In case you didn’t realize, we live in a multiverse1 of global street maps.
It all started back in the early 1980s in the offices of two startups who were both based in Sunnyvale, California. One was Etak, the original pioneer of in-vehicle navigation systems. The other was a little company called Karlin & Collins.
In the case of Etak, founder Stan Honey and angel investor Nolan Bushnell had the vision of building a ground breaking navigation system. Stan knew they had a large number of hard problems to solve in order for the system to be successful, and the need for a digital street map was only one of them. In the very early days Etak somewhat naively thought, “Maps? That’s the easy part — we’ll just get those from the government!”
Their assumption wasn’t totally lacking judgement.
It turns out that in 1965, almost twenty years before Etak was founded, the US Census Bureau had made the case for building a digital street map of the USA in support of the 1970 census. They called it GBF-DIME. The Bureau was a visionary of its time, realizing that such a map could not only be used in support of tabulating the national census, but it could also be used in many other areas, including education, transportation planning, emergency services and urban planning2.
Unfortunately Etak quickly came to realize that these US Census Bureau ‘stick’ maps didn’t quite meet the requirements of a navigation system. The data contained little information about curvature of the roads and highways were barely digitized. The quality of road connectivity — technically known as its topological correctness — left a lot to be desired. It was this hard reality that became the catalyst for Etak to get into the digital map business.
The other Sunnyvale start-up, Karlin & Collins, got its start in 1985 not because they’d invented a James Bond like navigation system like Etak, but because one of their founders, Galen Collins, had got lost driving in the Bay Area. Collins also saw the value in navigation, but focused on a much harder problem: developing a system that could provide turn-by-turn directions in addition to the map based guidance that the Etak Navigator provided. Collins’ desire for turn-by-turn directions added another whole level to the requirements — not only did you need to collect all the road geometry and street addresses, but now you also had to collect information about turn-restrictions and one ways. GBF-DIME definitely didn’t have that!
While Etak and Karlin & Collins discussed cooperating a number of times they rapidly became competitors.
Both companies realized ‘data was king’ and both companies started digitizing — the same cities, the same neighborhoods, the same streets, the same addresses. Not only in north America, but in Europe and Asia. All at huge expense.
Time moved on.
Etak was sold to Rupert Murdoch’s News Corporation who later sold it to Sony, who sold it to Tele Atlas. Tele Atlas was acquired by TomTom.
Karlin & Collins went through a series of rebrands, first to Navigation Technologies, then to NAVTEQ and later to HERE. HERE is now privately held by a number of corporate investors including BMW, Audi, Mercedes, Mitsubishi, Intel, Bosch and NTT.
Both TomTom and HERE built a global map database. Both developed a successful business licensing map data to automotive OEMs. With the advent of the internet, they also licensed their data for use on the web, initially to a fledgling web mapping company called MapQuest3. The MapQuest site relied solely on map data licensed from third parties. This ultimately became an opportunity for Google. They swooped in, launching Google Maps in 2005. MapQuest was left to atrophy by its new parent, AOL, who remained completely distracted by its acquisition of TimeWarner.
But even Google had to rely on third parties for digital map data. Google initially chose NAVTEQ as their primary source and later switched to Tele Atlas. But on October 7, 2009 Google made what was to be a very significant change. Hidden in their announcement of their new “report a problem” feature for Google Maps was a move that would send shock waves across the mapping industry. Google had dropped the use of Tele Atlas data in the USA and had replaced it with their own map. This thus became the genesis for a third global street map. Today Google Maps continues to maintain a map of the entire planet, albeit relying on the foundation of many third party datasets.
Meanwhile in 2004, around the same time that Google Maps got it start, an English gentleman named Steve Coast became frustrated with the UK’s national mapping agency, the Ordnance Survey. Unlike the US government, who released all their geospatial data for free with no license restrictions, the Ordnance Survey insisted on (significant) license fees. In response Steve launched the OpenStreetMap (OSM) project and two years later established the OSM Foundation. It got off to a slow start, but a few key catalysts gave it the momentum it needed:
- Contributions from organizations like AND (now Geojunxion), who provided some basic road networks; integration of US Census Bureau street maps
- Access to aerial imagery, providing the necessary backdrop for map editing, initially provided by Yahoo! and later Microsoft Bing Maps
- Money. In 2012 Google made the decision to start charging for access to its Google Maps API. The original catalyst for OSM was the lack of free access to good map data. Google’s move to add a paywall to its APIs added fuel to the fire. It precipitated moves by the likes of Foursquare, Wikipedia and AllTrails to switch from Google Maps to OSM. Many others followed.
