Strategy/Wikimedia movement/2017/Sources/Indonesia research findings draft May 2017

Summary of Key Opportunities & Findings: Indonesia & Brazil

Working Draft | Updated: June 6 2017 

Prepared by Reboot

Shared with Wikimedia Foundation

I. Introduction  

This memo captures key findings from Movement Strategy New Voices (formerly Track D) research conducted in Indonesia and Brazil in May 2016. It is a working draft, intended to inform Movement Strategy processes currently underway at the Wikimedia Foundation (WMF).

Key findings have been organized by the five key research themes guiding this study.

• Theme 1: General Information Seeking & Learning
• Theme 2: Internet Usage (Web & Mobile)
• Theme 3: Online Information Behaviors
• Themes 4 & 5: Awareness, Perceptions, & Usage of Wikipedia

This memo emphasizes emerging trends that may be most relevant to ongoing Movement Strategy conversations, with a focus on behaviors observed among younger respondents given their importance to the future of Wikimedia. The document starts with a synthesis of key trends, their implications for the movement, and concrete opportunities for Wikimedia to consider. It then summarizes the research approach and highlights key research findings. Opportunities for the Wikimedia movement have been highlighted throughout the key findings section, under each relevant research theme. All opportunities noted in this document are preliminary, and intended as jumping-off points for further discussion.

Given the urgent timelines of the Movement Strategy process, Reboot has done its best to surface top-level, draft findings to feed into Cycle 2 conversations. Separate and more formal research products will be produced that cover the breadth of findings from Indonesia and Brazil, by late June 2016.

While this analysis is grounded in findings from Indonesia and Brazil, it builds upon patterns from other contexts where Reboot conducted Movement Strategy / New Readers research, namely India and Nigeria. We hope this approach is helpful for WMF in rapidly synthesizing findings from across these contexts, and integrating them into the ongoing strategy process.

That said, Reboot believes that a focused comparative analysis exercise across the countries with which WMF engages could help the Foundation determine how to interpret and weight findings from different countries—both where it has conducted primary research, and beyond—based on the unique characteristics of each country (in its own right or as illustrations of similar contexts). Characteristics which WMF may use to map, compare, and categorize countries include: basic demographic and economic indicators; sociocultural characteristics; relevant macro trends (e..g. rate of urbanization, rate of growth in mobile penetration); and relevant political economy factors (e.g. degree of media freedom, levels of public sector corruption). Such an exercise can help WMF determine which countries (where primary research has been conducted) are more representative of certain broader trends being explored as part of the Movement Strategy process, and can help the movement determine unique and tailored strategies for different contexts (e.g., delineated by factors such as region, mobile penetration, or level of trust in traditional media, etc).

II. Synthesis of opportunities

Wikimedia in 2030: A Source of (not a Destination for) Knowledge

The central question surfaced through this research was: Wikipedia has all this content, so now what? Or to put it another way: How can the sum of the world’s knowledge be more relevant to and useful for people’s lives?

The research suggests that building a robust and respected destination for knowledge is no longer enough. Looking forward, in order to continue advancing towards its mission, the Wikimedia movement must become a source of—rather than a destination for—knowledge. It can do this by empowering and enabling learning, wherever and however people learn.

Some of this is currently possible via internal initiatives such as Wikipedia’s syndication feeds or is currently happening via external products such as Google’s Knowledge Graph, but many such applications seem to have been either i) experiments driven by individual community members’ interest in new technologies, or ii) developments driven by external parties, and sometimes opposed by Wikimedians.

It is our hope that this research can contribute to the development of a coherent movement strategy grounded in market trends and user needs, and embraced by the community as necessary, evolutionary steps in its model.

What follows is a summary of key market and user trends surfaced through this research, their strategic implications for Wikimedia, and specific resulting opportunities that the movement may consider as it continues to develop its strategy.

Key trends

The research in Brazil and Indonesia—and past work for New Readers in India and Nigeria—surfaced a few key trends that the movement should consider in the development of its strategy:

• People are decreasingly visiting individual websites or web properties; instead, they are getting content from peers or curators on the platforms where they socialize.
• People no longer expect their content to be mediated by “trusted institutions”; instead, they want their content curated by trusted individuals. Online, trust is often a badge bestowed by the crowd—in the form of high volumes of likes or followers—on individual experts or influencers for a particular content area, who may not be associated with institutions.
• People and organizations are increasingly content creators and curators themselves, but they need foundational ingredients from trusted sources to help them make their arguments, develop their brands, and fulfill their missions.

Strategic Implications for Wikimedia

The Wikimedia movement is well-positioned to take advantage these trends in service of its mission. But moving forward, the sum of all knowledge will no longer be a destination people go to—rather, Wikimedia content will be a source of and an engine for learning, in all the diverse and dynamic ways human beings learn and consume knowledge.

In 2030, this might mean:

• For other publishers, Wikimedia is a central provider of transparent and reliable information. Through more distributed publishing via APIs and other means, Wikimedia content is published through many platforms, from The New York Times to the next Snapchat. This enables Wikipedia to serve as a just-in-time encyclopedia: providing relevant context and information to people, when and where they are learning about or discussing the world around them.
• For other knowledge creators, Wikimedia is empowering their development of content. Through features that enable greater modularity and portability of its content, Wikimedia enables individuals and organizations to develop custom content packages for diverse use-cases. For example, an educational nonprofit can use elements of Wikipedia content—e.g. 10% of the most relevant aspects of its article on the Internet, among others—to form the basis of a digital literacy course.
• For the community, Wikimedia has increased the visibility (and therefore the attractiveness) of Wikimedia contribution across diverse communities and contexts. Through features and programs that recognize community members’ contributions, Wikimedia can highlight the expertise and trustworthiness of its content creators, curators, and moderators. This serves a few purposes: it appeals to users’ growing desire to know the individual source of their content, it generates awareness of Wikimedia’s robust (but often misunderstood) model, and it serves as a way to recruit other contributors (both through increased awareness of the possibilities of contribution, and the social validation incentives enabled by this approach).

