Champions acknowledging the link between gender balance and business impact
Even in Norway, widely perceived as progressive and egalitarian, gender imbalance in content is a reality – that’s what media company Amedia found after investigating its own output. The publisher also discovered evidence to support their belief that better gender representation makes business sense. After analysing stories from 19 newspapers published across a 21-month period, they found that publications which featured more stories containing women sources had higher readership among women. This interesting correlation was identified as part of a wide-ranging investigation into gender balance across 660,000 stories from 64 newspapers published over the same 21-month window – a massive endeavour, made possible through data science and automatic classification. Using natural language processing techniques, Amedia was able to automate the process of identifying and counting the number of women and men mentioned in its content. They first identified so-called entities referenced in stories, such as people, businesses or organisations, in order to be able to differentiate between, say, a business with ‘Nina’ in its name, and an actual person named ‘Nina’. The second step involved matching recognised names with Statistics Norway’s public database of men’s and women’s names. With this data in place in their centralised data repository, they were able to run analyses towards all aspects of data.
They found that across the 660,000 stories, on average, only 34% of names mentioned belonged to women. But further analysis of the individual newspapers revealed significant variation. They found that the content of smaller publications tended to be more gender balanced than larger, regional ones.
For instance, the top performer had a 42% share of women’s names. The differences in gender balance became even more apparent when they mapped gender across different story topics. In ‘Education’, which encompassed some 30,000 stories, 47% of all names belonged to women, compared to only 21% for the 37,000 stories under the topic “Disasters, emergencies and accidents’. The category dubbed ‘Society’, which covers topics such as communities, families, and welfare, was the only one in which women’s names were in the clear majority.
These differences in gender representation may well be a result of authority figures skewing towards being women or men in different sectors. However, they could also stem from journalists’ own gender biases influencing who they choose to interview.
Future plans include offering all of Amedia’s journalists and editors running data on gender imbalance. Some editorial teams already use a specially developed dashboard, which informs them about the gender gap in readership in real-time. (See AMEDIA)
Champions tracking gender balance
BBC 50:50 UK, Global
The BBC releases its 50:50 challenge impact results annually.In late 2016, BBC presenter Ros Atkins started to experiment with improving the gender balance of the sources featured on his TV programme, Outside Source. His team started to collect data on who appeared in each broadcast, identified subject areas and stories where women were underrepresented
and expanded their network of highly qualified women sources. As a result, women sources featured on the programme went from 29% to 51% in the space of four months. And while the BBC’s broadcast rankings went down 2% in two years, Outside Source’s rankings went up 25%. Since then, more than 500 BBC programmes have joined the project, dubbed 50:50, and more than 20 external media partners have signed up to replicate the idea.
The methodology behind the 50:50 project is relatively simple. Production teams independently collect data on the gender balance of their broadcasts, using a measuring system adapted to the nature of their programme. Data is shared each month among all those who participate. Given that the programmes which take part range from news to music to politics, the measuring system can be adapted to suit the nature of each broadcast. For instance, in a television show where the presenter is controlled by the network, he or she would be excluded from the final count, whereas a production team that can determine who presents the programme may include presenters in their figures. Atkins’ reasoning for allowing these adaptations is that teams are more likely to trust and act upon the data they collect if they have a say in establishing the methodology.
Naturally, the data collection has to be credible enough for teams to believe in the numbers, but the data itself is not the end goal of the project. Rather, it serves as an engine to help drive change and motivate participants to increase the number of women in content.
FINANCIAL TIMES UK, Global
When the Financial Times (FT) defined the size of its women audience for the first time in 2016, it found that it was relatively small and disengaged. Viewed as both a concern and a business opportunity, this realisation brought about several projects aimed at changing women subscribers’ perception of the brand, increasing women engagement and driving internal cultural change. One of these projects is the JanetBot, a machine learning tool that uses facial analysis software to identify the gender of people in images used by the FT. It shares gender classifications with editors via coloured on-screen flags and gender balance data via a Slack channel.
Within the organisation, the bot’s goal is to raise awareness of gender imbalance in home page pictures and encourage journalists to increase the number of images featuring women. It also serves as a tool to boost engagement among their women audience – FT analysis shows that women are more likely than men to click on stories illustrated with pictures of women. The FT has learned several lessons from launching the JanetBot. For starters, those who will be using the product need to be involved in the design process from the start. Once it has been developed, ironing out features that users find unhelpful can be hard and that hasn’t been possible with the JanetBot due to limited resources.
