Mappers take responsibility for what they’ve built
The union push at Mapbox is based on the idea that “maps are political.”
You ever wake up to find that the kid you went to high school with has suddenly become, like, governor?
That’s the feeling I’ve been having these days with Mapbox. I first got introduced to the location-data company when it was still something of a side project for Development Seed, a start-up running out of an old under-heated garage in an alley off of 14th Street here in D.C. Memories are fuzzy, but I think I first met Eric Gundersen and Bonnie Bogle, co-founders of Mapbox, at a conference in Chile back in 2010. Drinks, karaoke, and maybe a few games of “Big Buck Hunter” were involved? I kept up with Gundersen for a time and we’d talk about the power of mapping, with him saying things like “mapping has always been expensive, and the dudes that had the maps had the money.”
Now there’s talk of Mapbox going public at a valuation of $2 billion. And this week a majority of employees of the company declared that they intend to form a union, arguing that “maps are political.”
It’s an interesting story of tech workers — particularly tech workers at a platform company — being eager to take responsibility for what they’re unleashing into the world.
(Note: At some point “MapBox,” with the mixed capitalization, became “Mapbox.” For clarity and because I don’t know when exactly that happened, I’m sticking with Mapbox here.)
The roots of modern mapping
Development Seed was born at an exciting moment in the evolution of digital mapping. The world’s great shift online was producing a ton of “data exhaust” from web searches, mobile-phone tracking, and more. Plus people were beginning to volunteer giant caches of data through things like social media posts, reviews on sites like Yelp, and wikis.
That much of that data could be tied to specific locations in the offline world opened up exciting possibilities. One big moment was the creation of a project called HouseMaps.org, which mashed together Craigslist apartment listings and Google maps to make it easier for users to search street by street for a new place to live. It was a revelation.
With the new ease of creating maps, everyone started creating maps for just about everything. Excitingly, these were live maps, and collaborative. This was the time of OpenStreetMap, a wiki-based challenge to the maps that were owned by the British government that started when one person, by the name of Steve Coast, decided to survey his Regent Park neighbored, posted the results online, and asked others to do the same. There was a map to plot the aftermath of the devastating earthquake in Haiti and one to connecting citizens affected by wildfires in Russia so they didn’t have to wait for government help. There was the “iPhone Tracker” that revealed visually what your mobile device had recorded about your travels to and fro. A project called Ushaidi tracked post-election violence in Kenya. The Public Laboratory for Open Technology and Science set people up to plot out the BP oil spill from the Deepwater Horizon oil rig in the Gulf of Mexico, and a map called If It Were My Home compared its spread to neighborhoods, towns, and cities.
Even the Obama White House was on trend.
In the fall of 2011, the White House posted an “Excess Federal Buildings Interactive Map.” It plotted more than 7,000 federal structures across the whole of the United States, in a bid to build political momentum behind coming up with a way of getting rid of some of them.
Because no one person had to populate any of the maps, they could reflect truths that no one person or organization would be incentivized enough to capture. Internet theorist Clay Shirky talks about this as the Coasean floor, riffing off the work of the British economist Ronald Coase. The idea: that pre-Internet, there were projects that existing companies wouldn’t take on because the managerial effort required outweighed the benefits. But, the thinking goes, the Internet and the cheap tools it made available to users lowered the costs so that ad hoc groups would find it worthwhile enough to contribute some piece to such efforts, whether it’s collectively attempting to drive help to struggling communities or, as Shirky worried, perpetuating a virtual pile-on of someone who’s caused some offense.
But back to maps. The existing big map players, like Google, had no reason to plot a place like Sudan because they couldn’t sell advertising there. So for much of the world, it was an open playing field.
“This mapping thing is about to blow up,” Gundersen told me at the time, in 2011.
Mapbox was born into politics
Development Seed launched into that mix, based on the idea that, as Gundersen told me at the time, “Whenever you need a common operating picture, you need a map.” Development Seed worked in places that were in turmoil, as an attempt to move past the giant paper maps that organizations like USAID would tack up on the wall when some spot on the map was experiencing crisis. The company could instead put in the hands of aid workers zoomable “slippy maps” on iPads or other devices, constantly updated with real-time information. The maps were built on the back of the volunteer-run OpenStreetMaps, overlaid with all sorts of data, including things like satellite imagery from NASA.
The company would work with the National Democratic Institute to pull out of PDFs and then plot results from independent election watchers in the course of elections in Afghanistan, with New America to map drone strikes ordered by President Bush and President Obama on Pakistan’s tribal regions, and with the UN on ReliefWeb to collect and plot first-hand reports of humanitarian aid around the world. There were explicit political aims here: to ensure fair elections in Afghanistan, hold the U.S. accountable for drone strikes in Pakistan, and better distribute development aid in underserved places around the world.
