FluPortal.org

A Review of FluPortal and Recommendations for the Future

View the the full report or download a .pdf version.
1. Summary
2. Technology and Design
3. Blogging and Curation
4. Outreach and Communication
5. Appendix

Summary

by Rekha Murthy

FluPortal was launched in early Fall 2009 as a content and technology resource for stations and other public media outlets covering the H1N1 flu pandemic. The project was funded by CPB and led by PRX in collaboration with NPR. The project’s main goal was to support a coordinated, informed, and timely public media response to what was expected to be a major health crisis in the United States and around the world. It was also intended to lay the groundwork for timely and effective deployment of future crisis-related public media coverage, both on-air and online. This document shares materials and lessons learned from the FluPortal project toward that end.

The Site

At the project’s core was the creation of a portal providing public media outlets a central access point for information related to the H1N1 pandemic. A publicly accessible website was determined to be the best way to satisfy this goal. FluPortal.org aggregated and curated technical tools and best practices, reputable government and non-government articles and resources about the pandemic, and relevant public media coverage. While the website was public, it was targeted to the needs of the public media community.

Within the public media community, we identified two important audiences for the project and shaped the site around their needs: 1) news directors, editors, reporters, and producers planning and covering local outbreaks of the swine flu, and 2) web managers tasked with updating station websites with pandemic information.

As a result, the website was organized into four main sections:

• For Station Websites — Aimed at station web managers, this page highlighted a collection of free web tools and online resources to support coverage of H1N1 on local station websites.
• Reporter and Producer Resources — Collected H1N1 information, links, and resources to assist public media editorial staffs in their pandemic coverage.
• Public Media Coverage — Cataloged H1N1 reporting by public media outlets throughout the system.
• FluPortal Blog — A running conversation of ideas, information, and trends related to public media coverage of swine flu. The blog regularly highlighted the work of reporters, editors, public health professionals, and technologists through interviews and guest posts.

We implemented on FluPortal many of the same widgets we were recommending to stations, including a display of the NPR API, a blend of government public health information feeds created with Yahoo Pipes, and a Delicious widget displaying the latest public media and other H1N1 links curated by the FluPortal team. Josh Andrews’ report provides greater detail on the process of designing and developing the site.

A cornerstone of the grant was the expectation that data related to swine flu — particularly numbers and locations of cases — would be gathered by the U.S. government and made freely available in structured form. The launch of Data.gov in Spring 2009 “to increase public access to high value, machine readable datasets generated by the Executive Branch of the Federal Government” encouraged this expectation. So we had initially envisioned a major role for FluPortal involving the collection, management, and visualization of H1N1 data.

However, extensive research at the project’s outset revealed few reputable datasets, and none at the national level. Some states and localities provided specific numbers of H1N1 cases, and a few projects such as HealthMap aggregated World Health Organization alerts, news reports, and personal accounts to provide a general picture of the pandemic. But a comprehensive, structured, and regularly updated dataset never became available. In fact, in July 2009, WHO announced that testing and counting individual cases was too burdensome for healthcare providers and not essential for the tracking of the pandemic. In August 2009, the Centers for Disease Control and Prevention stopped counting specific cases of H1N1 in favor of weekly estimates of influenza-like illness as reported by state health departments via a new nationwide surveillance system (more info on the CDC website). That became the only data FluPortal displayed, in the form of an interactive map developed by NPR, regularly updated by their multimedia team, and embeddable by stations on their websites.

Team

As with many websites, the structure of the core FluPortal team paralleled to some extent the structure of the website:

• Project manager: Rekha Murthy — Communicate with funder and project partners, plan project development and outreach, support team members as needed.
• Technical advisor, developer, and blogger: Josh Andrews — Develop and manage the WordPress site, write and maintain the For Station Websites technology tools section and the Public Media Coverage section, write occasional blog posts on tech tools for crisis coverage, assist station staff in implementing tools on their own websites.
• Blogger and editor: Katherine Bidwell — Manage the site’s content, especially the Blog and the Reporter and Program Resources section, establish blogging practices and blog daily, reach out to potential guest bloggers and interviewees, track news of the pandemic to determine direction of site and blog, create a curated feed of notable public media and other pandemic-related links using the Delicious bookmarking tool.
• Outreach coordinator: Ken Mills — Spread awareness of the site among public radio, public television, public health, and journalist and community organizations, gather feedback on the site from target users via email and phone interviews and guide team in implementing the feedback, attend industry conferences, organize weekly team meeting, assist in webinar planning.
• Everyone — Write blog posts, review each other’s work, share ideas for site content, contribute to Delicious feed, present at project webinars, spread the word about the site.

