It’s been a surreal few weeks across the globe thanks to the coronavirus pandemic. And it’s only getting surrealer. One reassuring side effect is the surge of resources for dealing with the virus and its impacts surfacing from various individuals, collectives, and organizations.

Produced in good faith by experts and non-experts alike, many of these resources are crowdsourced, allowing people in far-flung locations to collaborate in ways unheard of in previous global disasters.

As a human with a social anarchist bent, I find this type of self-organizing incredibly reassuring. As a digital experience designer, though, I find a lot of these resources horribly broken in their user experience. With the stakes as high as they are, making these tools usable and ethical is critical. So here are some thoughts based on my own experience as a content strategist and contributor to sonomafireinfo.org (now abandoned), an all-volunteer response to the 2017 Sonoma fires. I welcome input and amendments, particularly from those of you who have actively participated in the creation of these resources, or who have more formalized disaster response experience.

Jump to…

Common types of resources

Resources take a range of forms and each lend themselves to different purposes:

Type of resource Apps and other digital tooling Dashboards and data visualization Communities Link lists
Participants required 1+ 1+ 2+ 1+
Skill level needed Moderate – advanced Moderate – advanced Low Low
Ideal use case Direct service, task completion Status tracking, education, enabling decision-making Peer-to-peer communication, emotional support, real-time decision-making Peer-to-peer sharing, education, connecting people to things
Other notes Generally requires ongoing maintenance of codebase Can be easy to create with the right technical know-how but can easily misrepresent or mislead, which can cause real harm. Often formed around a specific topic, region, skill set, or specific resource from one of the other categories. Very easy to create; can easily become outdated and/or hard to navigate.

Regardless of category, the best resources spring from on-the-ground needs. Those resources that come from people or groups who aren’t directly impacted by a disaster can easily miss the mark and may prove either not useful, or not usable. That doesn’t mean they shouldn’t exist, but extra care needs to be taken if you’re creating resources for a group rather than as part of, or with, a group.

I focus primarily on link lists in this post because the tipping point for when these lists become unusable seems to come much earlier than the others. But those building other types of resources should find plenty here worth thinking about, especially under platform considerations, responsible maintenance, and sunsetting.

Do some homework first

I can’t stress this enough: don’t duplicate efforts. People in need don’t have time to sift through competing lists. It serves no one to compete for airwaves or bandwidth with existing resources.

But isn’t more choice better?

No. I’m guessing there’s a viable argument for why I’m wrong here, but I haven’t heard it yet. Maybe? I just read an opinion on Facebook that suggested anything that can get in front of people is useful, especially given that need is often localized/contextual. I’m not entirely convinced, but open to changing my mind. I’d love the opinion of someone officially working in disaster recovery.

A collection of screenshots showing the variety of disaster response resources out in the world.
The proliferation of digital resources for disaster response may make it harder for impacted people to find the specific help they really need in a crisis.

So how do you figure out if there’s already something out there that that you might join, leverage, or build on? It’s not easy, but you could always start with Google. Get creative with search terms, using various permutations of keyword phrases, and consider how a person seeking help might go about it. Odds are, you’ll discover at least a few related resource lists, and then dig. And dig. You’ll quickly discover just how chaotic and redundant the space can get.

Don’t overlook Twitter, Facebook, and other social channels. These serve as 24-hour news networks in times of disaster and popular projects often make their way around the tubes. Include place names, population traits, and/or disaster-related hashtags in your search. Finally, you might take a look through Data.gov’s Disaster portal, and the Civic Tech Guide and, for coronavirus or COVID-19-specific projects, the Coronavirus Tech Handbook.

When you find similar efforts

I get it: your idea is ever so slightly different than what’s out there. You can do it way better. No one can find the one that already exists, whereas you’re sure everyone will flock to yours.

You may be right about any or all of those things. But you still should think twice about rolling your own. Ask yourself:

  • How can you lend your time to improve what already exists?
  • How can you help people in need connect to the existing resource(s)?
  • Can you make a greater impact on more people by supporting a direct service organization on the ground with your cold, hard cash?

There is almost always a viable (read: very good) answer to each of these questions. But you have to be honest with yourself. If your goal really is to help, think about how you can strengthen what’s already out there, rather than adding to the cacophony. If you’ve decided at this point that it indeed makes sense to use your humanpower to bolster or expand an existing resource, then go forth and build that bridge. You’d be surprised by how much more you can accomplish by pooling your energy, knowledge, and/or money.