This research suggests that the Wikimedia movement can exponentially grow its impact by using its rich, deep content to power knowledge creators and platforms worldwide. It can help people discover, learn, and debate in all the diverse ways and places they like to do so. Doing so also leverages the movement’s strengths. As an accessible, digital, and transparent community of knowledge, it is uniquely suited to serve as the world’s engine and enabler of learning.

Specific Opportunities

From the broad vision outlined above, the research surfaced specific opportunities to help make Wikimedia more relevant to and useful, based on patterns in how people are getting and consuming information. The opportunities outlined go from those that are most aligned with Wikimedia’s current products and strategy, to those that require new ways of thinking about Wikimedia’s model and direction:

1. Help people find what they need on Wikipedia through more precise search and topic guides / curation.

   Information-seeking is becoming increasingly task- and search-led, and less discovery- and browsing-oriented. To learn new topics, people are increasingly looking to content curators or presenters—whether via niche blogs or YouTube channels—to help them learn.

   This means that on Wikipedia, people often find the articles ill-suited for their specific needs: “too long” and “too hard to find / learn what I need”. Wikimedia can address this by directing readers to specific parts of articles that are most relevant to their task or their search interest, or developing topic portals that guide learners through a specific topic of interest. Portals would ideally be curated and maintained by passionate community members (or external content experts / influencers), since people are increasingly looking to people, not institutions, to curate their content experiences.

2. Meet user expectations and preferences for online content by getting more visual, real-time, and social.

   As young people and digital natives become a larger part of the online population—and influence both the behaviors of those around them, and the direction of technology platforms—norms and expectations for digital experiences are changing. Visual, real-time, and social aren’t just buzzwords; research found that they are the characteristics of content platforms that young people increasingly prefer. Young people are getting their news by following the Instagram accounts of news sites, looking at trending topics on Twitter, and then discussing their collected news with friends in WhatsApp groups.

   These behaviors are forcing media organizations to adapt how they package, distribute, and support conversations around their content—and Wikimedia will need to consider how it adapts its model in this changing landscape. Might it allow videos to serve as references, to appeal to how young people like to learn? Might it push alerts on rapid, concentrated edits to prominent pages to popular platforms, to help break news events that may not otherwise reach the global radar? Might it support conversations on messaging platforms through in-app article previews? Such ideas are worth considering as Wikimedia adapts to the content preferences of the next generation.

3. Empower learning in the diverse places and ways people learn, by making content more modular and portable.

   When Wikipedia first started, the internet was a very different place. There were far fewer platforms for learning—whether formal and academically oriented, or informal and personally oriented. There were also far fewer content creators, both online and offline. Today, the democratization of both knowledge and content creation / publishing has led to a mushrooming of learning platforms. Users, for their part, are enthusiastically learning in diverse ways, from how-to videos on YouTube to homework-help communities such as Brainly.com.

   While Wikipedia may no longer be a primary destination for learning, it has the potential to be a primary engine of learning, by providing relevant, trusted, and customizable content to other content providers. It can support resource-strapped nonprofits developing skills-building curricula for the unemployed. It can support after-school clubs that are interested in extending learning outside the classroom. It can support content experts that want trusted, easy-to-use content to support their curricula on diverse topics. Wikimedia has the potential to service all these communities and, in doing so, exponentially increase the impact of its content.

4. Build a community passionate about helping people learn, by working with niche content experts, curators, and ambassadors.

   Wikipedia has a brilliant community that is passionate about developing the most respected source of knowledge in the world—it now needs to invest in building a community that is equally invested in getting that knowledge out into the world. As noted previously (and described in more detail in the findings that follow), people today are increasingly looking to trusted individuals to curate their content. Also, research has found that a significant number of those that currently discount Wikipedia as a trusted information source have this impression due to comments made by an individual (often a teacher). Bottom line is that individuals matter, whether in creating Wikimedia readers, community members, or naysayers.

   Moving forward, Wikimedia should consider attracting and investing in allies and community members that focus not on just generating content (which is the focus of the community today), but on getting it out to people in the forums and channels they like to learn. This may be programs to support digital influencers in using Wikimedia content and integrating it into their work. It may be ambassadors in schools that empower students to teach their peers about how to access and appropriately use Wikipedia for school assignments. Cultivating allies and champions within the education sector—and particularly in institutions of higher education—can have outsize pay-offs, given the negative impact of skepticism generated within this sector.

5. Build partnerships to serve the hard-to-reach or marginalized, by leveraging the unique value and assets of the Wikimedia movement.

   In its first 16 years, Wikimedia has grown into a powerful and singular global force, with a diverse, passionate, and talented community. Indeed, its vision has inspired many, and drawn them into the movement. Its product has immense global reach. Yet there are still many—those digitally unconnected, those that are illiterate or with basic reading abilities—that Wikimedia’s core products cannot serve. Moving forward, the movement should consider partnerships to help i) expand digital access and literacy, and ii) to improve the accessibility of its content.