In addition to the core team were a few other light, occasional resources. These included grant management and administration by two senior staff, and and about 30 hours of technology resources (outside of Josh Andrews) to purchase domain names, communicate with PRX’s ISP to set up the WordPress instance, support Josh in maintenance and administration of the WordPress site, and set up the two webinars that were held.

Overall, we found that this team structure was well suited to the project’s needs and worked well for the individual members. Roles were clear, and any overlap proved fruitful; for example, Ken Mills, the outreach coordinator, assisted Katherine Bidwell, the blogger, in making contact with desired guest bloggers, and when his outreach efforts turned up interesting people he referred them to Katherine for a blog interview. The FluPortal team worked remotely, staying in touch using free or low-cost services including email, various instant messaging clients, Skype, Free Conference Call, Google Voice, and the occasional in-person meeting. Online, we used Google Docs and PBworks to collaborate and document our work. Please see the team members’ reports for additional areas of collaboration.

That said, there sometimes seemed to be one role missing: an H1N1 or health/science domain expert. The grant called for the site to curate existing data and information rather than identify new sources and new angles on the H1N1 pandemic. However, good curation relies on being able to evaluate content and sources for reliability and accuracy, especially in order to provide diverse material beyond government and major mainstream media like the New York Times. That in turn requires health and science expertise our team didn’t have.

Not only would such an expert have enhanced our ability to provide good information, they could also have improved outreach by knowing the major players — individuals and organizations — in the fields of health and science information. While NPR did provide some advisory support from its Health Desk and Multimedia teams, we would recommend a more regular/dedicated resource for future such projects. Health crises, like other crises, are prone to a flood of information, which can easily become dominated by redundancy and even misinformation. A site like FluPortal can play a stronger role in combating such tendencies by truth-squadding what’s circulating as well as surfacing deeper research and new content. So including a domain expert would also require a shift in mission for future projects. Please see Katherine’s report for more thoughts on this issue.

Capacity

All team members were part-time. The project manager dedicated an average of 10 hours per week, varying widely over the course of the project. The other team members each worked about 20 hours per week on the project. This was more than sufficient given the waning course of the flu; however, if the pandemic had become more serious, more time would have been needed for all roles. The key challenge for future crisis-related projects is creating a fixed resource structure for a dynamic situation. Crises from pandemics to natural disasters vary widely in duration and intensity. For H1N1, we had advance warning, and it developed slowly over the course of many months. While its course was not easily predicted, and turned out to be less extreme than feared, other kinds of crises could be even less predictable. Staffing agreements and structures need to account for this fact.

Another consideration is when to bring on various roles. We strongly recommend a phase of discovery and outreach to begin well before a site is launched. Advance outreach can identify real needs and desires from the target audience while simultaneously raising awareness and buy-in. While future teams should not hesitate to continuously improve and update a site after launch, we did learn that first impressions do count: people formulate their understanding of a site on their first visit. With that in mind, it’s far better to meet a substantial portion of user needs and expectations out of the gate, rather than promote a site too early and then try to attract attention for major improvements later on.

Gathering and Displaying Information

As the team members’ reports address in detail, much of the FluPortal project involved gathering and displaying information. In the absence of reliable national-level datasets (e.g., on swine flu cases or vaccine availability), information meant audio, video, text articles, and blog posts by and about reputable government, non-goverment, public media, and mainstream media sources.

For the Reporter and Program Resources section, information was selected that would help editorial staffs quickly get up to speed on the topic, with lists of major organizations and government agencies, fact sheets, and public media “explainer” reports. While FluPortal’s ability to provide local information was understandably limited, we did offer ideas and resources for approaching state and local coverage. The National Center for Media Engagement provided tips for engaging the local community as part of a coverage plan.

For the For Station Websites section, two major categories of information were included: 1) H1N1-related video, audio, widgets, badges, and other content that stations could embed on their own sites, and 2) a series of guides written by the FluPortal team to using various free web tools to add dynamic H1N1 information to station sites and enhance stations’ social media presences in service of keeping their communities informed and getting story ideas. Josh Andrews’ report goes into this section in more detail.