If you’ve found a real gap that needs filling and you still want to proceed, then please—I beg you—take some time to think through the following issues.

Platform considerations

Your skill set

We’ll always gravitate toward the tools we know, and this is good and logical. But keep in mind that what you know how to build may not be the most appropriate solution for the problem at hand. Weigh your capabilities against the need. If the only language or platform or technology you know how to use just isn’t appropriate to solve the problem, don’t bother trying to jam a square peg into that round hole.

Collaborators’ skill sets

If you hope to enlist others, think about what skills will be needed and how common they are. This is particularly important if you hope to have impacted communities contribute to and/or manage the resource (see Community Context below for more).

Time to launch

How quickly can you get something up and running? In a fast-moving crisis, time is one of the biggest constraints. Disasters won’t wait for you to work out bugs or learn a new platform. Ergo, see #1.

Speed of updates

How quickly can updates be made to the resource using the technology at hand? Will others be able to jump in and help, or does the platform require a highly specialized skill set? See #2.

Stability

Generally speaking, a crisis is not the time to test-drive experimental new technology. This is particularly true if your resource addresses a high-stakes need, such DIY face masks. What’s at stake if the tech fails? On the other hand, if someone can’t edit a Google Doc to add a link to a blog post because people are overwhelming the system, it’s unlikely that someone’s life will be directly on the line.

Portability

How easy will it be if you need to switch platforms or move your data? Of course, this also needs to be weighed against how likely it is that you’ll need to do this.

User considerations

There are a number of platform considerations that bleed over into user considerations. Ignoring these can seriously hinder your efforts to support those in need. Just ask Shadow, the makers of the 2020 Iowa caucus app that failed horribly in the field. When we say “user” here, I of course mean those impacted by disaster or crisis. These are humans, many of whom may be in the most vulnerable situation of their lives. There’s a scale of impact, certainly: people transitioning to remote work or shelter-in-place orders are dealing with very different stressors than those whose high-risk loved one is exhibiting symptoms of COVID-19. Add to this existing inequities experienced by various communities, and it becomes very easy to create unintended consequences that do more harm than good.

Community context

I know this isn’t a popular thing to say but technologists have a terrible reputation for making all sorts of assumptions about the communities they build for (see my earlier comment about building with). Communities traditionally impacted by inequity (and the cultural and physical violence that accompanies it) have very real reasons not to trust people from outside their community who bring offers to “help.” This means your wicked great app for, say, connecting displaced people with emergency shelter may be ignored by those who have a legitimate reason to think their documentation may be questioned on arrival.

How do you prevent this? You don’t do it by coming up with a bright idea, building it, and expecting them to jump on board (even if hundreds of other people in your cultural circle, or who are like you, do). You can work to head off unintended consequences by asking the impacted community you want to help how you can be of service to them.

You can also develop a set of principles to inform the project, such as the Principles for an Equitable and Effective Crisis Response. “Principles create conceptual boundaries that enable people to agree upon what should or shouldn’t happen,” says Greg Bloom, and organizer who leads the Open Referral Initiative. “They are our most fundamental working agreements—setting out ‘how are we going to do what we’re going to do?'” This is why West Oakland Air Quality, a project I lead with OpenOakland in collaboration with the West Oakland Environmental Indicators Project, includes the 17 Principles of Environmental Justice in our GitHub ReadMe doc.

Familiarity

You’ve previously given some thought to how familiar you and your collaborators are with the given technology you want to use, but what about the people you’re actually building for? Does your solution require complex manipulation of data (or even basic understanding of reading charts and graphs)? Does it make use of platforms many people from all walks of life already make use of, like Facebook?  The goal should always be to make your resource easy to use. Otherwise, it’s not particularly helpful in a time of crisis.

Mobility of users

Coronavirus is a somewhat unique natural disaster, in that it doesn’t necessarily force people out of their homes (in fact, it’s forced many of us—myself included—to shelter-in-place). But fire, flood, earthquakes, and other disasters typically displace people. And diaspora is a real challenge to first responders trying to disseminate information and resources. So think about how people will find out about your resource if they’re bouncing from shelter to shelter.