As a result, the bot’s purpose has now shifted towards raising awareness of gender balance among newsroom staff rather than data collection, reminding journalists to consider the gender balance of pictures early in the life of a story. The FT has also been experimenting with bots to track the number of women and men contributors featured in stories to help journalists achieve a more balanced split.
The bot She Said He Said, launched in 2018, keeps track of the gender balance in sources by counting men’s and women’s pronouns and names, sharing data automatically with FT teams. However, encouraging them to look at this data and take action has proved challenging. This has led the FT to explore a manual approach to tracking and improving gender balance instead, having signed up to the BBC’s 50:50 project this year. Although the experiment is still in its early stages, the 50:50 methodology seems to raise awareness of imbalances and engage colleagues more effectively than the automated approaches. (See The Financial Times Deep Dive)
Champions increasing the visibility of women
SOUTH CHINA MORNING POST Global
With a series of company-wide initiatives, the South China Morning Post (SCMP) is making a concerted effort to boost women’s readership and increase the visibility of women across all of its products. The development and implementation of the initiatives were put in motion following the SCMP’s 2018 Hackathon, which highlighted the gap between men and women readership, the disparity between women and men sources in stories and the difficulties in finding expert women sources in various fields. Building on the findings from the project, the SCMP launched a campaign asking readers, partners, and external contacts to nominate women experts via an online form promoted via direct outreach, social media and newsletters.
To date, the SCMP has received nominations for women with expertise as diverse as nanotechnology, economic empowerment, retail, cloud accounting, gender equality and healthcare. The nominated experts are verified by an internal team and made available to the entire newsroom in a searchable database. With regards to tracking gender balance, the SCMP has taken advice from the BBC’s 50:50 project and uses some of the same metrics and guidelines.
Collaborating with other news organisations provided the SCMP with invaluable advice and helped them avoid some of the problems others faced. “I was inspired by the response we got when we reached out to other organisations who are so willing to share their experiences,” says Laura Warne, the SCMP’s digital editor.
“Everyone can do better in this area. It isn’t about competition. It’s about improving the media landscape around the world.”
Bloomberg is pursuing the same goal and implemented a company-wide mandate in 2018 to increase the representation of women sources in both online and on-air content. The 2,700 journalists and analysts working for the editorial division in 120 countries were asked to get to know the women experts in their beats to help establish a database the entire organisation can draw on. It now features more than 2,300 names, up from 500 at the start of 2018. At the same time, Bloomberg is trying to increase the number of women who are media trained and approved to speak to the media by their employers via its New Voices initiative. As part of this, Bloomberg offers and funds media and communications training for senior women and other diverse executives from leading financial firms.
In 2019, the company is sponsoring training for 12 women executives in each of the programme’s locations: New York, San Francisco, Toronto, London, Dubai, Mumbai, Hong Kong and Sydney.
RINGIER GROUP Switzerland
The Ringier Group launched its EqualVoice initiative in November 2019 setting an important precedent within the Group. At its core is the EqualVoice Factor – a measurement tool which provides data on the proportion of women and men in articles across Ringier’s media titles.
The initiative, however, is more than simply measurement and tracking. The initiative includes plans to launch a series of events, a special magazine and online channels (across all titles) on equality. A group-wide list of female experts is also currently being developed and a nationwide call for a photo challenge is being launched to depict the reality of women and men in the world of work and to equip image databases with these images.
The EqualVoice initiative is chaired by publisher Michael Ringier and CEO Marc Walder, and supported by members of the Ringier Group Executive Board and a dedicated Advisory Board.
Champions tapping into women audiences with engaging content
FINANCIAL TIMES UK, Global
In addition to tracking gender balance and raising awareness internally, the FT has also launched several content-related initiatives to boost engagement among women subscribers. With its Long Story Short newsletter, the FT targets women subscribers with stories that interest them in a format that they have said they enjoy. Tailored to be a catch-up newsletter for busy women, it is sent out every Friday and combines the biggest stories and best reads in one email. For each edition, a different woman journalist from the FT handpicks the story, sharing her personality, expertise, and interests while offering a ‘behind the scenes’ look at FT stories and the reasons they have caught her eye. The choice of stories is also informed by data on what women have been reading during the past week and includes both news and features on a range of topics. The reason for broadening coverage areas features is that the FT heard from women readers that they perceive the brand to be ‘just about finance’. The tone of the newsletter is deliberately distinct from that of the core FT brand, more informal and conversational.
