Forty kilometers down the South African coast from Cape Town, tucked behind a narrow strip of mountains, sits Masiphumelele, a crowded township of 40,000 people. A drive through “Masi,” as it is known locally, reveals dusty turf dotted with modest, single story homes set closely but neatly together. Most of Masi’s residents live here, within the township’s official boundaries.

The landscape abruptly changes at the township’s northern end, at Masemola Road, which ends before a dense crush of metal shacks 30 to 40 structures deep in a low-lying area known as “the Wetlands.” The Wetlands is not marked on any official maps, yet the section houses a quarter of Masi’s residents. No paved roads or walkways penetrate the mass of wood and corrugated iron shelters. Many Wetlands residents are connected to power but must weave through the narrow crevices separating the shacks to access water taps and toilets. The walk can be dangerous, and the unmistakable stench of human waste hovering over the settlement suggests that many avoid it as much as possible, especially women and children, and especially at night.

The realities of Masi’s so-called “informal residents” are hard to reconcile with the hopeful vision captured by the township’s name, which means “we will succeed” in the regional Xhosa language. Yet it was in that spirit that Priscilla Gala, a Wetlands resident and community organizer, brought two members of the Cape Town social start-up Lumkani and several of its volunteers into a block of new shacks to install bright blue fire detectors and talk to residents about the risks of fire.

“If you hear it go ‘beeeeeeep’ that means there is umlilo in your home,” explained Lumkani’s co-founder, David Gluckman, to one young woman and her son, using the Xhosa word for “fire.” “If you hear it go ‘beep beep beep beep,’ that means the fire is close, at your neighbor’s.” The young woman nodded knowingly. Four months before, she, Gala and 300 of their immediate neighbors lost their homes in a fire that ripped through their Wetlands block. It left about 600 people homeless.

Lumkani’s mission is to prevent the devastation wrought by fires like the one that hit Masi last November. Its solution is a networked detection system that alerts people to fire by measuring how quickly heat rises in the home. The system then communicates the presence of fire over radio frequency to sister devices installed in homes nearby. When installed at a certain density—say, one device every 20 meters—the little blue detectors serve as a fire warning system for whole communities.

Fires are a common event in South African informal settlements and others in major cities across the Global South. More than 860 million people worldwide live in informal settlements, most of which resemble Masiphumelele: dense concentrations of flimsy shelters that cling to the peripheries of low-income urban neighborhoods. Slums grow quickly and haphazardly as new economic migrants arrive and build slapdash homes from wood, metal, and other locally available materials. Neighbors often end up living a couple meters from one another, thus an accident with a candle or cookstove in one home can quickly become life-threatening to hundreds of people.

Fire ranks consistently among South African shack dwellers’ top concerns, alongside crime, flooding, and access to water and sanitation, according to research from the Community Organisation Resource Centre (CORC), a national NGO that helps informal settlement communities mobilize resources to improve living conditions and security. Since the end of 2015, fires destroyed 4,000 homes, displaced more than 12,000 people and caused over 150 deaths in South Africa, according to local media reports.

One particularly devastating incident set Lumkani into motion. On New Year’s Day in 2013, three fires tore through informal settlements in South Africa's largest township, Khayelitsha, leaving more than 5,000 people homeless. Emily Vining, who helped found Lumkani, recalls her devastation upon reading the news. “Here we were celebrating the [New Year’s] holiday and 5,000 people were starting the year with nothing,” she says.

At that time, Vining’s friend, an engineer named Francois Petousis, happened to be working on fire detection technology for his graduate studies at the University of Cape Town (UCT). Vining impressed upon him the need to apply his work outside of the lab. “I told him, ‘You have to do something with your research’,” she says. “And I committed to helping him find funding if he agreed to take it forward.”

Nearly four years since that conversation, Lumkani has installed its fire detectors in more than 8,000 homes in informal settlements in South Africa and Namibia. As of November 2016, the start-up witnessed more than 20 fires strike areas where its detectors are installed, resulting in only 29 homes lost since its alarms were put in.

Technically-minded

Petousis, Gluckman, Vining and the rest of Lumkani’s early team admit that they did not have a clear idea of how or where to start in 2013. Only one member—Samuel Ginsberg, a senior engineering lecturer at UCT—had developed and launched a commercial product before. Slum fires were an issue that none of them had any personal experience with beyond local headlines, however. Vining recognized early on that the only way to learn how to tackle the issue was to get involved with the communities that deal with the threat of it daily.

“I didn’t want [us] to be another organization using traditional top-down development practices,” Vining recalls. “I believe the only approach to making a valuable technology is through deep engagement with end users.”

Fire ranks among the top concerns of South Africa’s informal settlement residents, along with crime, flooding, and access to water and sanitation.

This approach, called user co-design or human-centered design, is considered best practice within the field of development engineering. Nevertheless, it is not always instinctive to technically-minded engineers and development practitioners whose life experiences are often far removed from the problems they are trying to solve. Petousis says that he counts himself in that category—or at least he did when he began his work.