For the Public Media Coverage section, FluPortal addressed the ongoing public media system problem of knowledge sharing. System members — stations and other outlets — don’t easily know what other members are doing. The team searched for and invited submissions of stations’ H1N1 topic or “landing” web pages and posted them on FluPortal. In addition, the team bookmarked swine flu stories by public media outlets using the Delicious tool, and displayed those bookmarks as an RSS widget in this section. Later in the project, NPR and the FluPortal team worked with third-party content aggregator Daylife to create a custom, automated feed that displayed H1N1 stories pulled from about 30 public media station feeds.

For the FluPortal Blog, the team not only highlighted new additions and features on the other sections, it created original content by interviewing a wide range of public media and public health professionals on approaches to analyzing and covering the pandemic. Katherine Bidwell’s report describes the blogging criteria and process in greater detail.

Clearly, information organization and display was a major challenge for FluPortal, as it would be for any such future sites. The top issue was presenting large amounts of useful information in a way that was readable, navigable, and quickly accessed by people on deadlines. In his report, Ken Mills notes that information density was indeed an issue for some users of the site. Reporters on deadline don’t have time to sift through large amounts of information, and some felt the site was overwhelming. Josh Andrews reports on some of the ways the team addressed this in different site sections. FluPortal wanted to set itself apart from the surfeit of sites offering superficial and repetitive H1N1 information, but to do so effectively, additional design research and innovation is needed.

Another related issue was balancing static, “evergreen” information display with timely material. This arose most often with the blog, which is time-linear by design. However, many blog posts remain relevant now and for future use, despite being buried by newer posts. We tried to counteract the burial problem by linking to evergreen blog posts from the static sections of the site and from newer, related posts as well. In all, information organization for a site with elements of various time sensitivity is an interesting problem we’d encourage future projects to explore further.

Outreach and Measuring a Site’s Impact

It comes as no surprise that outreach is critical to a site’s success, especially if a site is created around a specific topic for a temporary duration. We began awareness-raising to the public media community as soon as the site launched, working with partner NPR and funder CPB plus mailing lists, direct email, and industry publications to get the word out. The CPB-funded National Center for Media Engagement was very effective in sending both public radio and public television stations to the site and encouraging webinar attendance. We held two webinars using GotoWebinar to explain the project and invite feedback and content submissions. The first webinar, in the early fall when the pandemic was on the rise, was much better attended than the second, in January 2010, when the flu had waned. Ken Mills attended two conferences which were essential opportunities to improve and assess awareness and attitudes toward the project. Ken made a large number of direct email and phone contacts as well — please see his report for more detail and recommendations.

One known challenge for all public media projects is the lack of an authoritative and comprehensive directory of public media staff. In the absence of this, an outreach role must be staffed with someone who already knows the public media system well, as Ken does. Our Appendix lists organizations in public radio and public television that can help get the word out to their members. We cannot emphasize enough the need for a centralized, current public media contact database. This would also enable better targeting of the people most likely to use and benefit from the project. FluPortal’s intended users were not necessarily program and news directors and general managers, who are easier to track down; they were editors, reporters, and website staffers who are less so.

In addition to public media, the team did extensive outreach to government agencies including the Department of Health and Human Services, the Center for Disease Control and Prevention, and the National Archives. We contacted non-government public health, science, community, and journalist groups including the World Health Organization, the Association for Healthcare Journalists, the Harvard School of Public Health, the Mayo Clinic, the Public Library of Science, United Way, and many more. Please see the Appendix.

As Ken discusses in his report, awareness raising isn’t enough. An effective outreach effort for FluPortal also involved one-on-one contacts to recommend specific sections of the site and invite feedback. These were accomplished by scheduling in-depth conversations with individuals at stations and other organizations. In the case of the technical tools and widgets, it involved recommending specific tools and supporting individual station web managers in the implementation of FluPortal’s offerings. By inviting interviews and guest posts for the blog, we also got across the message that FluPortal was sincere in its information- and advice-sharing missions.

Formal user testing was an area we steered clear of out of concern that this would turn the focus away from the site’s substantive content to feedback on aesthetics like color, graphics, and layout. Ultimately, however, design influences adoption nearly as much as content. For future projects, we would strongly recommend interviews that walk users through the site, crafted in a way that acknowledges the importance of design but gets into content issues as well. While we did conduct ongoing outreach and solicited feedback and regularly folded the results into site improvements, approaching this more systematically and earlier on could make the team’s focus clearer from the beginning. For example, we were unsure whether our public media bookmarks feed was getting noticed and used, or whether blog readers would like to see more story ideas, more government resources, or both. More site-focused user interviews might also have yielded ideas on how to better organize the abundance of information on the site.

