3.1. Getting DNI’s Numbers
According to Google, between 2016 and 2017 there were 209 media projects funded by the DNI initiative. The United Kingdom (33) and Germany (28), followed by France (15), Spain (15) and Italy (14) were the countries with the most awarded projects. As Google claims in its reports, by the end of 2020, Germany (71), the United Kingdom (69) and France (61) were by far the great beneficiaries of this initiative. However, some numbers do not add up. For instance, in the 2017 report Google indicates that Austria had four different projects awarded. Despite explaining both Round 1 and 2 in detail for the first report, Google left out the two initiatives that participated in R1: Nimeh & Partners with a prototype project for monetization, and Der Standard (Standard Verlagsgesellschaft), with a medium project to identify negative emotions in online discourse.
Table 2 shows the correct number of funded initiatives by round and country that we have compiled through the investigation.
The total number of projects presented by Google is 662. According to our research, there appear to be at least 17 missing projects from our database, which shows 646 as the total number of documented successful proposals. After investigating in detail all the names of the organizations involved, we realized that companies such as the Slovenian Event Registry or Honire Inovativne Informacijske Tehnologije, and others like the British Project for Study of the 21st Century or even Sky News, are nowhere to be found in the DNI website, although they do appear as part of the recipients in at least one of the three published reports. The missing data will be discussed in the following section.
As our evidence shows, the DNI Fund has continuously increased the number of European participants in each country; Cyprus (1) is the only exception to this statement as the Cyprus News Agency was their only accepted candidate in 2016–2017, and it has received no more money since the first initial round that year (although it appears in every report published by Google). Other countries, despite having smaller projects, have brought their media companies onto Google’s radar. Malta (1), for example, appears for the first time in 2020 with the participation of Mediatoday Co. Limited, a media company founded in 1999. Likewise, the two outlets in Estonia (AS Eesti Meedia, the biggest media group in the Baltic countries, and Õhtuleht Kirjastus AS, a media association) are present in one of the most recent rounds (5). Finally, it is worth mentioning that Lithuania (5), Luxembourg (2), Slovakia (4) and Slovenia (6) entered the DNI Fund in 2018 and have consistently increased their participation since then.
3.2. Footing the Bill
Although the reports only disclose information about the total amount awarded, we have managed to establish the full quantities involved in the three-year period based on the country destination. As it appears in Google’s report published in October 2017, the fund in 2016 was established in two different funding rounds.
Table 3 shows both rounds, where we can see how countries such as Lithuania and Latvia participated only in the first one, while Bulgaria, Croatia, Cyprus, and Slovakia appear in the second.
The 2018 report also had a funding breakdown, although this time it does not separate the amount received in each round (Rounds 3 and 4 comprise all the projects awarded in 2017). In
Table 4, we see a comparison between 2016 and 2017 in terms of funding. The difference between those years is shown in the third column.
Google’s final report was published in October 2020, and it summarized the total amount of money that was awarded by the DNI Fund Initiative. This is the first time that a disclaimer appears near the funding, which states: “6% of the DNI Fund went to knowledge sharing, reporting and overheads” [
13] (p. 5). There are no further specifications as to what those activities involve, or whether the 6% is deducted by Google from the award. We also do not know if the recipient is obligated to carry out the dissemination activities with at least 6% of the final budget.
Table 5 shows the awarded amounts by round throughout the whole DNI Fund initiative. Rounds 5 and 6, which took place in 2018 and 2019, respectively, have been calculated by deducting the total amount of the report for Rounds 3 and 4 from the total amount announced on the last report.
With respect to population,
Table 6 X divides the total amount awarded to each country by its population, thereby obtaining a ratio that aims to simulate GDP in terms of media funding per inhabitant. Despite obtaining fewer funds through the DNI, Bulgaria and Romania are the two countries with the highest ratio, followed by Latvia and the Czech Republic.
3.3. Targeting Media Issues
The first DNI report thoroughly describes the purpose of the fund and goes into details about the nature of the projects and the most popular topics among successful applications. Those topics were clustered into seven categories: “Intelligence, Data Management and Workflow, Interface and Discovery, Next Journalism, Social and Community, Business Models, Distribution and Circulation” [
34].
Table 7 shows the number of projects funded in each topic group in 2016 during Round 1 and Round 2.
However, these categories disappear once we reach the report of 2018. Instead, Google breaks down the funding into four new categories: Battling Misinformation, Telling Local Stories, Boosting Digital Revenues, and Exploring New Technologies. The official document does not specify the nature of this change, so there is no easy way of relating the old and new brackets. Furthermore, the report does not give any more explanation regarding the distribution of the projects within the four different categories. There was no online information referring to this matter until the third and last report was published. According to Google, once the fund reached Round 6 more than half of the projects (52%) were aiming to explore new technologies, while concern about disinformation was a key aspect for 21% of the ventures [
13]. Finally, 15% of the proposals focused on local journalism, and 12% were aiming to boost digital revenues [
13]. However, we do not know if those percentages relate specifically to Round 6 or if they imply a full analysis of the whole DNI Fund.