And Development Seed was also part of a broader conversation about the risks of mapping, like plotting the neighborhoods of Japan’s outcast Burakumin people. And activists using tools from the Public Laboratory for Open Technology and Science to map the city of Lima, Peru, declined to add the results to OpenStreetMap over worries that it would give the Peruvian government more leverage over disadvantaged neighborhoods.
Around 2012, the folks at Development Seed realized that the platform they were using in places like Nigeria, Congo, and Pakistan could be used in more corporate settings, including in the U.S. Mapbox gained so much momentum inside the company that it began to be spun out as a standalone entity.
The effort got a real boost when Google announced it was going to start charging for use of its embeddable maps once they gathered more than 25,000 daily views, which drove Foursquare into the arms of Mapbox. Google, too, lacked the crowd-driven element that helped fuel Mapbox, though it made a bid to add some of it by buying Waze. At about that time, Mapbox had only 30 or so employees, still working out of the humble garage in D.C., which the Economist described by saying that “too many computers switched on at the same time has been known to fry fuses.”
Things changed. In 2017, the Japanese company SoftBank led an $164 investment round in Mapbox. The Wall Street Journal characterized it as a “sizable bet.” But it made some sense, because SoftBank, and the $100 billion “Vision Fund” it had put together, was eager to win on emerging technologies like ride-sharing and place-based social sharing undergirded by location data; both Lyft and Snapchat were Mapbox customers.
It wasn’t all smooth sailing. For example, in 2018 a contributor to OpenStreetMap labeled New York City “Jewtropolis,” which auto-populated on Mapbox-powered maps used by everyone from Citi Bike to the New York Times. Mapbox didn’t go the route of pointing out that that was other people’s data, simply hosted on its platform. It scrambled and apologized.
And it kept growing. By 2019, the firm CB Insights said Mapbox was on track to become a “unicorn,” valued at $1 billion or more. In March of this year, Gundersen left the CEO role and became chief strategy officer and chairman of the board. The CEO became Peter Sirota, who had been with Amazon Web Services.
The union push
All of which brings us fairly current, to where there’s talk of SoftBank thinking of taking Mapbox public in a SPAC deal that would, yes, value the company at $2 billion.
And that’s giving Mapbox’s employees – which the company counts at about 460 — some leverage. (Employee discontent isn’t particularly attractive to investors.) And the workers are using some of that advantage in a bid to create a union under the auspices of the Communications Workers of America. Bloomberg reported this week that just about two-thirds of Mapbox’s employees based in the U.S. have signed union cards.
What’s particularly interesting here is the framing of the organizing push. Of course, there are employee quality-of-life issues on the table. But organizers are also putting the drive in the explicit context of wanting to take ownership over the technology they are putting out into the universe.
"More and more often we build beautiful maps and tooling only to find out they are being used for politically dubious or nefarious purposes,” reads the statement from the group. “We are seeking complete transparency into how customers use our services. Additionally, we will reserve the individual right to refuse to work on any projects that conflict with a worker’s ethics and to do so without fear of reprisal; as well as the ability to deny service to any customer completely if put to a majority vote by the workers.”
(While there are certainly class issues involved, compare this to the lead organizer in the driven to form a union in an Amazon warehouse in Alabama saying the impetus there was that workers were “hurting and confused.”)
That’s where this becomes about more than simply one set of tech workers forming a union. It becomes about a company and the people who drive it signing their name to the idea that what they build, they have to answer for.
“Mapbox technology is changing the world,” reads the statement from union organizers. “As the workers who bring this technology to life, we take responsibility for how it is used.”
LINKING AROUND
Now that he’s moving off the columnist beat, the New York Times’ Frank Bruni has second thoughts about his tone: “There aren’t as many clicks in cooling tempers and complicating people’s understanding of situations as there are in stoking their rage.”
Facebook is asking its Oversight Board for an advisory opinion — its first — on how residential information can be shared. (This is something of a one-way street: Facebook doesn’t have to abide by the recommendations the board volunteers.)
And MacKenzie Scott is giving away a ton of money, but at $2.74 billion, still a fraction of her net worth —a good reminder of the breathtaking disparity in what counts as real money in the tech world she comes from and the arts/advocacy worlds she’s pouring cash into.
Fascinating recounting of Mapbox's history. Hang around long enough and some of your friends become really successful, right? I'm curious to learn more about Eric and Bonnie's personal visions for their work and now their fortunes, and also about the leadership of the emerging union, which seems to have learned something from tech worker organizing at other/bigger tech companies.