Does Apple have a Scott Forstall problem?
tech.fortune.cnn.com
tech.fortune.cnn.com
Those who keep saying "Maps is great" need to leave an egocentric view and realize that there are millions of users all around the world with different experiences.
According to the apology letter they serve about 500 million searches in one week. That means two billion searches per month. What does this mean in therms of customer experience?
http://www.mtonic.com/applemaps/
According to this either 60% of the locations either incorrect or missing altogether. I'll be generous and propose that Maps, world-wide, might have an average of 5% incorrect or missing locations. I don't know if this is too low or too high. It's just a number that I pulled out of my imagination in order to get a sense of proportion.
If only 5% of the data is wrong and we have two billion searches per month, that means 100 million bad searches per month. I'll let the reader guess as to how many users that affects. It certainly is in the millions.
If data errors are larger than that the situation is far worst. Again, I'll leave it up to the reader to guess as to how much user anger would trigger Apple to post an apology letter on the front page of their site.
If you live here in California, and, in particular, the Bay Area, please refrain from posting how "Maps is great and it is beautiful". You do not represent the experience of the vast majority of users who had almost no issues whatsoever with Google Maps.
Oh, yes, on the whole "Maps is beautiful" mantra. Who the f* cares? 3D view? Who the f* cares? First make them dead accurate, then add eye candy if it makes sense. Accuracy is far more important than bling when it comes to maps. Nobody is going to want to use a beautiful map that takes you to the wrong place. And 3D view. Really? Get it right first. Then play.
As it says in the article, the same could be applied to Siri but Siri came with the "beta" warning label, a lot of people understand that voice recognition and that kind of pseudo-AI is really hard, and it was better than anything competitors had to offer, which, obviously, is not the case with maps.
Well, pretty much every 10.x.0 OSX release has had some nasty regressions in it, but they're generally not quite as easy to spot as the maps fiasco. What's worse is that this isn't a new product that starts off a bit crap with the view of improving it over time - there have been a few examples of these as well.
Instead, from the users' point of view, the maps update falls plainly into the "if it ain't broke, don't fix it" category. Users don't give a damn about contracts between Apple and Google. The Maps app has been there since iPhoneOS 1.0 and it's not had any drastic changes since then, despite its tremendous popularity. I find it baffling that they didn't release the new map material as an extra mode with more features (3D, turn by turn navigation), in addition to Google's map material for the classic 2D view until iOS7 comes along. That would probably have completely avoided the problem and given Apple an extra year to get their data into shape. The new features, although a little half-baked, would still have been seen as an improvement by users, whereas now they're just blinded by the seething rage of having the core functionality yanked out underneath them.
And it did now?
Please go back to stats class.
My simplistic 5% = 100 million calc serves to simply highlight that small numbers can mean huge problems. Even if I am off by an order of magnitude this is still a huge number of searches per month that are potentially sending people to entirely the wrong place, if at all.
Relax.
I have a theory about these people -- they are the ones who never realized that you could search Maps for place names.
If you search for street addresses, both Maps implementations perform OK. But the strength of Google's database is that you can type almost anything within reason into the Maps search box and get where you want to go. For my first few weeks of iPhone usage, whenever I wanted to find a nearby "Starbucks" or "Barnes and Noble" or $name_of_local_restaurant, I'd do the search in Safari, pick a likely result, and then click on the page's address link, assuming there was one, in order to open it in Maps. Eventually I realized that I could type the search string directly into Maps, and that made life a lot easier. If that stopped working or became less effective on the iPhone that I use for navigation in the car, I would scream bloody murder.
So my guess is that the people who aren't complaining about Apple's maps application are either living in metro areas that were carefully vetted by the data provider, searching for street addresses directly, or both.
Some places are genuinely fine: I'm in San Francisco and the maps really are pretty damn good (in SF, I'll actually use the word good). Of course, this is right in Apple's back yard, but the reports seem to indicate that most major U.S. cities are in good shape (mind you, I'm talking about maps & places, not flyover's 3D views, which are a whole different story). And, I legitimately find the link to Yelp's reviews more useful than Google's place markings, but Google's are more consistently available.
Not to say that this isn't a huge mess, but simply that there are places where "Maps is good" ("Maps is great" is a bit much no matter where you are!).
(Though I had less luck with Hong Kong. For instance, a search for "Big Buddha, Lantau Island, Hong Kong" and every likely English variant just finds the middle of Lantau Island, not the Buddha itself.)
Another example: in Saigon, the Mekong was not rendering as water. Imagine being in Manhattan and having big blank yellow spaces where the East and Hudson rivers should be.
In New York, the situation is fairly dire. Before, subway stops showed up on the map with the names of the subway lines so you could see that the nearest stop to you was an "R" or a "4/6" or whatever; now that information does not appear on the map at all.
If you use the map and ask for transit directions, that used to just work. Now you get presented with a list of apps it can launch for this purpose, but some of them are terrible and the ratings don't reflect this - you just have to try them at random and see. For instance, there's an app called "Transit" which tells you which stations you want to go between but it doesn't give the names of the subway lines so you don't know which platform to stand on to get the right train. Meanwhile, one of the very best apps for doing New York subway navigation and planning - KICKMap - doesn't even show up as one of the available options from Maps.