It also differs visually from the core product, featuring a different logo and colour palette as well as a more diverse range of people in the images used. It is not, however, explicitly branded as a women’s product. So far it has proved successful among the FT’s women audience, while also engaging men subscribers. Data from the third quarter of 2019 showed that it has a bigger audience among women than the FT average, and has higher open and clickthrough rates, up 3.5 percentage points year on year and 1.8 percentage points, respectively.
SOUTH CHINA MORNING POST Global
Another initiative that grew out of the SCMP’s Hackathon focuses on generating targeted content for women, finding innovative ways to distribute it and building a community around it through a diverse set of platforms. Last month, the SCMP launched Lunar. It showcases content from across the newsroom as a curated package “with news, interviews and in-depth features about women, by women and for women”. In line with the cross-departmental and collaborative approach that characterises the SCMP’s gender balance initiatives, the team in charge of Lunar also encourage newsroom staff to think about developing stories that suit the platform, and shaping them in a way that might increase engagement among women.
The SCMP conducted audience research in order to better understand what type of content appeals to women readers. Some of the topic areas that stood out were diplomacy, regional news and society, which covers a broad range of social affairs, including education, issues of equality and cultural trends. “Women are interested in a really diverse range of content,” Warne says. “We don’t want to limit this or be narrow in our approach. We’re looking at content that not only features women but that interests and affects them. We want to be very broad in our definition of this.” (See The Financial Times Deep dive P. 32)
LUSAKA SUN Zambia
Creating engaging content for women is also a priority at the Lusaka Sun in Zambia, a standalone publication targeted at the country’s lower income class, and launched by the Daily Nation in January this year. Rather than focusing on political coverage, which is common among other Zambian news outlets, the Daily Nation’s executive editor Mary Mbewe says The Sun set out to highlight topics such as social issues, injustice and trade. Some 70% of the paper’s staff are women and one of the publication’s main audience category are traders, the majority of whom are women. Having such a high number
of women employees was a deliberate decision. “I feel that women are more empathetic to what is happening in the areas that we are targeting,” Mbewe says. “It’s easier for them to talk to fellow women in the compounds, for instance, or at the market, and for those people to open up, and tell their story and their struggle, and how they are managing to make it.”
As part of one of The Sun’s policies aimed at increasing gender balance, they publish at least one positive and inspiring story per week which features a woman as the main subject. The initiative has generated good readership, according to Mbewe, and led to more women coming forward to share their stories. To further encourage women’s participation and visibility, The Sun also launched a platform where they can share their stories, images or leave their information to be contacted in the future. In addition to the website and an epaper, which is currently under development, The Sun is working on a platform for mobile which will make it easier for its audience to access the paper’s content.
At ARA, a Catalan daily newspaper, three journalists launched ARA Feminismes, an initiative aimed at creating and distributing content with a focus on gender perspective. Wanting to find a broader audience for the gender balanced stories they were already producing,
the three journalists behind the idea, Lara Bonilla, Marta Rodríguez and Thais Gutierrez, started work in the months leading up to 2019 International Women’s Day on March 8 and timed the launch to coincide with the event.
Now, ARA Feminismes actively distributes its content on ARA’s website, a weekly newsletter and via a Twitter and Facebook page with some 3,500 and 7,500 followers respectively. The grassroots initiative has garnered support from management and the trio behind it, who are leading the project in an unofficial capacity in addition to their other work, offer advice to colleagues on how to write gender balanced content. They encourage the newsroom to submit stories that would fit the ARA Feminismes brand. As part of the project, the team also launched several calls to action on the website in a bid to solicit input for stories from women. One of these led to a successful interactive article about problems and issues new mothers are experiencing, giving them an opportunity to voice their opinion about a rarely discussed topic in mainstream media.
Having found that women experts are more hesitant to speak to the media, regardless of their competency compared to their counterparts who are men, they next want to put together a database with women sources which the entire newsroom can draw on. Another company-wide initiative includes an analysis of gender balance in the newsroom and content, conducted by an external company.
DEEP DIVE : THE FINANCIAL TIMES
By Kesewa Hennessy, Digital Editor for Audience Engagement, Financial Times
Why the Financial Times began women engagement projects
The Financial Times’ (FT’s) gender projects started with data. This has proved crucial in helping us understand our women audience and work towards meeting their needs.