   This may include support of wifi access points in public places, developing customized content for digital literacy courses, or providing “simple” or “lite” versions of its content. This may mean community members that provide support services (e.g. adapting Wikipedia content for specific use-cases or curricula) for socially oriented organizations. And all this may require partnership with governments, civil society groups, educators, and other new allies.

   There is, however, the risk for extending itself too thin, particularly as Wikimedia pursues new partnerships. The movement will need to be clear on how specifically it will evaluate different opportunities to reach “every single human being” and what the opportunity costs or trade-offs associated with different approaches may be.

Finally, the success of the above strategies will be contingent on Wikimedia’s ability to clarify its brand and establish understanding of its processes, both their strengths and limitations. In a world that is increasingly skeptical of institutions and savvy to bias, Wikimedia’s model can provide important lessons and neutral, factual content. But, as research across several countries showed, Wikimedia’s content is often not trusted. There are many ways to approach this challenge, and many great efforts already underway, but tackling this head-on will be critical for the movement’s success moving forward.

In addition to clarifying its process, Wikimedia also has an opportunity to increase brand awareness through events and promotions targeted at those outside the Wikimedia community. For example, creating Festivals of Knowledge elevate local knowledge. Think edit-a-thons meet street fairs that bring together local knowledge institutions like museums, cultural institutions, musicians, and social organizations to celebrate local knowledge and culture while also providing on-ramps for people better understand the relevance of the open knowledge movement.

III. Research background

Research objectives & approach

The overarching objective of this project is to inform the WMF Movement Strategy process, by providing primary research on and input from new and existing readers in low-reach markets, where the movement is not currently well-known. Primary research conducted in Indonesia and Brazil seeks to:

• Gather and synthesize relevant local behaviors and trends that will impact the future of the Wikimedia movement over the next 15 years. Specifically, the research examines behaviors and identifies trends related to how people seek, use, and come to trust information, particularly that which is found online.
• Understand the current awareness, perception and usage of Wikipedia for potential and existing readers. Compare awareness, perception, and usage with other platforms for information-seeking and learning to determine Wikipedia’s comparative advantages and limitations—and to identify forward opportunities.
• Learn from local experts and institutions in relevant fields (knowledge, communities, and technology) on how WMF can better advance its mission in the specific low-reach market over the coming 15 years, as well as better serve users within it.

Field research in Indonesia and Brazil was conducted in two field-sprints each over 10 days in duration, and was led by a team by Reboot.

In Indonesia, the research was supported by the WMF’s Global Partnerships Lead for Asia and three Indonesia researchers. In Brazil, the team was accompanied by WMF’s Director of Strategic Global Partnerships, the Global Partnerships Lead from Latin America and the New Readers Product Manager, as well as, three Brazilian researchers.

Research methods included a light literature review, semi-structured ethnographic interviews, user observation, and expert / key informant interviews.

Overview of Research Sample

In Indonesia, field research was conducted with 61 respondents and 3 key informants in West Java. In Brazil, field research was conducted with 52 respondents and 5 key informants in the state of Bahia. Respondents covered a range of characteristics that met the study’s target sample criteria with less than 10% difference in terms of distribution across gender, age, representation from those in the education sector, and level of internet access. The respondent characteristics break down as follows:

	Indonesia	Brazil
Location	48% Bandung (Urban) 36% Sukabumi (Peri-urban) 16% Cianjur (Small Town)	53% Salvador (Urban) 29% Alagoinhas (Peri-Urban) 18% Santo Amaro (Small Town)
Internet Access	48% Unlimited Access 29% Moderate Access 19% Limited Access 9% Unconnected	42% Unlimited Access 27% Moderate Access 20% Limited Access 11% Unconnected
Age	11% 15-17 26% 18-25 35% 26-35 32% 35+	7% 15-17 30% 18-25 37% 26-35 43% 35+
Gender	57% Male 43% Female	49% Male 51% Female
Occupation	40% within the education sector: 61% students 26% educators 13% school administrators 50% employed outside the education sector 10% Under/unemployed	33% within the education sector: 67% students 17% educators 16 school administrators 42% employed outside the education sector 25% Under/unemployed
Language Ability	18% High 82% Moderate-Low	7% High 93% Moderate-Low

IV. Key findings

Theme 1: General information seeking & learning

How do people look for information and learn?

Finding 1

The rapid growth of messaging apps has blurred the line between online and offline information-seeking and -consumption.

• The popularity of messaging apps (e.g. WhatsApp, Line, Telegram) has skyrocketed in recent years, due to various social, practical, and economic reasons. As of late 2016, WhatsApp was installed on 58% of smartphones in Indonesia, an increase of nearly 300% from the previous year. In 2015 Brazilian users of WhatsApp jumped from 38 million to 42 million in under two months.

  In Indonesia, it is common for people to have multiple messaging apps installed on their phone—each for connecting with a different social circle. For example, WhatsApp is mostly perceived for professional coordination and Line to connect with friends. While in Brazil, WhatsApp still reigns as the most common for people to connect across their multiple social networks.

  These apps enable people to pursue existing social behaviors—both Brazilian and Indonesian cultures are collectivist and highly social—with greater ease (i.e., access to individual family members and friends wherever they may be), greater efficiency (i.e., enables the rapid and widespread collection / dissemination of information), and at decreased cost (i.e., telcos often offer free or low-cost access to messaging apps to entice customers).

• Getting information via messaging apps is seen as equivalent to information passed by word-of-mouth—just faster and through a broader network. These apps are seen as a way to extend existing, offline social behaviors, and not necessarily by “being online”. This is especially common in those with low digital literacy and/or access, as they tend to primarily use the internet through messaging apps (which are a common entry point to internet usage).