And we know this because they didn't setup relationships with any of the local data providers in each country. They just setup a single one with Yelp.
- Park Hyatt Tokyo - Tokyo Railway Station (the MAIN station) - Paddington Station in London - Melbourne Aus and Frankfurt Germany Airport roads - Hong Kong is misplaced http://www.ausbt.com.au/6-reasons-travellers-should-be-wary-...
The list goes on across the net. I support several other posters in that only an Apple fan would say that the maps are good overall. All you will hear from these people is how TomTom had a bad map 5 years ago, or how Garmin is missing a 3 year old suburb in their area. The fact is that in this instance, Apple stuffed up severely.
It was a brave decision of Apple to go ahead with this version of maps - they would have known full well just how bad it was. Maybe it's the kick in the butt that Apple needed to catch up quickly. I expect that within 6 months, the situation will be a lot better for Apple.
This test makes a bad comparison. See here:
The ordering in making that fateful Google decision must have been first to decide if their own alternative was going to be close to ready. At that point somebody aggregated the data from the Maps team and made a forecast that said they would be.
Further, the failure of Maps has made iOS 6 itself a failure. You can tell from the marketing materials that the new Maps application was the headline feature of iOS 6. You can reverse engineer their thinking - we are dropping Google so this new release of the OS needs something special to hold it up. That 'something special' was the 3D view.
People are forgetting that Maps was marketed as a better application than Google Maps.
So that leaves a few points where it went wrong:
a) whoever provided the input to the group that decided to drop Google in saying that they would be ready
b) whoever decided to drop Google based on that info
c) whoever was responsible for delivering the finished Maps product
d) whoever was responsible for pinning the success of iOS 6 on Maps
e) whoever was responsible for testing Maps
f) whoever was responsible for raising user expectation by marketing Maps and iOS 6 as being 'better' and 'the best'
As head of iOS and a member of the exec committee, a lot of those paths lead back to Forstall.
According to a followup article ( http://vore.cc/post/32503374905/old-maps-vs-new-maps ) the old Maps app fails this exact test at least as badly. It just fails a bit differently. Whereas the new Maps app often says it can't find anything, the old one would for a similarly sized group of searches silently err on the side of finding a match that was wrong.
The problem is very simple: The same city name shows in in multiple countries and the old map just didn't know which one he wanted.
I'm sure TomTom and Apple cleaned up the data for areas where their programmers and bosses live. Even in Seattle, I searched for "Pizza" and it didn't find the place down the street [I had to search for "Pizzareia" for it to show up].
When you see reports of major airports being missing or mislocated, you can't use the "Well my local airport shows up fine" defense. It's like an inverse Tu Quoque fallacy [you hit a problem, I didn't, therefore the problem isn't there].
http://www.fastcodesign.com/1670760/will-apples-tacky-softwa...
“But before Forstall, it was Steve Jobs who encouraged the skeuomorphic approach, some say. “iCal’s leather-stitching was literally based on a texture in his Gulfstream jet,” says the former senior UI designer. “There was lots of internal email among UI designers at Apple saying this was just embarrassing, just terrible.”
[0] http://www.fastcodesign.com/1670760/will-apples-tacky-softwa...
Textures and styles aren't skeuomorphism, skeuomorphism is imitating real-world objects to provide a sense of familiarity in execution. Not a notes app that looks like a notepad, a notes app that FUNCTIONS like a notepad.
It’s consequently all about style, not about function (at least when it comes to real-world objects). Wikipedia has examples like flame shaped lightbulbs, artificial film grain or spoke patterns on hubcaps.
It seems that when it comes to software, skeumorphism can also refer to functional adoptions of virtual interfaces to their real-world counterparts†, but I think it would be misguided to exclude mere style here.
I mean, in the end this is a stupid semantic argument. Using leather textures is what it is, whether it’s skeumorphism or just style, but you two certainly do not have the facts on your side.
—
† It’s pretty clear why that is the case: it’s only possible to make lightbulbs look like flames, not to make them work like flames. Software provides more freedom and makes it possible to make something virtual work like the real thing.
"skeuomorphism is when a product imitates design elements /functionally necessary/ in the original product design, but that becomes ornamental in the new product design."
So yes, it is about function.
And also, the three things you mention are directly related to the function of the original objects. While these things are clearly used ornamentally in modern devices, their existence/use was inherent in the design of the modern devices' predecessors.
Your examples are skeumorphic because they use functional aspects of the original. "Find My Friends" is not skeumorphic because even if such a thing as "Find My Friends" existed in a previous form, the leather has nothing to do with its function.
If that sounds confusing to you, then I don't know what else to say. You're wrong?
For example, the rich Corinthian leather of a desk calendar functions as an arm rest and also functions as a binding mechanism for the calendar pages within. The stitching of said leather functions to hold the piece together.
And again, where is the original product design for Find My Friends? Why did I never hear about anyone using a leather-bound location diviner before?
However, that doesn't mean every thing -- like the leather stitching in "find my friends" -- that Apple does involving natural textures is skeumorphic.