Low connectivity/power cases

Similarly, many disasters cause damage to infrastructure that may leave populations without internet or cellular access. This was the case in Puerto Rico, for example, when Hurricane Maria left thousands without cell service. If your tool requires the cloud to function, think hard about how practical it will be when in use on the ground.

Responsible maintenance

Perhaps the biggest challenge facing creators of disaster response resources is keeping the resource up to date. In fast-moving crises, situations change dynamically and captured information quickly becomes outdated. Wrong information can cost lives if it misinforms or keeps people from the help they need.

What needs updating

There are some common types of information that change rapidly:

  • Official instructions and status updates
    Recent disasters like the Sonoma and Paradise fires in California have led to massive changes in how local authorities coordinate with each other and communicate those efforts out to impacted citizens but local governments aren’t always great at communicating out consistently.
  • Open/closed status of shelters, drop-off points, and other locations
    This includes specific hours a resource can be accessed in addition to whether or not it’s actively functioning.
  • Running lists of donation or volunteer needs
    Needs change rapidly based on on-the-ground logistics and the wrong information can at best distract distribution centers and coordinators and at worst, delay or otherwise block their work.
  • URLs
    Links change as online resources merge, move, or disappear altogether. If someone needing help keeps running into dead-ends, it causes more distress and undermines your goal of connecting people to resources.

How to keep information updated

If your resource relies on accurate information to be useful, you have an obligation to carefully think about how you’ll keep this information updated. When people are in precarious conditions that impact their safety, relying on inaccurate information can be deadly. This isn’t an exaggeration, which is why I’m repeating it. If you send someone to a shelter in the midst of a fire and it’s been closed and evacuated, you could make the situation worse for them.

There are three basic levels of updates:

  1. Low-touch
    A lot of folks who spin something up on their own rely on people to submit updates until a more sustainable system emerges. Including a form where visitors can send updates is helpful here. Bonus points for keeping a link to the form in close physical proximity to any items likely to need frequent updates (yes, repeating the link next to every resource item in a large collection is a good thing).
  2. Medium-touch
    A more proactive way to keep information updated is to use direct phone and email outreach to the sources themselves. This is a large part of what I did during the 2017 Sonoma fires to ensure the information we were providing was accurate. This can be time-consuming, though, and those channels can be unreliable in a disaster. An alternative (or supplement) is to actively scan channels like Twitter for updates from those on the ground. But again, relying on information that isn’t obtained directly from the source can lead you to disseminating inaccurate or false information. Use with caution. Evaluate the veracity of your sources.
  3. High-touch
    The most accurate way of keeping information updated is by relying on contacts out on the streets to verify locations and needs. Obviously, this requires a lot of people power. It can also be risky during natural disasters or other situations where on-the-ground safety is iffy.

Inevitably, you’ll need to use some combination of the above approaches. Thinking through the implications of each will help you plan ahead to set up lines of communication for gathering the information you’re trying to disseminate.

Messaging and content usability

People under pressure or in shock can’t process as much as they can under normal circumstances. It’s especially critical that your content accommodates this state of mind. The following recommendations reflect the content best practices most applicable to disaster-response resources.

Be transparent.

In times of crisis, people need to be able to quickly and easily evaluate the credibility of information sources. Always identify yourself and do it visibly. In most cases, you should include a noticeable disclaimer that the resource is not from an official entity.

Keep URLs short.

You want the destination to be memorable. You also want to leave room for valuable messaging when sharing the URL through character-constrained channels like Twitter.

Put the most important information first.

This applies to the page, within each section, and in headlines. This should almost always be the term a disaster victim is most likely to use when searching for help, or some variation of it.

Be clear.

Use plain language. If you can include or link to language translations and/or versions formatted for assistive tech, do so.

Be brief.

Use shorter sentences. Only provide the most important information. If you want to include additional context or deeper reflections do it elsewhere and simply link to it, so those who want that level of detail can find it. Trying to include the kitchen sink will get in the way of the critical information people need most.

Make scanning easy.

Long paragraphs and dense text is hard to read, particularly if you’re stressed out and in a hurry. Keep text sizes large, and break content into chunks. Ensure your design meets standard color contrast requirements for users with low vision or who may be viewing screens in less-than-ideal lighting conditions.

