org/about-us )

In our project we primarily focused on the European data set the provided by GTP. This decision was made based on considerations regarding feasibility, the relevance due to the geographical ‘closeness’ of project members as well as the quality of data. When comparing the data sets for the US and Europe we were able to identify a larger number of Twitter handles in the European data set. Furthermore, the European context has been explored less than the US one.

Initial Data Sets

This report draws on a variety of data, which are detailed below:

• Google European Revolving Door data set which included 80 people who worked for Google and switched to positions in governmental organizations or have been working for governments and are now working for Google based on data scraped from LinkedIN.

• The Twitter accounts of above individuals where Twitter handle could be found and matched

• Google scrape with the name of above individuals as keywords
• Non-exhaustive LexisNexis recherche based on the names of above indivudlas

Research Questions

The aim of this project is to profile Google and its influence over European government institutions. Hence, individuals who have worked for both Google and government institutions of the European Union and its member countries are studied. In the research context, influence is operationalised as digital connections on Social Media platforms and social ties. Therefore, the ensuing research questions is as following:

---

What are the online connections between people who are (or have been) working for Google and for a governmental organization within the EU and to what extent can they be taken to infer real world ties?

---

In order to tackle this overall question we split up into two sub-groups to work on two different sets of questions, which could then be linked to each other. The first group dealt with digital connections online and particularly on LinkedIN and Twitter. The second group focussed on digital evidence of real-word ties such as newspaper articles and pictures.

Methodology

In this project we wanted to clarify the uncertainty and messiness of basically any research endeavor in the context of social media research and therefore focused on the significance of methodological and theoretical considerations as much as on the potential results of our analysis. In order to appreciate both of these pillars we decided to aim at suggesting a methodological approach to tackle our overall research question and illustrate that by providing a case study-based application of that methodology. When scrutinizing online connections and to what extent they can be taken as indicators for ‘real world’ [offline] ties requires us to in a first step qualitatively define who and what is at the center of attention in order to scrape data on online connections (e.g. following-follower networks, retweets, mentions, etc.). The process of making sense of these connections was also based on a more qualitative approach by looking at e.g. newspaper articles that co-mention certain actors or event reports that document the co-presence of certain actors in order to determine the existence of any ‘real world’ ties.

In order to organize the complex data and produce networks that allow us to visualize encompassed aspects and dimensions, it was evident that the research efforts could not only proceed on a quantitative level but must also be complemented and enriched by fine-grained qualitative investigation. We therefore follow a mixed-method approach. Below figure illustrates the employed methodology. The first row stands for the the initial data set that was the starting point of our investigation. The dataset is readily available on the RevolvingDoor website, thus there was no query design. From the data set we know that the data was scraped from LinkedIn. We employed a mix of qualitative sighting of the data and visulisation techniques to simplify the data to enable quick and easy interpretation of the complex data. After this first exploration of this data set, we went back to the stages of query design and data capture to complete, clean up and complement our initial data set. Next, we designed a Google query and scraped the websites and web suffixes theguardian.com, dailymail.co.uk, prweek.com, and gov.uk to scrape for the individuals identified in the dataset. This produced a list of links, which underwent a preliminary content analysis. Next, the LinkedIn profiles of prominent individuals of the dataset were qualitatively analysed on LinkedIn and subsequently on Twitter. This data was supplemented by a LexisNexis recherche with the keywords 'Google + European' and 'Google + UK'.

Findings & Discussion

The European Google revolving door data set is smaller in comparison to the US one. Figure one shows the employment moves of the individuals studied. Mainly individuals from UK government (purple) and European Union institutions (red) changed their jobs to work for Google (dark turquoise). While the individuals in the UK have worked in various governmental institutions, the individuals moving from the European Union to Google have mainly worked for the European Commission and European Parliament. However, 10 Downing Street has seen the most individuals leaving for Google in comparison to the other UK government institutions observed in the dataset. On the other hand, a smaller amount of people went from Google to European government institutions. The individuals leaving Google start to work for the European Commission and the European Parliament as well as UK government institutions. There are very few cases of individuals who worked for Google and European government institutions at the same time. These cases usually are individuals working for Google and advising a government institutions on a particular issue for a limited time period.

Below figure presents a consolidated view of figure one. Hence, the two main groups shifting to Google positions are from the UK and the European Union. Only a small amount of individuals from French or Polish government institutions moved to positions with Google and even a smaller amount of individuals from Italy, Spain, The Netherlands. Only 1 individual of the dataset moved from German government to Google. The same applies for Rumania. On the other hand, a smaller amount of Google employees moved to government positions.

A timeline of the individual employment moves between Google and European government institutions is depicted in the below picture as well as an illustration of the Revolving Door phenomenon. For example Elizabeth Haugland Dupuy has worked for a UK governmental institution, changed to Google and then back the a UK government institution. The figure replicates previous findings that mainly individuals from the UK and the European Union leave for Google (dark blue). Interestingly, few individuals moved from government institutions to Google in 2011. Exploring the factors that drove this shift is important and is likely to provide additional insights, but is unfortunately outside the scope of this time and resource constrained project.