• For many, using apps is often seen as different and distinct from using the internet. Especially those with limited digital literacy or internet access typically did not think of “using apps” (messaging and social media apps in particular) as “using the internet” (which is associated with using a web browser). Even those with moderate internet access and literacy did not always make the connection between using mobile apps and being online. The terms “internet” and “online” may be decreasing in relatability for such audiences, who simply refer to what they do by the name of the app they are using (i.e., “I’m Facebooking” or “I’m Zapping” (for WhatsApping) may have no mental connection with “I’m online” or “I’m using the internet”).

Finding 2

Small, hyper-targeted networks have become hugely popular, as it has become easier to create “groups” on messaging apps. Because it is so easy to do so, people are now forming such networks for diverse use-cases.

• By allowing people to organize their connections into “groups”, messaging apps have led to the proliferation of small, hyper-targeted networks for a broad range of personal and professional applications. The popularity of established social networks is waning—one 17-year-old Indonesian respondent said “Facebook is for old people”. People want greater control over what information they share (and what conversations they have) with whom. Messaging apps allow this control via customized, hyper-targeted networks for specific purposes.

  As a result, people are now establishing networks organized around specific topics (e.g. hyperlocal / neighborhood-level news, sports teams), types of relationships (e.g. family, colleagues, classmates) and/or shared experiences (e.g. participants of a specific event). Most people (beyond those with extremely limited internet access / literacy) belong to many messaging groups, which all represent different aspects of (or communities in) their lives, whether those communities are long-established (e.g. family) or temporary and transient (e.g. a semester-long class).

Opportunities

• Identify and communicate around highly relatable use-cases. Identify specific hyper-targeted networks whose purpose and composition may lend themselves to finding utility for the types of content Wikipedia provides. Consider using these high-likelihood, high-impact use cases to ground communications campaigns or materials in ways that people are increasingly finding and sharing information.
• Optimize Wikipedia content for sharing. Examine the types of content people are sharing in instant messaging groups, and consider making the ingredients of Wikipedia articles more modular and easily shareable, highlighting those ingredients with the attributes (format, design, etc) that lend themselves to sharing.

Finding 3

The rise of messaging-based networks has profoundly impacted how people gather, share, and determine the value of information in their daily lives. People are getting increasingly used to information coming to them (versus seeking it out), then validating it via app-hosted networks.

• Information—particularly that which is topical or sensational—tends to spread rapidly via these hyper-targeted networks, even if information dissemination wasn’t the intended purpose of these networks. People share all sorts of information via messaging groups, whether that which is related to the group’s purpose or otherwise. Because people belong to multiple groups, they may take information shared in one group and distribute it in other groups, so long as they perceive it to be of interest to recipient groups.
• The dynamics of messaging-group information-sharing has contributed to the sense that:
  1. The information being spread via messaging apps is urgently important. Because users receive push notifications from these apps, it contributes to the sense that the news being shared via these channels is urgent and important. Each time there is a new message, users get a notification—the study did not surface a single respondent that had these notifications turned off—and, as scientific research has shown, checking incoming messages leads to an increase in dopamine in the brain. This makes people literally feel good by checking information alerts pushed to them—and the physical sensation creates a positive feedback that sustains the habit.
  2. The information you need will come to you, so long as you are connected to the right people. Who you follow and are connected to is becoming just as—if not more, for some—important than the sites that you follow. There may be a growing trend towards using people as content navigators and curators.
  3. Information, most commonly news media, is true only when it has been validated through multiple sources, which is reinforced by growing mistrust in formal institutions, and the increased ease of digital search. This can be achieved by:
     • seeing the same facts in multiple places (which is easier for those with higher internet access / literacy, as they can search for a news story and corroborate it via different sources), or
     • seeing the same news story passed around in multiple groups to which you belong (which is more common for those with lower internet access / literacy).

Opportunities

• Help people follow “exciting” (e.g. articles getting uncommon editing traffic) or more interest-specific (e.g. a specific football team) Wikipedia content. Identify the types of edits to Wikipedia articles that people would want to “follow” and get alerts on (e.g. via a messaging app). These may be edits related to political figures, game scores, or news that may inform or help settle bar bets, etc. Help users create customized alerts for specific types of article updates (e.g. IFTTT-style, where it’s based on conditional statements) that then create the sense that Wikipedia is a highly relevant, live resource they would want to integrate into their regular communications stream and share updates with their peers.

Finding 4

People are increasingly getting new information primarily from individuals, and supplementing it with information from institutions.

• Information is now about getting it from the right people, rather than the right organizations. This has led to the sense that you should follow the right people—for those less digitally literate, via messaging groups; for those more digitally literate, in the form of specialist social media accounts—to get the information that you need. Information on websites is increasingly becoming a supplemental source of information—not to learn, but to verify or deepen knowledge.

  Those with high levels of digital literacy are able to build networks of peers (beyond those they know offline) to exchange information. They have developed the skills needed to discover and connect to content-producers they trust. In Brazil, for example, an entrepreneur chooses to follow industry bloggers because she respected the content they produced, but more importantly, because she could connect with them to ask specific questions. Those with lower digital literacy, on the other hand, may have more difficulty finding content specialists they trust beyond their personal networks and must rely on information that is passed to them through word-of-mouth, both on and offline.

• As a result, it is becoming increasingly rare, especially among younger and more digitally savvy users, to browse the internet or go to specific websites. When people—especially those younger and/or more digitally connected—do initiate searches for information, it is typically either i) to verify or learn more about something they have learned about from their peers, typically through messaging apps, or ii) to learn about something that is directly and immediately relevant to their personal needs (e.g how to do something, such as cook a specific dish).