Regarding the profile of the companies that apply for funding, we can affirm that most of them are identified as “commercial media” (381). Universities and research centres (17) and publicly funded media (16) barely represent the 5% of the total of funded initiatives, followed by non-profit media (42) and individual participants (41). The latter have been counted as “independent actors” despite presenting or creating projects that involve or implicate existing media. Finally, in the category of “other organizations” (145) we have gathered different institutions that cannot be classified as media organizations, such as the Bulgarian company Damocles Analytics, which specializes in AI-driven political insights. The French private meteorological consultancy Meteo Consult also belongs to this category, as well as the German micropayment solution PayPeanuts or the Polish E-Kiosk.
Figure 1 shows the different kinds of companies that have been funded by the DNI.
3.4. Guess Who?
The last pages of every report are dedicated to all the funded organisations, classified by country. However, there have been some fluctuations between the three documents, and several media groups or companies have changed names or even disappeared. To give an example, Futurezone, an Austrian hightech, is presented at the end of the 2017 report as the developer of FutureBot, a framework for personalized news distribution via messenger. Yet, the Google News Initiative website shows Telekurier Online Medien as the awarded company, and the digital report establishes both companies as two very different initiatives. Belgium and its media group De Persgroep appear several times on the reports as well, with different names such as “De Persgroep Publishing” in 2017, or “De Persgroep Advertising and Mediahuis Connect” in 2018, as if they were separate entities or belonged to dissimilar subsidiaries of the company.
Secondly, it seems that Google easily overlooks the existence of companies that belong to the same parent corporation, such as the Czech Regi Radio Music, which belongs to the Lagardere Active group. Rasmus Nielsen Holdings presents a similar issue: PoliWatchBot, a medium project funded in Round 3, appears on the report as a proposal by the Danish newspaper Altinget. However, this company, as well as Mandag Morgen, which presented a project in Round 6, are not always identified as part of the media holding group. By appearing as independent companies, the list of recipients tends to be longer, and this is likely to mislead the reader.
The Spanish case also seems confusing: Grupo Prisa (Promotora de Informaciones S.A.), for instance, is a large media group with several subsidiaries around the country. In the reports, Google states that the Spanish media holding obtained funding for Rounds 1, 3, 4 and 6. However, Ediciones El País S.L. obtained money for a large project in 2016 (Round 1) regarding HD video, and another large project in Round 6 for a branded content inventory optimizer. El País belongs to Grupo Prisa, which Google never explains in any of the reports, nor publishes on the website. The same occurs with Prisa Inn (Round 3) or Prisa Radio (Rounds 1 and 6).
More examples include the 2020 report, where Sweden’s media appears, followed by 13 different companies funded by the DNI. Nonetheless, at least five of them belong to the Bonnier Group, such as Papereed, which presented in Round 5 a solution for automated and data journalism based on audio, or Kvällstidningen Expressen and its prototype in Round 3 (although this appears online as part of the Bonnier Group). Other initiatives like the ones brought by Bonnier Business Press or Bonnier News seem simpler to identify thanks to the name of the parent company already being present in the subsidiary.
Finally, our analysis includes individual proposals as part of the total projects accounted for. Sometimes they appear as duplicates, such as the German One Eighty, created by Victoria Schneider and awarded in Round 1. We have managed to create a full list of recipients (40) that were awarded with individual grants for prototype projects, as shown in
Appendix A. All Rounds had at least four projects that benefited individuals: Round 1 awarded projects from the Czech Republic, France, Germany, Latvia and UK; Round 2 had proposals granted from Finland, Germany, Italy, the Netherlands, Portugal and the UK; prototypes participating in Round 3 were from the Czech Republic, Germany, Ireland, Latvia and Slovakia; Round 4 had projects from Cyprus, Germany, Hungary, Ireland, Italy and the UK; participants in Round 5 were from Austria, the Czech Republic, Italy, Poland, Romania and the UK; and Round 6 granted funding to Greek, Portuguese, Spanish and British projects.
3.5. Blurred Data
Gathering information to complete this research became a rather difficult task in terms of transparency. Previous investigations have shown that Google does not maintain a “centralized public repository of all of its journalism sponsorships” [
38] (p. 2), and neither does it disclose the full size of each award. Almost 20% of the funded projects presented some issues regarding the total amount allocated to their proposals, and they have been categorized in our database as “not disclosed”. From the projects that did show a specific amount, 38.4% were listed as an undetermined range, matching the full size of the award for each category (EUR 50k in prototypes, EUR 300k for medium and up to EUR 1m in large projects).
Google keeps an active website that compiles most of the awarded companies and a brief description of each project. Unfortunately, it no longer includes all of the projects funded. We do not know, for instance, what happened to the French publishing house E2J2 that was present on the 2020 report because we have been unable to find any references by the DNI Fund to this project. The German regional newspaper Rhein-Zeitung is nowhere to be found, something that happens as well with Charged in the Netherlands, Honire in Slovenia, and the British Sky News, among others. Their presence in each year’s reports swells the ranks for each country, but there is no more information on the DNI website.
Last but not least, we must comment on the lack of consistency in the updates maintained by Google, which hindered our attempts to keep track of each project and classify them properly. For instance, there is no way of differentiating Round 1 from Round 2 projects in the DNI Fund online page. However, people like Danny Lein, founder of the software company Twipe, have published reports and articles online trying to gather as much information as possible regarding these funding opportunities [
40].