As above findings indicate the Revolving Door phenomon seems to be most pronounced between the UK and the EU. We therefore conducted a qualitative analysis of some of the prominent individuals of the UK - EU Revolving Door and how they connect to two of the most prominent British politicians: David Cameron, the British Prime Minister at the time and Tony Blair.

Shared Education

Similarly we wanted to explore if the individuals in the initial data set went to the same universities and conducted a further analysis based on the intitial Revolving Door data set. The first figure provides an overall view of all universities that these individual went to. The following figure on the right-hand side provides a more focused view showing the universities that most people went to. These are unsurprisingly the University of Oxford, LSE or Harvard, but also other universities such as Sciences Po, Warsaw University or école national d'adminstration.

Shared Workplaces

Keeping the UK - EU focus, we investigated the connections on a more granular level by visualising what the shared workplaces are. Again the image on the left-hand side provides the complete analysis while the image on the right-hand side shows the most prominent.

Shared Skills

To further cement our findings, we also looked at shared skills. If individuals move back and forth between Google and the European governmental institutions, then it is likely that their jobs will always require similar skills. Below figure image details the skills the selected individuals in the initial data set had listed in their public LinkedIn profiles. Closed profiles were not included in the analysis.

Shared Interestes

However, the individuals in the EU revolving data set share not just a past or current employer, but potentially also interests and maybe even follow one another. For that purpose a hashtag and mention analysis of the actors’ public Twitter account was conducted. Below figure illustrates what the individuals talk about by means of hashtag analysis. They share hashtags, such as YouTube, Financial Times, EU, TTIP, or The Economist for example. The TTIP connections illustrates how the free trade negotiations connects the Google and governmental issue space. Similarly the Silicon Valley Open Doors Conference 2016 connects the European actor space with Google. The YouTube node is very big and worthwhile exploring. Given the limited time frame of this project, this was unfortunately outside the scope. The more pressing question is who influences these individuals on Twitter.

For that purpose we studies who these public Twitter accounts follow. The most followed account is Google’s. Given that we used the public Twitter handles of the individuals in the EU Google Revolving Door data set many of whom work for Google this seems logical. The employees follow their employer. Looking at the community clusters, we were able to identify the areas they are most interested in: they follow people/accounts from the industry, from policy (both people and institutions), institutions (both EU inst. and international) and finally UK related news, people and institutions.

Approaching the influencing question from the other end, we conducted an analysis who follows the EU Google Revolving Door individuals. As below figure illustrates again four clusters can be identified. The yellow UK-based clusters shows several prominent individuals, such as Joanna Shields, Nigel Huddleston and individuals tweeting under the handle hubmum and theobertram. The red coloured cluster entails US-based issues and individuals, such as Eric Schmidt and Demis Hassabis. The blue French cluster is smaller while the green E.U and international cluster is slightly bigger. In the green space we can see that individuals owning the handle angesteen, marcvanderham or radumagdin are prominent. Eric Schmidt for example has been on the business advisory board to the then Prime Minister David Cameroon. It might therefore be possible that the Google Revolving Door phenomenon includes selected few individuals.

To explore this notion, we analysed the extent to which these individuals follow each other. Below network graph shows that indeed there seems to be an online circle following each other. Eric Schmidt features very prominently in this circle. But other familiar names can be discerned, such as Demis Hassabis, Marc Vanderham, Joanna Shields, and Twitter accounts, such as theobertram, angesteen.

Shared Publicity

Previous findings illustrate shared education, workplaces, skills and interests. Governmental officials are public figures are more or less well-known and thus present in the media. This seems another fruitful perspective to take: what evidence of the Google Revolving Door phenomenon can be found in the media? For that purpose we conducted a LexisNexis query with the keywords ‘Google’ and ‘European’ for the time period between 1^st January 2010 and 7^th July 2016. Below table present the result for the search query. It features several names mentioned in previous analyses. For example, Eric Schmidt is no only a prominent person in the Google Revolving Door network on Twitter, but he was also on the business advisory board to David Cameron as the Google query showed. Similarly, Rachel Whetstone worked for the Conservative Home Secretary Michael Howard while Susan Hunter was a senior policy advisor to Tony Blair and then later became the head of public policy for Google UK. .Joanna Shields is also present in afore mentioned Twitter network. These findings provide additional support for the Google Revolving Door phenomenon in Europe.

Conclusions

Using the initial LinkedIn EU Google Revolving Door data set as a starting point, we found evidence of the Google Revolving Door phenomenon in Europe. Individuals who work for Google change jobs to work for European governmental institutions and vice versa. Employing data and methodological triangulation we substantiated this claim and show that the online Google Revolving Door phenomenon extends to offline real life social ties. The triangulated data sources included LinkedIn, Twitter, Google Query, and, LexisNexis. Methodologically a range of analytical techniques were combined, such as computational hashtag and mention analysis, network, content and visual analysis. Not only does this project substantiate previous findings of the EU Google Revolving door, extend it from online influence to offline influence, but also illustrates the value of data and methodological triangulation in substantiating findings.

Further research needs to be done into why many individuals left their governmental jobs to work for Google in 2011. Similarly, the role of YouTube as a connecting node in the shared interest space on Twitter between the EU Google Revolving Door individuals needs to be explored.