Opportunities

• Appoint topic ambassadors / content curators to help people discover and navigate content on topics they are deeply passionate about. These content curators should be on the platforms that people are using to discover, discuss, and share content, rather than just being on Wikipedia. They can share their passionate for certain topics in a way that is more authentic and relatable (and aligned with how people currently want to receive information), and package bite-sized pieces of Wikipedia content and mix it with relevant content from other sources, for sharing in other platforms.

Finding 5

The value of information is increasingly assessed based on indicators relating to the individual producing the content, rather than the institution with which they are associated.

• The trend towards individually-based information gathering has led to new indicators of value—a blend of utility and reliability of content—for information producers. Ways of assessing value include:
  • Closeness of social connection: The closer to me they are, the more likely it is they know what I want / need (utility) and the more I can verify the content with our shared social connections (reliability)
  • Volume of social validation: The more followers they have, the more likely it is that they are providing content that is useful to other people (utility) and I will trust the wisdom of the crowd to determine that they are valid (reliability)
  • Intersection of closeness of connection and volume of validation: The more that it is my friends and peers that are validating a particular source (a blend of the above characteristics), the more I know this is a valuable source for me.
  • Independence of content-producer: When creditable professionals or “experts” (i.e. journalists or educators) self-produce content independent of institutions, I can trust it because this content is produced free of institutional bias and influence.

Finding 6

The unconnected have little control over their information diets and rely heavily on information disseminated through word-of-mouth and public intermediaries (e.g. television, radio, and public service institutions).

• Those that are unconnected are mostly pushed information with limited ways of seeking for the information they need beyond word-of-mouth. To reach the unconnected, social institutions are sending staff into communities to pass information about service offerings directly. This physical information exchange is effective, but can only reach a small part of the community. The most effective organizations have identified community influencers and leaders to pass information to—making the most of word-of-mouth model.
• As an extension of word-of-mouth, television and radio are still main information sources for those with low-no internet access. Often, those with limited access do not question the validity of these sources, rather they rely each to deliver both the practical and news information they need in their day-to-day lives. Those that rely on TV as a main information source tend to recognize and trust specific TV personalities before channels or media organizations. They will make sure to tune into specific educational or expose programing throughout the day to learn from these personalities.

Theme 2: Internet usage (web & mobile)

How do people use the internet?

Finding 7

Both Brazil and Indonesia are largely regarded as a mobile-first countries, with mobile penetration driving many digital behaviors and habits. For mobile-reliant internet users, their mental models and behaviors related to the internet are different from those that come from their computer-first peers.

• Mobile penetration in Indonesia and Brazil is extremely high. Mobile penetration is around 130% (meaning many people have more than one SIM card) in both Indonesia and Brazil. As of mid-2016, smartphone penetration was 47% in Indonesia and 57.8% in Brazil. Mobile phones are more affordable and easier to learn (and get support from peers on) than computers. As a result, many Indonesians—particular those that are not getting exposure to or opportunities to use computers at school or at work—are using the internet for the first time via mobile phones, and mobile phones may be their only connection to the internet.
• This, however, doesn’t necessarily mean internet usage is predominantly mobile-first... yet. It is difficult to determine just how much internet behavior is mobile-first or mobile-only at present—but it is clear that it is growing. While a 2014 Baidu study found that 55.2% respondents in Indonesia first experienced the internet with a computer, and an E.Life Survey in the same year estimated that 47% of Brazilians access the internet via computer, more recent studies suggest that mobile is a more prominent means. Both in the Baidu study and more recent ones, it seems that smartphones are becoming increasingly popular for accessing the internet. Statistics, however, differ for just how much mobile accounts for internet usage. In Indonesia for example, some studies show that 70% of web page views come from mobile (StatCounter, January 2016) while others reference 36% of internet users (Nielsen Indonesia, 2016).
• Part of the reason for imprecise statistics may be that people don’t always know they are “on the internet”. Those that are using messaging apps, for example, may not consider the usage as “being on the internet”. This is exacerbated by the fact that many mobile operators offer low-cost or free bundles for messaging apps, which can be interpreted by users as “free”.

  Brazilian mobile operators Claro and Oi offer zero-rated access to Facebook and Twitter to mobile data subscribers, as a way of winning customers. The operator TIM was the first to introduce a plan in which any files transferred on the WhatsApp platform didn’t incur data costs. As a result of these bundles, using “the internet”, by contrast, is largely perceived to cost money.

• Mobile-first users tend to be heavier users of apps, particularly those whose usage is discounted or free. This typically means messaging or social network apps. Over time, this impacts how they understand and are able to navigate the internet. This impacts how they like to recieve content (prepackaged and with specific UI or visual design preferences), and it may impact the degree to which they are comfortable navigating the internet (without a curated, directed experience like those provided by apps).

  Those with moderate to unlimited access to internet prioritize downloading and use apps that save them time. They choose apps that aggregate information for them (i.e. those that pull together multiple news sources) or serve a very specific function (i.e. banking or sports tracking apps). Cost focused users are more selective of apps they download. They are careful of not downloading (or keeping) apps that waste data, memory, or screen time (i.e. that are for browsing mindlessly: Instagram, Facebook).

Opportunities

• Be proactive and strategic, but discerning, about the implications of a mobile-first or mobile-led generation. There is a lot of discussion about a mobile-first generation, but many of the headlines and quantitative studies do not cover the specifics of what people do or don’t do on mobile, and what their preferences are. Before investing too heavily in an all-mobile strategy, organizations like WMF need a more granular level of understanding of habits, preferences, and trends relevant to mobile behavior, and to invest in heavy user testing around new product opportunities. For this work, coordinated research and testing among different functional teams interested in the opportunities (and limitations) of mobile may lead to more cost-effective work.