In 2016, for the first time, we were able to define the size of our women audience using the inferred gender model. The data confirmed that this audience was relatively small and disengaged. We saw this as both a concern and a business opportunity. Follow-up research helped us understand why this audience was not engaging with the FT (for example, women readers told us they found the FT a masculine product) and how we could try to change this. The inferred gender model also allowed us to see which stories women were reading, further increasing our understanding of this audience.
We refocused the remit of the role of Digital Editor for Audience Engagement to include setting up and running a series of gender projects which have two broad aims:
- Externally: To change women subscribers’ perceptions of the FT, increase their engagement in the long run and build loyalty in the audience.
- To better understand what our women audience consumes.
- To encourage culture change among colleagues to better meet the demands of that audience.
These projects are aimed at engaging existing women subscribers rather than reaching a bigger women audience for the FT. But we also hope that a possible corollary of some of these projects will be to change the perceptions of women who don’t yet read the FT.
To encourage culture change among colleagues to better meet the demands of that audience.
One of our key metrics for determining women’s engagement with a story, topic or section is the proportion of women readers and how this compares with the overall benchmark. We are always aiming to increase that proportion and raise the overall benchmark.
1.2 percentage point increase between Q1 2017 and Q3 2019.
- Average percentage of women readers for an FT story
3 percentage point increase between 2017 and 2019 year on year.
THE FT’S WOMEN ENGAGEMENT PROJECTS
The project: We decided to promote stories that women want to read in prominent slots; that is, stories that we think are likely to exceed the average percentage of women readers for an FT story (‘overindex’) based on past engagement with that topic.
HOW DOES IT WORK?
The home page editors and Digital Editor for Audience Engagement together select one piece a day likely to do well with women. The story is promoted at the top of the home page, in a popular daily news digest email with a high proportion of women readers, and on social media. We then track how our Project XX choices perform with women.
WHAT’S THE GOAL?
- Externally: To change women perceptions of the FT by trying to ensure women are more likely to spot the kinds of stories they want to read.
- To learn more about women subscribers by tracking the performance of Project XX stories and to share the information with relevant teams.
- To encourage culture change among relevant teams. For example, home page editors to promote a wider range of stories and to consider ‘success’ in terms of metrics other than page views; commissioning editors to commission a wider range of stories.
HOW DO PROJECT XX STORIES DO WITH WOMEN?
The average proportion of women readers for a Project XX story is several percentage points higher than for an average FT story. The goal is to continue to increase that difference.
WHAT WERE THE CHALLENGES?
Primarily, bringing about culture change on the home page, one of the most valuable parts of the FT. Here are a few examples:
- Encouraging home page editors to accept a new practice, especially one with the potential to be seen as interfering with news judgment and independence.
- Encouraging a focus on metrics other than total page views (e.g. % of women page views).
- Promoting stories not seen as ‘core’ FT (often described as ‘niche’, ‘soft’, ‘fluffy’).
- Broadening ideas about who ‘the FT reader’ is and what they want to see on the home page, using both data and a wider range of journalistic perspectives.
- Challenging preconceptions about what women FT subscribers read –and equally about what men subscribers read (however high the percentage of women readers, FT stories above our benchmark number of total page views are read by more men than women).
- Gaining an understanding of the practices and decision-making processes of the team that runs the home page. These are not always clear to those outside the team.
- Establishing workflows, as well as rapport and trust between the home page and AE teams.
- Embedding a controversial practice with status somewhere between compulsory and voluntary, ultimately relying on colleagues’ goodwill.
WHAT WE’VE LEARNT FROM PROJECT XX:
- Cultural change can be slow but it happens. The kind of stories home page editors consider home page material has shifted as they’ve seen Project XX stories perform extremely well in terms of page views.
- Project XX has helped expand the range of stories that perform well for the FT. In particular, it has helped show home page editors that work-related topics previously seen as ‘niche’ will often gather high total page views as well as over-indexing with women. Greater interest in promoting new kinds of over-indexing stories in turn encourages commissioning editors to produce more. These topics include fertility, menstruation, menopause, parenting, sexual harassment, racial discrimination and LGBTQ+ issues.
- Incentives help win editorial support. Examples include the opportunity for additional promotion, as well as information on which stories could attract more readers for them and insights into what home page editors want.