- ‘Paid Editing’ — whereby corporations funded enormous numbers of edits to OSM. This started in 2017 and started to mushroom in 2019:
I’m not sure if Steve Coast’s original aspiration was to develop the ‘Wikipedia of Maps’, but it certainly turned out that way. Thanks to all the hard work of its millions of contributors, today OSM provides a beautifully rich global map:
Some nine years after Google launched its own map of the US in 2009, Apple Maps embarked on a similar journey, launching their home grown map of northern California in September 2018. Today Apple Maps has extended its coverage to many countries, including the US, Canada, most of western Europe, Australia, New Zealand, Israel and Saudi Arabia4. Filling in the gaps with OSM and other third party data Apple effectively has its own global street map5.
But it doesn’t stop there. In the enterprise mapping world Esri has not been standing still. Esri created a global atlas which they call the “ArcGIS Living Atlas of the World”. Just like a paper atlas it contains many maps. And included in the list of maps is a highly curated global street map built from collating and aggregating numerous sources from around the world. The purpose behind this map is a little different from the other players. It’s designed to help Esri users get more value out of their investment in ArcGIS. If you want to be cynical — it’s to keep their users in the ArcGIS ecosystem.
So let’s review our multiverse of global maps. In no particular order:
- Google Maps
- Apple Maps6
- Esri ArcGIS Atlas of the World7
So now we have half a dozen organizations, all creating essentially the same thing. A global street map of the planet.
It’s as though there are six different organizations creating six separate sets of identical roads for you to drive on. Or perhaps it’s like having six different electrical companies, each creating an entirely separate electrical supply network to your house. Is that crazy or what?
But wait — there’s more!
Ladies and Gentlemen, now we also have the Overture Maps Foundation!
The Overture Maps Foundation (OMF) was officially announced by the Linux Foundation on December 15, 2022 and has Amazon Web Services, Meta, Microsoft and TomTom as founding members. Clearly some heavy hitters. Together they will be “Powering current and next-generation map products by creating reliable, easy-to-use, and interoperable open map data”
Err, so how and why did this happen?
Well, having talked to a few people in-the-know, I think I can distill it down to three things:
Let’s Start with Money…
Building a high quality global street map from scratch is expensive. Super expensive. As I’ve said in prior posts: you had better start with a number greater than 1 that ends with the letter ‘B’. The hard work only starts after you’ve built the map. Now you’ve committed yourself to spending beaucoup bucks to maintain it. Only a very few companies on this planet have the financial means to do this and even they are under extreme pressure.
And as it stands in today’s economic climate pressures are now much, much greater.
Stock prices of all companies in this business have dropped — and for some — precipitously:
For a comparative benchmark consider the fact that the NASDAQ composite has dropped about 35% since its all-time high in November 2021. Alphabet, Apple and Microsoft are in this ballpark, but Amazon, Meta and TomTom have performed decidedly worse.
And there have been some hard realities for each of the OMF founding members:
- Microsoft: Back in the Balmer days Microsoft was fairly bullish about mapping (remember ‘Bing Maps’?9). They even started going down the path of building their own map. It wasn’t until Satya took over that reality hit: the bulk of Bing Maps’ assets were sold to Uber. Uber got serious about maps for a while, but then their own reality hit. Let’s just call that one ‘Travis’.
- TomTom: Since the heyday of a $2,000 navigation option for your shiny new vehicle, TomTom’s world has been shrinking inexorably. They got some solstice from licensing their data to Google Maps and later to Apple Maps, but now the Google revenue is gone. And Apple Maps is continuing to expand its coverage, so TomTom’s revenue from Apple has got to be shrinking fast.
- Meta: For Meta, well, we all know it’s not been easy. 13% of their staff got laid off in 2022. I’m sure that just like Satya, Zuckerberg has no appetite for investing heavily in a global map.
- Amazon: in 2022 Amazon has been implementing its largest cuts in its 28 year history and Jassy has just announced there will be 18,000 layoffs coming. Enough said.
So clearly each founding member must see OMF as a way to combine efforts and reduce costs.
Second Topic: Control…
Clearly Google Maps is a factor. It was a big factor for OSM getting its initial traction. But what’s wrong with OSM? Doesn’t that give these players what they need?
Well apparently not.
The founding members see specific challenges with OSM and they all relate to strategy and control.
While they can have some influence, no one company or group of companies that works with OSM has the ability to:
- Direct OSM’s strategy: what information is mapped, where it is mapped and in what order
- Define how information is represented in OSM: each country or region can effectively define their own data models independently of other countries
- Set QA processes: OSM leans toward manual and ground-truth verification processes; using input from massive sensor networks (e.g. vehicle sensors) to detect change or map errors is not widely endorsed
- Prioritize internationalization: not surprisingly local map editors tend to favor their local language (thus all the German labels in the OSM map of the Berlin Zoo above)
- Prevent vandalism: whereby malicious edits make their way into a widely published product
The last one is interesting. It caused a number of companies to get together to launch the Daylight Map Distribution organization which essentially puts OSM data through a data scrubber. Every day millions of contributions are made to OSM by thousands of people and it’s impossible to check everything in-real time. To quote Daylight:
“Some of these contributions may have intentional and unintentional edits that are incompatible with our use cases.”