“The truth is, it feels inappropriate for me to be developing solutions [like this],” he says. Growing up in South Africa, Petousis has been a daily witness to deep social and economic inequality. He, like many South Africans, describes the country as an extreme example of haves and have-nots, wrought by decades of racial segregation under the apartheid government in the 20th century. Today, the majority black population continues to face disproportionate levels of poverty, unemployment, and crime, as well as poorer access to services and quality education.

As an engineer, Petousis admits that he questions whether he is equipped to solve social problems. “I live in the Cape Town that is, generally speaking, very privileged. I can drive 20 minutes into [an informal settlement] and often do, but I haven’t spent more than 48 hours at a time there,” he says. “What do I know about life there [or] the right solutions to community challenges?”

When Petousis began researching slum fire prevention under Ginsberg, he was encouraged to approach the issue through an engineering lens. “[Ginsberg] believed there should be a simple technical solution to solving that problem,” Petousis explains. They both quickly realized that there was not. Most of the available literature on fire detection focused on U.S.-based systems, where home fire awareness and prevention is commonplace, unlike in South Africa. This literature pointed to smoke as the best indicator of fire—an idea that was quickly dismissed by a local NGO, which pointed out that informal settlements are smoky environments as residents often cook over open fires.

“From our early [informal settlement visits], we realized that if we used smoke detectors, they would be sounding all the time,” Petousis says.

Recognizing that they would need more local input, Petousis began researching human-centered design approaches, while he and Ginsberg refocused their technical efforts on heat detection. They ultimately improvised a novel, low-cost solution that differs from standard heat detection systems in that it does not use a silicone-based thermal resistor. (Lumkani has now patented its solution but declined to elaborate on technical specifications because the team is in the process of making design modifications.) The pair then commenced product testing with the South African Bureau of Standards and other independent assessors. Nevertheless, as designed, the solution had no pathway out of the lab and into the informal settlements.

Community direction

Engineering and design literature encouraging user engagement may be instructive to the technically minded, but the success of it comes down to relationships. In this respect, Petousis found himself short of a starting place. Cape Town alone has dozens of informal settlements, which are demographically and politically diverse. For an outsider with no direct ties, it would be impossible to know which community, if any, would be receptive to partnering on a project.

For guidance, Vining reached out to CORC. CORC is visible throughout South Africa, supporting community leadership in mobilizing resources and funding for development initiatives, like installing public toilets or building safe public spaces. “All of the work we do is driven by the leadership within each community,” says Thandeka Tshabalala, CORC’s technical and livelihoods coordinator. (The organization’s motto is “Nothing for us without us.”)

CORC is also one of the few organizations that conducts demographic surveying in South Africa’s informal settlements and collects data on the communities worst affected by slum fires. The organization agreed to help Lumkani’s early team find a community partner to develop the fire detector. In early 2014, the leaders of UT Gardens, a small settlement of about 400 homes in Khayelitsha, volunteered for the task.

Twice a week for eight months, Lumkani’s growing team met with CORC and UT Gardens’ 15-member steering committee to discuss the community’s experience and concerns with fire. “Through those conversations, it became clear that just having [a warning device] for one household would not be meaningful,” Vining explains. “A shack can burn in two minutes, and after that [the fire] gets beyond what one person can control.” Community members wanted a system that would automatically alert neighbors and fire responders in the event of a fire. Lumkani’s household surveys of 70 UT Gardens residents validated this. This feedback ultimately informed the development of Lumkani’s signature feature: the device’s networking capability.

Even with UT Gardens’ input, it took Lumkani several attempts to arrive at a satisfactory design. “We were still too technically-minded about it at first,” Petousis concedes. During the design process, Lumkani’s team prioritized alerting neighbors and getting people to safety, since fire spreads quickly in informal settlements and many slum dwellings do not have easy access to roads or water, limiting the effectiveness of emergency responders. The most feasible solution was a community alarm that would communicate with individual household devices and broadcast the presence of fire from a central loudspeaker.

“It wasn’t a sleek solution. Communities would have to wait for us to install this central device before the alert network could work,” Petousis explains. “We felt that it would create this uncomfortable power dynamic where communities wouldn’t be empowered to bring in this system themselves—they would have to wait on us.” The team scrapped the idea.

An impromptu brainstorming session on Gluckman’s porch inspired the idea of networking the fire alarms to one another. Petousis says it was as if all of the team’s deep involvement with the UT Gardens suddenly gelled: “We just realized as we were sitting there that morning that we knew what we needed to do.”

The team did months of testing on sensor sensitivity, sound level, radio frequency, and networking bugs—mostly within their own homes to avoid straining community relations with too many false alarms. By November 2014, the first 250 devices were ready to launch in UT Gardens, and within two weeks, the system was put to the test when an outdoor cooking fire spread to a nearby home. The alarms alerted neighbors, who were able to contain the fire. Only one home was lost and no one was injured. Just over a year later, those same pilot devices alerted the UT Gardens residents of a much more serious incident—an arson attack that consumed 13 shacks before it could be extinguished but resulted in no loss of life.