Finding 8

As mobile internet usage dominates, people’s level and range of internet access (and therefore digital literacy) is highly correlated with their income.

• The cost of mobile data is still relatively costly in Indonesia. Most Indonesian consumers buy pre-paid data bundles. A recent study found that an Indonesian worker being paid minimum wage would have to work just over 5 hours to afford 1 GB of data (which cost, on average, USD 2.84 in 2015). [Note: More data on mobile coverage, networks, and cost is available at this site, based on 2015 data.]

  In Brazil, government has been working with telecoms since 2014 to lower the price of data. But, with half the population being low income earners, data still continues to be costly for the average Brazilian. Some studies cite that it would take a minimum wage worker (about $1.05UDS/hr) about 34 hours to pay for a 500mb mobile data plan (about $35.80USD).

• Cost of mobile data is a barrier to getting connected for low income users. A recent PwC study found that mobile data costs need to decrease by 65% in Indonesia and 68% in Brazil for it to be generally affordable—a prepaid data plan that allows up to 500 MB of use a month, and costing 5% or less of a person’s gross monthly income—for currently unconnected populations.

  Low income users tend to buy data “pay as you go”. They purchase it when they can afford it, use it straightaway, and go without data for a period of time thereafter.

• For those that only have data intermediately, access to affordable wifi is key for getting online. In Brazil, low-income families are accessing “under the table” cost-sharing models which allow internet to be accessible in the home. Some models are as simple as sharing a router and password with a downstairs neighbor while others more complex with small scale entrepreneurs setting up neighborhood enterprises and selling cheap radio internet to neighbors. In Indonesia, having wireless internet at home is uncommon.
• While public wifi networks are growing in availability, these offerings are limited to urban centers and still largely inconvenient, defeating one of the key benefits of mobile internet: information that you want when you want it. To mitigate some of the financial burdens associated with getting online, lower- and middle-income Indonesians in urban areas commonly access internet through public wifi spots, which are increasingly common and openly available in public places, and the norm (though requiring a password) in many offices, schools, cafes, and restaurants and those with computers will often tether their phones. In more rural areas, this pattern is less common.

Opportunities

• Reduce the barrier of affordable internet through product and partnerships solutions. Work with institutions and local leaders to support access to content with or without the internet. For example, sponsored access through public wifi points or partnering with public libraries or universities to develop offline alternatives only needing periodic updates.
• Wikipedia Zero can play a role in alleviating the financial burdens associated with accessing information online. WMF should consider whether and how to integrate Zero into the Indonesian market.

Theme 3: Online information behaviors

How are people accessing, evaluating, and using information online?

Finding 9

Both Brazilians and Indonesians are highly comfortable with using translation tools and apps to make the most of their online experience.

• High-quality specific online content in Portuguese and Bahasa Indonesian can be difficult to find. Outside of national and local news and information, respondents found it challenging to find useful information in local languages they can use for school, work, or their personal interests. Respondents with high digital literacy commonly mentioned that “the more specific their question the more likely the answer would be in English.” As a result, those that spoke English sensed that that to get the most out of the internet, one must at least be comfortable with written English.
• People with limited written English proficiency do not see it as a barrier to using the internet in English. Generally, those with limited English ability browse the internet in their local language. When searching for specific information though, they may need to look to English results. Regardless of language ability, most users with medium digital access and confidence (or higher) were comfortable using translation tools (whether the browser based or app-based tools such as Google Translate).

  For Wikipedia, those with higher digital literacy and confidence figured out how to toggle between languages, although most users tended to go back to the search page and click on the English and local language search results to view articles in both languages.

• Those with above-average written English proficiency tend to browse internet in English, and access more content from international sources. This is seen as being more efficient and effective for getting the information one needs, as there is a greater volume of content in English.

Finding 10

When consuming text online, people tend to be quicker and more task-oriented (“scanning for what I need”) but they tend to consult multiple sources. When consuming visual content online, people tend to take more time and be discovery-oriented (“exploring for what I like”).

• People tend to proactively seek out and read text-heavy webpages and articles when they have specific questions they need answered. This may be when they are looking for specific information for school or work, supplementing information they have received via other channels, or just have a specific question they need answered. When looking for this information, they tend to scan the page until they come to what they came for, and to ignore the other context.

  As people gain digital literacy, they also gain control of their information diet—which counterintuitively opens up a wide information landscape. To navigate these complexities, they set up tools to cope with the wide availability of information.

• When searching for textual information, the stakes tend to be higher, since the information will be used to complete a specific task. As such, they are more likely to triangulate across multiple sources. The quality of information becomes more important, since the completion of the task (whether for school, work, or to add to an ongoing social conversation) will be judged by external sources, and often those in a position of authority. As a result, it is even more important to cross-check the information in multiple places. In doing so, the way of consuming content is still largely “scanning-based”.
• When not seeking to a certain task, people (skewing towards younger populations) are surf the internet via visually-led apps or content sources, e.g. Instagram or YouTube. They use the platforms’ Explore (Instagram) or Up Next and Recommend (YouTube) features to discover new content. It feels like a lighter, more fun way to discover new content.
• In Brazil, video, accessed through YouTube, has emerged as a prefered destination for students and how-to learners. Video is ideal because it converges both the depth of textual information and the appeal of visual content. Video consumers tend to search for specific topics on YouTube and are able evaluate videos more quickly than text. To do so they look—not just on the credibility of the source—but at the production value, charisma of presenter, and specificity of instruction.