For a site like FluPortal, measuring impact is complex. Unique visits is not sufficient, given the specificity — and therefore limited size — of the target audience. Indicators of the site’s reach and impact include:

• Google Analytics — visits, time on site, bounce rate
• Direct outreach efforts via email, phone calls, in-person conversations
• Webinar attendance
• Emails via the site’s Contact Us form
• Blog comments
• Mentions in media including other sites and blogs
• Mentions on and traffic referrals from Twitter
• Appearance of FluPortal widgets on station sites

A few findings: Average time on site was high at 4 minutes, a good indicator of engagement. The For Station Websites (technical tools) section got the most unique visits after the home page, with the Blog and Reporter Resources nearly tied in next place. FluPortal did get mentioned in various media, including the Current newsletter, the Center for Social Media’s blog, the Association of Health Care Journalists’ blog, Daylife’s blog, and the Nieman Foundation’s Guide to Covering Pandemic Flu. FluPortal.org and specific blog posts appeared regularly on Twitter, which was the top referring site after Google (the National Center for Media Engagement’s site came in third). Please see the team member reports for more detail on various additional indicators.

As for station participation in FluPortal, we found that direct contact was the most successful. Station, program, and non-pubmedia staff were responsive to direct outreach efforts. Some guest-blogged or agreed to be interviewed, presented at our webinars, and even shared their H1N1 landing-page stats with the team. We received a few dozen emails telling us about various H1N1 programs and initiatives at various stations. Our first webinar, held in October when the pandemic was still growing, was very well attended, with 75 attendees. See Ken Mills’ report for more detail.

In contrast, online participation was virtually nil: only a handful of blog comments, and only two signups to a Google Group that was made private to public media and public health professionals only. Some suggested that a FluPortal Facebook page or Twitter account would have encouraged more participation. We chose to avoid these so as not to cultivate an identity separate from stations’ own offerings, since the site was intended to support stations, not be a destination in itself. Getting public media people to communicate with one another without significant facilitation is a challenge not unique to FluPortal. One reason might be the limited time and resources with which many in this field operate. Having a dedicated outreach coordinator to make direct contact is essential in this situation. However, we’d recommend future projects continue to explore ways to encourage broader participation among our time-stretched public media community.

What Stations Want for FluPortal and Future Sites

Public radio and public television stations in the United States vary greatly, and so do their needs. It’s impossible to create a one-size-fits-all service that everyone will want and use. Yet our active efforts to gather feedback and contributions did give us a general sense of what worked and what worked less, as well as a better understanding of the issues important to consider for future such projects. Please see the team members’ reports for additional insights.

The success of the For Station Websites section demonstrates the desire among stations for more easy-to-use technical resources. Not only was that section the most visited, dozens of station web managers told us they used that section extensively to build their own H1N1 landing pages. For the Public Media Coverage section, more than a dozen stations contacted us directly to let us know about website landing pages and broadcast specials, which we posted there. All four main sections had low bounce rates and high average visit durations, good signs of engagement with the material.

As noted repeatedly in this report, we confirmed our initial understanding that station staff at all levels are too busy to devote deep attention to a complex website. In other words, if people can’t understand a site’s purpose and offerings quickly, they will move on. Awareness of FluPortal.org was high among both public radio and public television people, though fewer had actually visited the site, and even fewer said they used it significantly. Generally, they correctly understood the site’s purpose, although one challenge from the outset was to make clear that FluPortal was not a destination for stations to link to, but rather a resource for them to make themselves online and on-air destinations for swine flu information. A few people did say that the breadth of the site’s mission and offerings was a turnoff, that they would have preferred a smaller, more focused offering.

Throughout the project’s duration, Ken Mills and the rest of the team asked station staff what kinds of content they most wanted to see. For example, we learned that while stations were somewhat less interested in knowing how other pubmedia stations were covering the pandemic, they were often more interested in story/angle ideas and feedback on their own work. Since many station people considered H1N1 to be a national and international story, frequently station people asked for ways to localize coverage. FluPortal addressed this by posting advice and ideas from station staff and other public media organizations.