- Share personalised data with colleagues. To engage colleagues (and to help the new project take root in the early days), we let people know how their stories fared with women readers. Good news is always welcome.
- Face-to-face communication is important with new/ controversial projects. Understanding colleagues’ team culture, and hearing their ideas (and concerns), is an essential part of the project design process.
- Start with a small trial. You can extend it, adjust it or drop it if it doesn’t work. You’ll always learn something.
- Leadership matters. Signalling and support from senior editorial colleagues, explaining what’s happening and why it matters, is important. Some acknowledgment of colleagues’ extra efforts helps too.
- Data shows us that our preconceptions about women (and men) are often wrong. While stories on ‘women’s’ topics (fashion, style, beauty) over-index with women, the same stories are always read by a higher percentage of men. This is also true of articles on topics from managing your career to femtech. This reflects the makeup of our subscriber base, which is about 75% men. Some of the FT’s more specialist financial content over-indexes with women (financial regulation, fund management).
LONG STORY SHORT WEEKLY NEWSLETTER
The project: We wanted to target women subscribers with FT content that we know women read in a format they told us they liked.
HOW DOES IT WORK?
We send out a newsletter every Friday, combining the biggest stories and best reads in one smart email. It’s handpicked by a different women FT journalist each week.
WHAT’S THE GOAL?
To meet the needs of women who told us they tended to be time-poor, wishing they could engage more with the news. We approached this in several ways:
- Format: Tailored to be a catch-up newsletter for busy women. It offers them an opportunity to read the stories we think they shouldn’t miss, over the weekend.
- Content: The curators’ choice of stories is informed by data on what women have been reading during the past week. Because women readers told us they saw the FT as ‘just about finance’, we include both news and features on a range of topics – the best reads as well as the biggest stories. They are all FT stories we think are relevant to women, which might otherwise have been missed.
- Curators: To make the newsletter feel more relevant to an audience of women, a different women journalist curates every week. This highlights our women journalists and offers women perspectives on the news.
- Tone: Deliberately distinct from that of the core FT it is informed but also informal, engaging and conversational; designed to be an enjoyable read. Each curator brings her personality, expertise and interests to the week’s edition, offering a ‘behind the curtain’ look at FT stories and the reasons they have caught her eye.
- Visually: To set it apart from core FT, we chose a specially commissioned, hand-drawn logo and a different colour palette. The images feature a more diverse range of people.
HOW IS LONG STORY SHORT DOING? *
It has a highly engaged audience that is more women than the FT average.
- 8.2 percentage points more women readers than the average FT newsletter.
- Average open rate is 3.5 percentage points higher.
- Average clickthrough rate is 1.8 percentage points higher.
- Month-on-month growth in subscriber base since launch in early 2018.
*Data from Q3 2019
WHAT WERE THE CHALLENGES?
- Designing a newsletter format for an audience with which we were relatively unfamiliar.
- Finding a new design and tone of writing appropriate for the audience, while retaining an FT feel.
- Working with a wide range of authors and still maintaining a consistent feel.
- We wanted to ensure this newsletter had a diverse line-up of authors.
- Establishing a new editing platform to enable a more flexible design.
WHAT WE’VE LEARNT FROM LONG STORY SHORT
- Explicit ‘women’s’ branding is not essential. This product engages women, even though it’s not branded anywhere as a women’s product. We don’t know what effect ‘women’s’ branding would have, positive or negative.
- Men read ‘women’s products’ too. This newsletter is designed to engage women but also engages men subscribers.
- Direct feedback is limited. The audience for this newsletter doesn’t interact directly with the authors so our feedback at present is largely restricted to engagement metrics. We’re not sure why, though it could be because it’s hard for readers to build a sense of a personal relationship with such a large number of rotating authors.
The project: We wanted to find a way to get a better gender balance of pictures on the FT home page.
HOW DOES IT WORK?
JanetBot is a machine learning tool that uses facial analysis software to identify the gender of people in photographs in FT stories. The bot shares gender classifications with editors via coloured on-screen flags. It also shares gender balance data through a Slack channel.
WHAT’S THE GOAL?
- Externally: To increase women engagement. Our analysis shows women are more likely than men to click on stories illustrated with pictures of women.
- Internally: To raise awareness of gender imbalance in home page pictures and to prompt FT journalists to take action to increase the number of pictures featuring women.