Finding 11

Young people are increasingly defaulting to visually-led browsing. This is leading organizations to prioritize visual content and platforms in their strategies, even for activities that are typically not designed for such formats.

• People (skewing toward younger populations) are increasingly spending time browsing and sharing visual content online. As noted above, most of their time browsing is done using specific apps that prioritize visual content. The weight given to visual content is even impacting how they select apps to share visuals: some respondents said a key reason they preferred WhatsApp was its ability to share high-resolution images.
• Organizations are increasingly adapting to younger audiences’ preferences for visual content, even for materials or activities that have not traditionally lent themselves to visual presentations. News organizations are publishing content on Instagram, as some youth are using it as their primary news-gathering platform. (One 17-year-old respondent said, “Instagram gives just the right level of content” when explaining why he prefers following the Instagram feed of his favorite news site rather than going to the actual site. The feed featured largely photos of politicians with news summaries of less than 100 words, and typically a few different hashtags to promote discoverability.)

  NGOs are fundraising via Instagram, and even publishing basic financial audit information via the platform in efforts to show their transparency. This is both to attract young audiences, and recognizing their importance as early adopters in getting other generations in their networks (parents, younger siblings) online, and knowing that their behaviors will likely influence others. (This behavior, assumed by these organizations, was observed through field research in Indonesia and in other New Voices countries.)

Opportunities

• Get visual and distributed to meet users where they are. Wikipedia must consider how it needs to adapt its format, design, and distribution model to better meet the needs and habits of the next generation. The nature and extent of these changes can fall on a spectrum—from a fundamental redesign of the platform, to establishing “lite”, “junior”, or more visual versions of the platform, to simply incorporating more visual content—but this should be considered in ongoing platform design and content strategy conversations.

  The movement should consider how it might move away from thinking of the web platforms and properties as the central organizing body, but think of the way to fulfill the mission via third-party platforms whereby its platform is merely the central repository of content. (This, of course, is how many news and information platforms are thinking today, but Wikipedia’s unique design and production model may make this more challenging to execute in practice.)

Finding 12

The trustworthiness of a piece of information (or its source) doesn’t necessarily determine its utility. Young people especially assume that most information online is biased, and they adapt how they validate and/or use information accordingly.

• Many factors have contributed to a loss of faith in the trustworthiness of news sources. These include:
  • Indonesia and Brazil’s history of media control. Under the New Order (President Suharto’s military dictatorship from 1966–98), the media was tightly controlled. Since mass media liberalization started in 1999, the country has made significant strides in the diversity, quality, and professionalism of journalistic content in the country. The country’s media environment is “partly free” and it ranks 124 out of 180 countries in Reporters Without Borders’ World Press Freedom Index.

    Brazil has more of a spotty history of media control, with on- and off- periods of government censorship under the rule of President Getulio Vargas from 1930 to 1945, and from 1951 to 1954. De facto censorship was reinstated again during the military regime (1964–1985). Brazil ranks 103 out of 180 countries in Reporters Without Borders’ World Press Freedom Index. [Note: Other New Voices Global South countries rank as follows: Nigeria 122; India 136; Mexico 147.]

  • Widespread knowledge that ownership of Indonesia’s media is highly concentrated, with a few corporate oligarchs that promote their business and political interests via their media platforms, which are increasingly multi-channel. In this era, local news organizations have suffered and media empires have become increasingly powerful and centralized. This phenomenon is common globally but especially glaring in many emerging markets, including other countries studied for New Voices Global South countries (Brazil) and New Readers (Nigeria, Mexico).
  • Increased use of the internet for information-seeking and consumption, where any person or entity can put information online. Particularly for those that i) are not browsing in languages with high volume of digital content, and/or ii) are looking for information on highly specialized topics in their own language, the likelihood of them coming against low-quality, unverified information grows higher.
  • Increased prevalence of sensational and untrue content on the internet, as online business models are oriented towards gathering clicks at whatever cost. Combined with the previous factor, it has led to the sense that the internet can be a wasteland of junk content.
  • Growing popularity of the term “fake news”, which has increased since the 2016 US presidential election and has spread globally. Respondents used the term “fake news” to describe any content that was sensational, and the term seemed to be increasingly correlated with information found on the internet.
• Young people assume bias in news sources (online or offline) and online information sources in general—the degree to which they are comfortable with bias depends on the extent to which the bias serves their interests. When bias aligns with their worldview, it’s recognized but unquestioned. When the bias may not align with their worldview, they seek to discuss it within their networks (largely via messaging-app groups, where they are getting the news). Recognizing that news information can’t be trusted, they tend to scan the headlines—taking in many shallow data points—then fill in details of stories through peer conversations (both on/offline) with people they trust, as means of passively finding “truth.”
• As a result, the indicators of trust are changing, and becoming more individually focused. While older generations may have relied on institutional reputation to determine the validity of information being produced, young people now are looking for indicators of individual reputation. An individual’s number of followers, likes, shares and “personality” of the brand are indicators of a source's trustworthiness. Young people look to their networks to help them verify sources—each “like” represents the opinion of a person.
• People are looking for content that is representative local events, movements, and perspectives and see a gap in the availability of this content on the internet. This content not only makes people feel seen and valued—which in turn builds trust—but helps make the leap to understanding new knowledge. For example, when researching social movements on the website of international organization a respondent was better able to understand the Black Lives Matter movement in the US when he contextualized it against is knowledge of similar Brazilian movements.