In other words: they’d cause a major PR headache for any company that used the data.
My favorite piece of map vandalism actually took place not in OSM but in Google Maps, back in 2015. Some enterprising chap contributed the following edit:
Last Topic: Interoperability…
I whined about this in my post ‘Why Geospatial Data is Stuck in the Year 1955’. Given OMF’s recent announcement and its planned working group on data schema it turned out to be rather prescient. They clearly see the same issue.
The thing is it’s still very difficult to use data built by one mapping organization in another mapping system.
So, for example, the way in which the city of Los Angeles defines, say, their data for building addresses is very different from the way that the city of Turin or Osaka might do it. So somebody building a global map has to deal with all these differences. My analogy is there is no equivalent to a standard shipping container in the geospatial world. As a result everyone is hurting — it’s incredibly inefficient and it’s a huge impediment to progress.
In my mind map-editing tool makers should enforce standard data models on their users — or at least very strongly encourage it — and make it insanely easy for their users to adopt.
But alas, that is not the case.
The resulting pain is thus felt by any organization that wants to use a high quality global map in their product.
So, in summary, the reasons for the birth of OMF seem to be valid and defensible.
But the Question is, Will it fly?
Let me start by pointing out that the ask of OMF is high.
If you want a proper seat at the table — by that I mean a seat on the steering committee — get ready to cough up $3M per year and to dedicate at least 20 full time engineers. If you assume the fully loaded cost of an engineer is, say, $250,000 per year, then you’re being essentially being asked to contribute $8M per year. No small chunk of change even for a wealthy organization.
At the same time, OMF is keeping its focus very narrow: streets, building outlines and basic information about places (or ‘POIs’). As far as I can tell they’re not even focused on street names or addresses at the outset. This narrow focus will increase the chances of success, but will this shallow foundation be enough to be useful? I guess you have to start somewhere.
There’s also the complex question of how the data will be licensed and what effect it will have on potential data contributors. This is too big a topic to cover in this post, but I suspect there will be some practical and very real challenges, particularly around “share alike” clauses.
To succeed OMF is going to need a lot more participation.
I think to increase the chances of success governments will need to contribute data at a frequent cadence and in volume. The good news is there is no membership fee for governments. But this isn’t a case of “if you build it they will come”. Additional incentives will be required as, like everyone else, governments are strapped for resources. Do they have the time or the money to contribute? What’s in it for them? So far OMF has not made any grand announcements about sharing back with the communities it serves.
What about the other global map makers? Will they join?
- Google? I don’t think they have any reason to join. They already have what they need, including a vibrant community of 120 million contributors via their Local Guides.
- Apple? I think they’re too focused on their own world. I’m guessing they’ll play the ‘wait and see’ game.
- HERE? Maybe. Being privately held it’s difficult to know their financial situation. Somehow I doubt it, especially given their announcement at CES yesterday about ‘UniMap’ which is their own ‘2.0’ mapping initiative’. We’ll see.
- Esri? My guess is they’re probably the most likely to join. But will they take the lead and get all their government customers to participate? One can only hope.
So my final question is this:
OMF — is it ‘OMFG’ or just a big ‘MEH’?
Grab a drink. Get the popcorn. Let’s all watch and see.
1 Note: I said ‘multiverse’, not ‘metaverse’
2 See ‘The GBF/DIME System’ published by US Bureau of the Census in 1978. Digitization courtesy Google.
3 At this time HERE was known as NAVTEQ and Tele Atlas was yet to be acquired by TomTom. There was also third company in the mix that licensed map data, Geographic Data Technology a.k.a. GDT. It was acquired by Tele Atlas in 2004.
4 See Apple Maps expansions, courtesy of Justin O’Beirne:
5 It is interesting to note that Apple is one of the top ‘Paid Editors’ on OSM. For more details see Jennings Anderson’s 2021 Update on Paid Editing in OSM.
6 I know, I know — you could argue that Apple Maps is really a hybrid map and not a true global map of its own making.
7 Unlike the other players Esri is not creating any map data from scratch.
8 Stock table:
9 OMG — it still exists!
- The Overture Maps Foundation
- The US Department of Commerce — Bureau of the Census for their paper on GBF/DIME and Google for digitizing it
- Kaile Bower for her paper “Looking Back and Ahead: A History Of Cartography at The Census Bureau and What The Future Holds”
- Justin O’Beirne for his wonderful site chronicling the evolution of Apple Maps and Google Maps
- Reference for Business for chronicling the history of NAVTEQ
- The Linux Foundation
- Wikimedia for their tireless efforts to keep Wikipedia supported and for letting people reference it
- OpenStreetMap (OSM)
- Jennings Anderson for keeping track of OSM edits
- Best of OSM
- Stan Honey and Marv White for their insight on Etak’s discussions with Karlin & Collins and for being such an inspiration to me as a young engineer!
- Google for being a catalyst to help push the geospatial world forward