WHAT WERE THE CHALLENGES?
- The underlying facial recognition software/database appears to have particular trouble classifying non-white faces, reflecting biases reportedly common in AI.
- The bot requires frequent correction’ of classification from users, which is beyond newsroom resources.
- Convincing colleagues to look at the automated data and integrating this task into their workflow.
WHAT HAVE WE LEARNT?
- To design a product that secures newsroom engagement, editorial users need to be involved at the inception.It can be hard to iron out features that users find unhelpful once a product has been developed and launched. In this case, limited resources mean it has not been possible to do so.
- JanetBot has therefore proved more useful in raising awareness of the gender imbalance (via the on-screen flags) than as a data-collection tool. We hope to use JanetBot’s awareness-raising capabilities in future projects, switching the focus from published output to reminding journalists to consider the gender balance of pictures early in the life of a story.
- It’s hard to get newsroom users to change their behaviour using automated processes. We are now aiming to move towards a more manual approach.
SHE SAID HE SAID AND 50:50 PROJECT
The project: We wanted to help FT teams include a better balance of men and women contributors in their output.
What’s the goal? To build on FT analysis showing that articles produced by teams that quote a high proportion of women are well read by women. In addition, teams that quote a high proportion of women also feature more women in their pictures.
How does it work? We began with automated sharing of data with colleagues, intended to raise awareness of the gender imbalance. This year we’ve shifted our focus to a more manual approach:
- She Said He Said (launched 2018):A bot that tells us the gender balance of sources quoted in published FT articles. It uses pronouns (‘she said… he said…’) and first names to determine whether a source is a man or a woman.
- 50:50 Project (launched 2019): A simple, voluntary system of self-monitoring, where teams track the gender balance of their contributors in published stories. Launched in partnership with the BBC as a follow-up to She Said He Said.
WHAT WERE THE CHALLENGES?
- She Said He Said: Getting colleagues to look at automated data in the first instance, and to act based on this data.
- 50:50 Project:
- Identifying a colleague in a journalistic role to take the lead, which is helpful in encouraging other journalists to take part.
- Finding ways to keep momentum going among participants, especially when their percentage of women contributors starts from a low base, improves only slowly, remains static or gets worse.
- Personal communication with thosetakingpart is proving vital, to identify and correct any misapprehensions about the goals of the project and to discuss ideas. Six commissioning desks have so far opted into the project. However, these desks deal with many of the FT’s hundreds of reporters and editors. Finding an effective way to communicate with such large numbers is a challenge.
WHAT HAVE WE LEARNT?
- The manual approach is more engaging. At this early stage in the 50:50 Project, it seems to raise awareness of imbalances and engage colleagues more effectively than automation. It also appears to be a better way to get discussion going, in turn raising awareness across the newsroom.
- Work with people who want to be involved. 50:50 seems to ‘give permission’, a focus and structure to those already keen to bring about change.
- Personal contact with colleagues matters. The voluntary and team-led approach allows colleagues to take ownership, which appears to increase commitment. Where the project has at first been interpreted as a top-down edict, enthusiasm has sometimes been limited. By contrast, when teams have been invited to discuss their difficulties and share ideas appropriate to their own work, enthusiasm and engagement has increased.
DEEP DIVE : AMEDIA
With input from Jostein Larsen Østring, Amedia’s Vice President for Development
WHO ARE AMEDIA?
Owner of more than 70 titles, Amedia is Norway’s largest publisher of local newspapers and reaches a daily readership of some 1.8 million people.
WHY DID AMEDIA BEGIN DATA ANALYSES ON WOMEN ENGAGEMENT?
In 2016, Amedia began a concerted effort to analyse the data coming from their newsrooms to better understand their audience and the content that audience was paying attention to. In the first instance, they spent around two months working with chosen newsrooms, looking at different topics the publication covered in order to enhance the experience for their readers and better understand how to balance free articles and those behind the paywall. The analysis was expanded to other areas, including gender data. One of the main learnings Amedia has taken away from these projects is the importance of relevance when it comes to creating content for its readers and subscribers. A number of factors influence what stories readers are interested in, such as their age or geographical location, meaning Amedia newsrooms need to produce a variety of content that is relevant to these different audience segments in order to keep them engaged. In the case of gender, Amedia looked at articles from 19 newspapers published across a 21-month period and discovered that publications which featured more stories containing women sources had more women readers. This correlation was identified as part of a wide-ranging investigation into gender balance across 660,000 stories from 64 newspapers published over the same 21-month window. It revealed that, on average, only 34% of names mentioned in a story belonged to women.