The top concern of stations was the assurance that information posted on FluPortal.org was from a trusted source. This feedback led the team to post brief biographies on the About page so our identities and qualifications were known.

In our many discussions with public radio and public television program directors, general managers, news directors, and digital staff, an interesting picture emerged that can inform future crisis site efforts. The grant for FluPortal was based on an assumption — or aspiration — that public media stations view their mission as going beyond news gathering and broadcasting and into the role of conveners in their communities. WDET in Detroit is an example of such a station: it reached out to local healthcare providers and community groups and hosted community meetings about the swine flu pandemic. This was the kind of station we had in mind when we worked with the National Center for Media Engagement to share advice on community engagement and get the word out about FluPortal.

But many stations, particularly public television stations, don’t seek to be conveners. For them, could FluPortal still be useful? Yes. WAMU, a public radio station, is not a community convener but has a robust local news-gathering operation and website. Its team made ample use of FluPortal, particularly the web guides.

While many stations have designed their websites to be news and community information destinations, many simply post program schedules and contribution and contact information. This difference in approach is greatly influenced by the wide variance in financial and staff resources available to each station. While no website like FluPortal can fit every station’s needs, awareness of the differences can influence the site’s content and outreach strategy.

Key Successes, Challenges, and Open Questions

Was FluPortal ultimately a success? In many ways, yes. We set out to create a website that would invite wide and deep usage as well as discussion and feedback about future sites. In the process, we achieved a very good model upon which future sites can build. FluPortal showed what is possible with a small but dedicated team to aggregate materials and raise system awareness around a specific topic in a relatively short span of time. On the technology side, we found ways around the public media system’s lack of skilled technologists by focusing our technical guides on tools that non-technologists can implement, too. These guides were written specifically for H1N1 and crisis-related sites, but are also easily applied to many other kinds of websites. They were very well received by station web managers. Widgets by our team and NPR are of high quality and can be quickly repurposed for future needs.

Editorially, even without a reporting mandate, the FluPortal blog offered daily unique content pointing readers to a wide variety of angles and resources. Our outreach efforts raised awareness of the site and revealed a desire for such resources from some stations, though they also revealed real limits to stations’ abilities to deeply engage with such a site. Traffic was not as high as we would have liked, but NPR Health Editor Joe Neel noted it would likely have increased dramatically had the pandemic become more serious. As with nearly every public media project, FluPortal emphasized the need for an up-to-date, comprehensive directory of public media staff. A centralized public media announcement mechanism would also aid short-term sites like FluPortal in raising awareness quickly and in notifying stations when they’re ending.

The communication challenge also suggests the need for one destination site, a “CrisisPortal” perhaps, that stations know to turn to every time a major crisis occurs. Such a site would also address the issue of resource allocation. Crises are by nature impossible to predict, and as the pandemic started to wane, our team was challenged to find a new direction for the site for the duration of the grant period. If future sites are situation-specific, like FluPortal, we would strongly recommend a more flexible funding and resource model. Alternatively, a consolidated CrisisPortal could allow a dedicated team to distribute their efforts across situations in accordance with their shifting intensity.

We did anticipate the temporary nature of a pandemic by creating, where possible, materials with future applicability — the WordPress build, the technical guides, the blog best practices, as well as some of the posts themselves. We have also provided a list of government and non-government organizations that could be of use for pandemic flu and beyond. FluPortal’s post-pandemic relevance raises the importance of archiving the work created during such projects. We are grateful to the National Center of Media Engagement, which will host FluPortal.org as an archive indefinitely following the project’s conclusion on March 31, 2010.

Overall, FluPortal became a part of the growing discussions about crisis response both inside and outside of public media. Key players in these discussions are projects like the CPB-funded Station Action for Emergency Readiness (SAFER), Ushahidi, and CrisisCamp.

In the (somewhat) final analysis, does a site like FluPortal have a place in public media? We, and many of the stations we spoke with, think it does. This is particularly true if a central destination site is developed to avoid duplication of effort and information across crises. Such a site could be an important node in an overall effort to get the system to collaborate, to pool resources, and to exchange ideas and information. Crises come in many kinds and time frames. During many crises, the ability to activate public media’s deep and broad presence across the United States to put out quality, relevant information can be of tremendous value to our communities. This is what our public service mission is about.

Team Reports

The next sections of this document are reports by the individual team members: Josh Andrews on the technical aspects of the site, Katherine Bidwell on the editorial aspects, and Ken Mills on outreach.