Use boldface and color sparingly, to create landmarks on the page so readers’ eyes naturally gravitate to the most critical info. When everything is boldface or color coded, the screen gets too busy and readers can’t figure out where to look first.

Break long lists into smaller categories or topics, based on how those needing aid are likely to seek it out. Consider adding a table of contents at the top with quick links to each section. Don’t forget to add a link within each section so readers can quickly return to the table of contents.

Order categories logically. Usually—but not always—this means alphabetically. In disaster situations, it may make more sense to list categories by urgency (e.g. Government Announcements, Safety Maps, Shelter Locations, Volunteer Opportunities…). Within each category, consider the implications of listing items alphabetically vs. reverse chronologically (e.g. most recently updated at the top).

Make searching easy.

Once a list reaches a certain size (generally a couple of pages; though that’s completely anecdotal), people start to need a keyword search option.

Screenshot of a list of example volunteer projects.
With no categorization or tagging in place, a reader has to scroll through endless pages in search of a relevant item on helpwithcovid.com.

Your approach is highly dependent on your platform and how well you’ve thought through the content you’re producing. Taking search into account from the get-go can make your resource infinitely more usable. If your platform allows tagging, create a controlled vocabulary with which to tag each item. You’ll almost certainly need to expand this list as the crisis evolves, but you want to keep it controlled and not open (i.e. anyone can create a new tag), to prevent competing synonyms that might land an item tagged with one word but not its synonym. We don’t want a disaster victim selecting free food and getting 1-2 results, while there are 9-10 others tagged with free meals that they’ll never see.

Include some indication of the last time content was updated.

If you’re updating in regular intervals, an unobtrusive sentence stating “Updated every hour on the hour,” or “Last updated on {date} at {time and time zone} will do. If items are being updated piecemeal or on the fly, include “Updated: {date} at {time and time zone}” as a parenthetical to each item.

Limit text-heavy graphics and video and if you do use them, make sure they’re accessible.

Video can take up major bandwidth and in a disaster, infrastructure may not be able to support playback, rendering content unusable and people in need roadblocked. Text-heavy graphics keep content from people with visual impairments, assistive tech, and people who don’t speak the language. If you insist on using video or text-heavy graphics, be sure to include text captions or transcripts when distributing.

Include contact information

If your resource is useful, people will want to contact you. Be prepared to field inquiries about the following:

  • Specific items or data you’ve included in your resource, even if they came from outside sources. People will expect you to be able to answer questions you probably can’t.
  • Updates and additions they feel should be included. You may or may not agree with them, but they will come in.
  • Journalists may reach out wanting to reference the resource in their coverage of the disaster, or to get in touch with a source on the list.
  • Other queries already answered in the resource itself. Be patient with these, as folks often miss seemingly obvious things when they’re stressed out or in a rush. Remember that some of these folks may be disaster victims who haven’t slept or eaten recently, or who are scared and in shock. Be gentle and kind even when you’ve answered the same question 23 times and it’s plainly listed on the site.

Sunsetting

This one is easy to overlook, given the urgency with which disaster response technology is often created. Some quick tips and questions to answer:

  • Have a plan for what you will do with the resource once the immediate need has passed.
  • For how long will content be maintained and updated?
  • Where will the code/native files be hosted and for how long?
  • What disclaimers do you need to include to let people know the project is no longer maintained?
  • If you intend to take your resource offline altogether, will there be a public record anywhere?

For more ideas about life-cycle planning, check out FabRider’s tips for network-centric resources.

Final thoughts

At some point, we’re going to move from triaging the spread of this coronavirus to managing it. That won’t be happening anytime, soon, though. This means we’re going to desperately need these tools and resources—and new ones, too. And there will certainly be additional additional crises around the globe on top of it. But as the tech community continues to charge ahead to help (a good thing, if done intentionally),  it needs to take very seriously its obligation to do so responsibly.

This list is not exhaustive by any means and while I’m an expert in content, I’m no expert in disaster response. I welcome revisions—particularly from those working within the disaster response and recovery field. You can leave a comment below or contact me directly.

This post was last updated on Mar. 19, 2020.

Stay tuned in…

My email update is where I tend to share early explorations and thoughts on how we can design digital experiences that work for real people, while contributing to healthy organizations and systems. It’s safely occasional (like a sundae). Join us!