UX Innovation > In-Vehicle UX Blog



What3words: Car Use Questionable

by Derek Viita | May 02, 2019

Automotive, more than any other channel, tends to get flooded with solutions from other domains regardless of their relevance for the car.  Some tech has stuck around (Bluetooth).  Some tech did not (NFC for mobile media).

One of the newer and more intriguing solutions in the car space is what3words, a geocoded "navigation language" which divides the entire planet into 3x3 meter squares, and gives each square a unique 3-word name.  what3words (W3W) is then used by businesses, government agencies, and consumers via its website, mobile app, or API.

The what3words address for Strategy Analytics Milton Keynes, UK

W3W deftly answers current needs in certain problem spaces.  But despite several years of relentless marketing, it is still not a revolutionary product, and will find rough seas on its journey to scale.  Especially in the car.

Let's start with an overall summary, then narrow our analysis to the car.

First, a broad view:  W3W is the Esperanto of navigational languages, in every good and bad way.  Like Esparanto, W3W has some great simplicity-related value props going for it.  For W3W these are:
- Simpler voice entry.
- Simpler automated logistics for emergency localization, delivery, and so forth.

Of course, simplification always comes at a cost of some kind of fidelity.  For Esperanto, costs include a lack of cultural sensitivities, and a poor relationship with non-Roman languages.  In the case of W3W, the cost for simplicity is the removal of context and relativism.  When all context is stripped away, one has no way of knowing whether an address is in the general area the user is interested in.  Or if 2 addresses are near each other, in the same town, or even on the same continent.  Is "lance.string.crater" near "lances.string.crater"?  Well, no. One's in Alabama, the other's in Alaska.  And a user must perform a multi-step task on W3W's web portal or mobile app just to confirm that.

Also, because the system is based on a 2-D grid, multiple destinations at the same W3W location (e.g. a building with multiple floors) still require separate address lines.  So although W3W enables an undeniable improvement for voice entry, and is an outstanding solution for unlabeled locations in rural areas, it does not solve for complex delivery or wayfinding in cities.  And note, too, that the world's population is migrating from rural areas toward cities.  Not the other way around.

Now, let's take all these pros and cons into the car...

W3W has decided that the car is low-hanging fruit, because of two very true things:
- Manual destination entry in on-board satnav is largely a mess that consumers do not like.
- Voice is still the most-ideal HMI modality for the car, because it allows a driver to keep their hands on the wheel and eyes on the road.  And W3W is tailor-made for voice-based destination entry.

So W3W has made a very aggressive push to get themselves into the dash.  Mercedes-Benz and Ford have been the 2 highest-profile partners announced thus far.

But here's the rub:  W3W does not solve any car-specific problems.  It just repackages them.

To enter a W3W address manually in the car, a user must still type it using a touchscreen or jogdial-enabled keyboard.  The experience is still frustrating and time-consuming.

To enter a W3W address via voice, a driver must still use the in-car microphones and connections.  These are still unreliable in the wild, and contribute to loads of recognition errors.  Consumers have grown so impatient for usable in-car voice controls, they now greatly prefer mobile or home-based voice assistants, and do not see automakers as innovators in the space.

And as with any voice system (home, mobile, and especially car), error correction is still painful.  To that, W3W-in-the-car presents an even greater challenge.

Imagine while driving on the motorway, you use the car's voice control system to enter the W3W address "hike.clean.both" (located in a field near Stirling, Scotland) into your satnav.  Unfortunately, the on-board system recognizes this address as "like.clean.both" (in the Firth of Forth near Aberdour, Scotland).  Can you correct that first word on the fly, or do you have to nix the whole entry?  How is that process easier than correcting a UK postcode?  How is this easier than correcting GPS coordinates?

W3W does have a future, especially in the automated delivery and emergency response spaces, and especially for rural areas where addresses are unlabeled or difficult to communicate.  But despite its marketing for Asian mega-cities (Tokyo for example) it does not solve any of the growing problems for dense urban environments.  And despite its partnerships with Mercedes-Benz and Ford, until embedded voice systems are drastically improved, it does not solve any problems in the car.

Previous Post: Uber's Weakness; Tesla's Strength
Leave a comment