Analyst commentary on issues, trends, and developments in the automotive telematics, navigation, and infotainment markets.

Two navigation system suppliers have been kind enough to provide me with demonstrator units, and I won’t denigrate their desire to improve by naming them here.  But I will use them as an example of what must happen in the navigation space, before navi systems can be seen as mass-appeal devices delivering a user experience worth spending money on.  Only those companies who get the experience right will be able to hold any kind of margin as things develop and more alternative solutions become available.

Over the weekend, I had trouble with both devices (a PND and a mobile phone-based system).  The trouble came not in routing, or traffic alerts, or telling me which lane to be in, or any of the truly complex tasks a navi system is called upon to deliver.  The problems were in entering addresses - the most fundamental activity for any navigation device.  Almost every consumer study, notably the J.D. Power surveys, have shown this to be the single most problematic task facing navigation software designers.  Clearly it has not yet been solved.

For everyone who thinks this is just a short-term problem, and voice entry will solve things – it will not.  These are logic-based problems, not technology problems. The US has a particular logic to creating and describing addresses.  Most US residents interpret addresses the same way.  We usually start with a street address, street numbers before the street name, and street address before the city name.  Occasionally we have the right zip code.  Rarely there is a "street line 2" for building or suite or floor.  Streets often have compass headings (East, West, etc.) and those headings can be before or after the primary street name.  A bit of thought (by a native American), or some quick focus groups and/or surveys could easily find the most used paradigms.

Getting this logic right is an essential requirement of navigation device software. Every typical (and sometimes atypical) US address format and entry approach need to be allowed and accommodated by even the most lowly navigation system.  Designing a system for global markets is frankly of no concern.  Each country will need tailoring of the interface – that is the only way to get a good experience for all users.  Compromising an essential element of usability - address entry - in the name of development cost savings is a shortsighted approach that will come back to bite those who attempt to save a couple dollars, euros, or yen.

To be a bit more specific, for all the navigation software makers out there:

I was going to 2771 W. Pico Blvd, Los Angeles 90006 (Home of Papa Cristo’s Greektown Grill – a fantastic Greek place, for those of you in the area.  For those who are not, look at  www.papacristo.com).

The PND provides the choice of entering an address by city first, or by street first.  One would think that is a great feature to have, since it allows for more user discretion.  Except, if you choose street first, it picks the last city entered as the search city.  That means the formerly useful “intelligent street search” is now your enemy.  If the street you enter does not exist in the city the device is checking (a city which is not made clear to you), you are not allowed to enter that street name.  Frustration ensues. 

Further frustration is generated because the device apparently did not allow entering zip codes – it chooses them by itself.  So when I gave up on “street first” and tried “city first”, entering Los Angeles, it gave me the single choice of LA 90012.  Well, unfortunately, my address was in 90006.  90006 was nowhere on the listed options.  Now what?  I finally chose the number entry touch pad instead of the default alphabetic entry pad, and entered 90006.  It accepted that because, yes, it actually was “Enter by City or Zip”, a choice not listed on the main menu.  Then I needed to learn that W. Pico Blvd had to be entered as Pico Blvd, with no W.  Of course, in LA, there is a big difference between East and West, and the street address contains the W – we do not think of Pico Blvd, then decide on E or W.  At this point, were it not my job, the device would have been returned to the store.

The phone-based system had a different problem.  We were going to 1000 Cerritos Ave. in Anaheim, CA (a hockey rink).  We did not have the zip code.  The phone-based system allows you to fill in the address form starting at any point: address, city, state, zip.  Great, except when it decides there must be a “street 2” entry.  It felt 1000 Cerritos Ave. needed a street 2 entry.  After giving up and using my PNW (Personal Navigation Wife) to find the hockey game, I sat down at home and managed to realize I was somehow using “Find by Intersection” instead of “Find by Address”.  A very easy mistake, and also very easy to correct – if the screen had given me an indication I was in “intersection” mode and not “address” mode.  Unfortunately, it looked and acted identical to the address mode.   A small-screen problem?  There was room for an advisory telling me to “Type in a Street” – that advice is obvious.  The entry mode I had erroneously selected was not obvious.  User Interface Designers please remember – the user is never wrong.  This might seem like my problem, as I chose the wrong mode.  Nope – it is the device’s problem, for not guiding me to the right mode (or at least making it clear to me what mode I was using).  I also had to learn that zip codes can not be entered if you enter a city and state. This was frustrating the one time I actually had a zip (90006 for Papa Cristo’s). 

In general, I can say the mobile device was better at address entry logic than the PND.  For example, when doing address entry, it assumes a number will be first in the address entry area and has number lock on for that entry, until the space key is pressed.  Sadly, neither device was at a level where my mother could use it, which can be one standard for defining a good navigation experience.  There should be zero frustration at the very start of a navigation process.   Users will be quite forgiving of challenges with things they see as difficult or complicated.  But there is no tolerance for problems with things they see as easy to achieve – like entering an address. 

Can this really be so hard?  The profession of User Interface Design has many tools for working through user logic.  The Cognitive Walkthrough and the Use Scenario Test are a couple.  I was a human factors engineer in a prior life, and such scenario testing is not too difficult.  It has to be done with the input of native residents of the country for which the system is being tailored, and it can not try to accommodate more than one country.  Paper-based scenarios and flow-charts are good enough, so long as the logic is thought out.  It would be great to follow-up with real usability testing, but that isn’t necessary to get it right.  All navi device makers can be assured they will get plenty of usability input from their customers.  They must be ready to quickly adapt, and focus on what users say is important not what the managers or engineers think is important (i.e. more features is not the answer).

I am quite certain 2007 will be the year of navigation, where the holiday season will see a navi device in many gift boxes.  It will also be the year large numbers of people get disillusioned and frustrated with the devices.  This is both good and bad.  The truth will come out regarding who has spent the time to design a good user interface, versus those who spent their money on marketing.  Depending on their temperament and which device they purchased, buyers will either a) throw the device away and never get another, or b) decide they need to upgrade to a device with better usability.

At the very least, we will have a much wiser and more discerning group of potential customers.  This means such flaws in interface logic will quickly be weeded out.  The companies who get the Navigation User Experience right will be the ones who can differentiate themselves, and maintain margins the longest.

Telematics will be an Increasingly Important Piece of the Bigger Picture
Author: David Alexander
Principal Analyst, Commercial Telematics
Posted: Thu, 19 Jun 2008 16:50:33 EDT

Telematics Detroit 2008
Author: Dominique Bonte
Principal Analyst, Telematics & Navigation
Posted: Tue, 27 May 2008 07:41:52 EDT

Navigation and LBS at CTIA
Author: Dominique Bonte
Principal Analyst, Telematics & Navigation
Posted: Mon, 7 Apr 2008 16:51:36 EDT

Navigation Highlights at CeBIT 2008
Author: Dominique Bonte
Principal Analyst, Telematics & Navigation
Posted: Tue, 11 Mar 2008 09:36:08 EDT

The Mobile World Congress 2008: Pedestrian Navigation at Last!
Author: Dominique Bonte
Principal Analyst, Telematics & Navigation
Posted: Mon, 11 Feb 2008 14:32:37 EST