WHAT WAS THE GOAL?
The ultimate goal is to better serve all segments of the population who choose to read Amedia’s newspapers. Improved gender representation also has the potential to attract more subscribers and increase revenue.
- Externally: To use the data to attract women readers, deliver more clicks in the short term and encourage more women to read and subscribe to Amedia’s titles in the long term.
- To encourage newsrooms to use the data to understand which stories are more relevant to women, and act to increase engagement among women subscribers
- To encourage culture change in Amedia newsrooms, making it possible to implement the desired change and have a real impact
HOW DOES IT WORK?
Amedia used natural language processing techniques to automate the process of identifying and counting the number of women and men mentioned in its content. They first identified so-called entities referenced in stories, such as people, businesses, or organisations, in order to differentiate between, say, a business with ‘Nina’ in its name and an actual person named ‘Nina’. The second step involved matching recognised names with Statistics Norway’s public database of men and women names. With this data in place, they were able to run analyses on content published by Amedia newspapers.
WHAT DID THEY FIND?
While Amedia’s analysis of 660,000 stories revealed that, on average, only 34% of names mentioned belonged to women, a closer look at the data revealed significant variationin aseemingly consistent picture.Forone, gender balance differed across the 64 newspapers that were part of the analysis. The top performer had a 42% share of women’s names in its stories, while the newspaper at the other end of the list had a 28% share. In general, the content of smaller publications tended to be more gender balanced than that of larger, regional ones. Given that the latter more often work on regional and sometimes national topics, this could indicate that the authority figures they interview are more often men, which could be as much of a contributing factor as the choices Amedia’s journalists make. The differences in gender balance became even more apparent when Amedia mapped gender across different story topics.
In ‘Education’, which encompassed some 30,000 stories, 47% of all names belonged to women, compared to only 21% for the 37,000 stories under the topic ‘Disasters, emergencies and accidents’. The category dubbed ‘Society’, which covers topics such as communities, families, and welfare, was the only one in which women’s names were in the clear majority. Again, these differences in gender representation may well be a result of authority figures skewing towards being women or men in different sectors. However, they could also stem from journalists’ own gender biases influencing who they choose to interview.
Lastly, Amedia discovered evidence to support their belief that better gender representation makes business sense. After analysing stories from 19 newspapers published across the same 21-month window, they found that publications which featured more stories containing women sources had higher readership among women.
HOW DO WOMEN RESPOND TO THESE STORIES?
Amedia used their data to identify stories and subject areas which are more relevant for women readers. This helped them understand what they need to do in order to produce content that is more relevant to women and increase the number of women subscribers.
WHAT ARE THE CHALLENGES FACING AMEDIA?
The biggest challenge is to ensure that the data changes everyday behaviour and is used by editors and journalists. The response amongst staff differs dramatically. Some embrace the new information while others ignore it. Most people in the newsroom do understand the value of the data and are using it to think about how they can better reach their audiences, but others remain focused on short-term deadlines and their specific areas of interest – they choose not to prioritise this initiative. The key to success is to drive internal culture change and convince everybody within the organisation that this is a priority. That will take time, but within a year or two Amedia hopes that this work will lead to significantly more women sources appearing in stories and more engaged women readers.
WHAT HAS AMEDIA LEARNT?
- It takes time to get to a position where data is informing the choices being made by people working within an organisation.
- Findings are being introduced to news editors who will have a responsibility to translate the findings into action.
- Gender balance differs across publications and topics. Smaller titles are generally more gender balanced than larger, regional ones. Women are better represented in topics such as health, education, and society, but are featured much more rarely in categories like emergencies and sports.
- The differences in gender representation may well be due to authority figures skewing towards being women or men in different sectors. However, they could also stem from journalists’ own gender biases influencing who they choose to interview. In short, the answer is not clear cut.
- All newsrooms are different and have different priorities. The project’s impact will really start to be seen next year, when a significant number of newsrooms have had time to use this data to change gender representation in their stories.
- Newsrooms are spending a lot more time on gender discussions, and data analysis is driving an increased focus on gender balance.
- Work to reach shared gender equality goals is also good for business, leading to more readers and subscribers.