Opportunities

• Allow customization of the article characteristics displayed to let discerning users judge for themselves. As people are getting increasingly confident in assessing online content, the indicators they use to do so differ. Wikipedia could consider revealing some of the metadata (e.g. number of edit(ors) and / or recent edits) behind different articles to give users greater visibility and control in how they assess content. This metadata may relate to data that Wikipedia already captures on articles (e.g. a “locked” article may signal greater trustworthiness) or new indicators that mimic what users are used to via other platforms (e.g. “likes” of an article).

Themes 4 & 5: Awareness and Usage of Wikipedia

What is awareness / perceptions of the Wikipedia brand? How does Wikipedia’s design, model, and/or programs currently support or inhibit online learning?

Finding 13

People tend to see Wikipedia as a finished product produced by a technology company, not a continuously growing resource and social movement led by a community and shepherded by a non-profit. As such, it is judged by the metrics by which they would judged other international tech giants.

• Although many people knew Wikipedia is an editable platform, they do not think of the actual content as adaptable and expandable. When people spoke about the content of Wikipedia, they tended to speak as if the content was static—not making the connection between how Wikipedia works and the state of the content. To them, the platform is seen as a final rather than an ever changing resource. As such, when they see mistakes on Wikipedia—one respondent recalled a time where she saw Lorem Ipsum text in a whole paragraph on an article—they tend to judge it harshly, noting that it is a sign of Wikipedia’s lack of professionalism.
• Most respondents believed that Wikipedia was run by a for-profit technology company—and one that was pretty boring and that lacked transparency. The inability to see the people and/or process behind the product negatively impacted people’s affinity for Wikipedia. Some respondents said that they liked flashy and innovative technology brands and organizations that were “transparent”, meaning that they could see the staff, processes, and even workplaces in which the product was made. One respondent referenced an article that talked about Google’s headquarters as why he trusted the brand.

Finding 14

Wikipedia’s open contribution model is poorly understood, and therefore viewed as a weakness.

• Many that know Wikipedia refer to it as an open source platform that anyone can edit, but they view the collaborative aspect of Wikipedia only in terms of the first stage of content contribution (the addition of content) and are unaware of the existence of possible, subsequent stages (the review, refinement, deletion or content, and the conversations that happen around open). Lack of understanding of the lifecycle of contributions and the benefits of Wikipedia’s model lead them to mistrust the platform. Most working in academia have no idea about the processes Wikimedia puts in place to increase the quality of content on its platforms.

Finding 15

People with any awareness of Wikipedia are using it for broad context or specific facts on people and places. As they learn that anyone can edit, they state distrust in the content, but still use it for the same purposes as well as verifying information found elsewhere on the web.

• Wikipedia is frequently described as a “dictionary”—it is useful for understanding what something is, and sometimes for understand what a word means. When googling nouns—typically people, places, things—people generally accept the Wikipedia page that comes up first in the search as a satisfactory definition of what they set out to find. This was the most common use of Wikipedia across all respondent groups.
• Those with low awareness of Wikipedia don’t demonstrate mistrust in Wikipedia’s content. They have few to no trust issues with the information that they find on the website. Those that have lightly used Wikipedia in the past and who have little awareness of its editing model or brand are quick to trust the content they find on the site.

  Levels of trust fluctuate for two reasons. First, those that have experienced errors in content either intentional errors (i.e. an Indonesian respondent referenced seeing a whole paragraph of just gibberish letters), or more commonly, out-of-date information (i.e. a telenovela plotline not up-to-date). Second, when they learn from others that the content is not trustworthy.

Finding 16

Although people may say they do not trust Wikipedia, they still find it useful—some even find it essential—in surfacing the information they need.

• Mistrust of Wikipedia is a learned perception—usually reinforced through higher education institutions. Those affiliated with institutions of higher education levels tended to be more more skeptical of Wikipedia, stating that professors would not accept Wikipedia as a source. Having an authority figure saying definitively that a source should not be trusted obviously impacts perceptions. But those with lower-levels of education (i.e. high school) had higher levels of trust in Wikipedia’s information often because their teachers told them that Wikipedia was an acceptable source. In one high school, researchers saw Grade 11 textbooks where Wikipedia articles were cited as footnotes for textbook content.
• And while many respondents said they cannot fully trust Wikipedia, this does not stop them from using the information. (This aligns with findings captured in previous themes that show that people don’t necessarily have to trust information to find it useful.) Especially skeptical audiences—typically, those that have a reason to be skeptical, e.g. students whose usage of information will be judged by academic standards—may not necessarily trust the information contained in Wikipedia articles, but find it a handy resources to get reference materials.

Opportunities (for several above findings)

• Reveal people and process to build trust. Wikipedia’s people and process is its greater asset, but it’s currently perceived as a liability. Consider how to reveal how Wikipedia is made in an easily accessible way to readers, and in ways that speak to readers’ growing mistrust in institutions, increased faith places in individuals, and other trends described in this research.

Finding 17

Even those who have used and trusted Wikipedia throughout their education, leave Wikipedia behind after completing their degrees. The value proposition beyond academia remains muddy.

• Many respondents were able to articulate how they have used Wikipedia in the past to support them throughout their education but were less clear about how Wikipedia could support them in their day-to-day life. This was influenced by the fact that Wikipedia is perceived to be an academic resource connected directly with their education. For example, a women in Indonesia who had used Wikipedia extensively for gathering background information for her university assignments didn’t know she could look up her favorite Korean drama star—only thinking that “serious people” could be found on the site.