Awakenings

Embracing the community partnership model enabled Lumkani to build a more effective and readily accepted product than what its team of engineers could have devised on their own. It also presented its own substantial set of challenges, particularly around Lumkani’s early business model.

Lumkani’s team recognized early on that cost would be a significant factor in the success of its product, owing to the limited financial means of its customers. Partnering with CORC alleviated some of the pressure, because CORC agreed to cover 80 percent of the devices’ cost in UT Gardens. UT Gardens’ leadership assumed responsibility for collecting the remaining 20 percent from residents—about 20 South African rand (US$1.30) per person.

This is a standard arrangement for all of CORC’s Cape Town initiatives, Tshabalala explains. Recognizing that informal settlements have limited financing options for costly community initiatives, CORC has a dedicated fund that pays for 80 to 90 percent of its supported projects. For Lumkani, this arrangement promised to help its devices achieve critical mass quickly, which is important for the success of all new products, but particularly one reliant on density to serve its purpose. Indeed, CORC financially backed the installation of 1,600 devices in Lumkani’s early community rollouts, while also supporting Lumkani by coordinating awareness campaigns through CORC’s national network, orchestrating informational sessions in the communities, and helping community leadership galvanize resources to invest in the fire detection system.

But this model has also proven to be slow moving for a young company that has its sights set on eventual profitability. “Our role as an NGO is to be a facilitator. We wait for communities to approach us [about a project] before getting involved, and that takes time,” Tshabalala explains. For CORC, this approach is very much in line with their mission to help informal settlement communities help themselves by allowing them to determine their own priorities. But Tshabalala acknowledges that projects rarely progress speedily. “Sometimes, from the first time [Lumkani] presents to a community, it takes three months before they get a proper response, because the community leadership needs to explain the project to [the other residents] and why this is something they should contribute money to,” she says.

New initiatives can be a tough sell financially, even when there is demonstrable need and funding support, Tshabalala says. The 20 rand each household pays for a Lumkani detector, for instance, is not an insignificant sum for many slum residents in Cape Town, where an average day’s work pays under 200 rand and steady employment is rare. Gluckman concedes that the challenging part of getting several hundred community members on board is that it demands that people weigh future risk against immediate daily necessities, like food and communal taxi fares.

“The cost of a fire detector is perceived as expensive to someone who doesn’t see the risk of fire today,” he says. Nevertheless, in the communities Lumkani serves, shack fires rank consistently among their top concerns, he adds.

If the logical reasoning is complicated, community politics are often more so, which has offered valuable if sometimes painful guidance for Lumkani’s implementation approach. “People are dealing with difficult issues in the communities we work in,” explains Clive Nqiwa, who took over for Vining as Lumkani’s field coordinator in 2015. “Some are able to pull together and work in such unity. [In others], sometimes it’s necessary to work around people who have a different agenda.”

Nqiwa worked alongside Gala throughout the entire 1,700-device rollout in Masiphumelele in early 2016. Both he and Gluckman marvel at her effectiveness as a community organizer. “I had always been told, if you want to go into Masi, you have to speak to Priscilla,” Gluckman says.

Masi had been on Lumkani’s radar since the company began, because a variety of factors including patchy electricity connectivity, lack of water taps, poor road access and a windy geographic setting render the community especially vulnerable to fire. When the team was finally introduced to Gala in 2016, she expressed that without financial backing form an organization like CORC, the community would not be able to self-finance fire detectors. Lumkani offered to launch a crowdfunding campaign on Indigogo to cover the costs. Gala then arranged a meeting for Lumkani with Wetlands community members and turned it over to Lumkani to make their presentation.

“She didn’t say a word in the meeting—she just let everyone else speak and ask questions,” Gluckman recalls. He says he believes it is her unimposing, quiet leadership that has made her so effective in mobilizing projects and garnering community trust. “She’s repeatedly been involved in good development projects and has created the right relationships, so she is able to make things happen quickly and seamlessly.”

The success of Lumkani’s original model hinged on getting hundreds of members of each community to prioritize future risk over immediate daily needs.

Gluckman says he recognizes that fully-funded projects are a much easier sell than ones requiring financial contribution from community members. Nevertheless, Lumkani’s partnership with Gala and Masiphumelele stands out as a collaborative example among others that have been more complicated, and even outright contentious at times.

Fortunately for Lumkani, the team has yet to have a partnership fall apart; nevertheless, they do not want their success to hinge on their ability to skillfully navigate community politics. There will inevitably be communities where fire poses a significant threat, but where there are too many political obstacles for Lumkani to work effectively, Nqiwa says. “We can’t get involved in those issues,” he adds. “We have to protect our image and our company, and part of that is recognizing that not every relationship is going to work.”

With that in mind, Lumkani continues to pursue community-level partnerships with organizations like CORC and the Red Cross, but it has shifted its core business model. The company now focuses on leveraging its technology and the user and community data its devices collect to encourage private sector players and government agencies to offer new products and services to informal settlement residents. Fire insurance and emergency response are two examples.

The approach, Gluckman says, is just a different path toward achieving the same mission.