0: http://www.dailytech.com/Apples+chief+Ahole+Forstall+Keeps+S...
The exact opposite of that would be Bob Mansfield.
Its also a "device centric problem": I'm always struck when talking with someone from Apple about how device centric they are thinking (running algorithms locally that should run in the cloud, etc.). The cloud is eating the world and device centric thinking, while its done them well for a while, is increasingly incorrect.
-- Seriously? This is "analysis"?
A "a self driving car with tons of lasers spinning on it"? Learn before you speak. The noticeable problems are not subtle: clouds covering cities in the UK, buildings not being named on College Campus, etc. Stadia 0.5km out of place. They are obvious, silly, dealbreaker problems for Soccer Moms and Sociology FFS. Algo's are not going to solve this cr@p, its Cartpgraphy 101...and in any event, Google has to hand fit the data that they do have, for a variety of reasons, to sat-nav visuals (including incosistent basemaps).
Algorithms are indeed solving a lot of these problems for Google. They are reading the street signs and shop signs for peet's sake!
In short, I don't think you've thought deeply about what you're writing.
Maps, however, is crucial to most people that use an iPhone. After telephony and web functions, it's probably the next most used feature. So, any failure by it, even if it's slight, is amplified. Yes, you can use Google Maps on the web or another app (MapQuest should really be capitalizing on this opportunity to reclaim some relevance), but it's one of the default icons on the home screen and where most people are going to head first. It would be like the dialer not being able to call certain phone numbers or the text message not working with certain recipients. It's a vital function and has to work.
I would say Siri is probably more of a failure than Maps, but because of it's non-essential nature, that feature is able to get away with what Maps can't.
Citation please. Nobody I know uses google maps, even when we travel (which is a lot).
There are some videos of non-native English speakers trying to use Siri on their new 4s (before additional languages were added), which are quite funny actually.
The same goes for the new Maps, which seems to have most of its issues related to countries outside the US.
But especially with iOS6, it's actually the first time for me personally where I feel like I'm downgrading my device instead of upgrading it: youtube app - gone, maps - replaced with inferior alternative, ...
Aha and there is class action lawsuit: http://dfw.cbslocal.com/2012/06/12/class-action-lawsuit-file...
Here is the quote:
Ryan asked Siri, “What is a class action?”
Siri told her, “I found this meeting.”I think it’s pretty useful and use it sometimes.
1. Call my wife at her office. 2. Text My Wife - I'll be home in 15 minutes. 3. Set a timer for 15 minutes. 4. Where's the closest Mc Donalds.
Obviously I'm not looking for a magic speaking encyclopedia of knowledge, that being said my primary use case for Siri is in the car, and for 90% of what I need to do it works fine.
BTW I just asked Siri "What is a class action" and it asked me if I wanted to search the web for "What is a class action" and it found the definition. I'll give it this though, whether automated or manually Siri's been pretty consistently tuned to get better and better as its gone along which was likely to happen as it got a larger and larger dataset to work with.
It's no AI and you'll have trouble if you treat it as such, but if you think of it as list of occasionally useful things that it can do, then it's pretty handy.
And Siri responded by asking whether I wanted to search for "Shang Tsung from Mortal Kombat", getting the spelling exactly correct.
Since iOS 6 (or at east I never noticed before?) there is also an info button on the Siri screen which shows example querys/commands. This hels a lot!
This, I guess, would apply to every non-english country and seems like a quite an oversight rendering music controls via Siri useless. Since english words and expressions became natural parts of every language, and are also pretty common in artist, movies, brands, places, etc, Siri should always check back with its english dictionary regardless of language settings.
Me: "Suche nach Häcker News"
Siri: "Das Web nach Hecker News durchsuchen"
Googles Search: "Meinten Sie: Hacker News"
Case in point - Steven Sinofsky and Windows 7.
I dont know what other project they can put someone like Scott Forstall on without even indirectly becoming related to OS development at Apple.
http://articles.businessinsider.com/2011-05-07/tech/30043798...
Makes sense to some extent at a company like apple. Maybe you get two chances but if you don't have the foresight to see how important maps are to people you might not be the best fit for the job.
Google has done this repeatedly with products like GOOG-411, a data collection exercise that was promptly shut down once it had served its purpose. Apple is taking a risk here because Maps is not some 'free' offering like GOOG-411, but they've done the math and will come out ahead when the dust settles.
I think the new Maps app is fantastic; the data accuracy leaves a bit to be desired, but I notice that those doing the most high-profile complaining point to errors in rural England, rural Ontario, etc. Here in the Bay Area, the data is working fine for me as well.
It is not shocking to me that a map database that is a few months old is less mature than one like Google's, which is the product of a decade's effort, and also the product of 5 years' worth of improvements riding the back of iOS users' crowd-sourced data. Now Apple has that data coming into its database, instead of flying out the door to Google's.
Maps is likely to improve, and fast.
In any case, blaming Forstall is idiotic and talking about firing him is Fortune-level idiotic. He made great things at NeXT, Mac OS X is great, Siri is great, and Maps is a work in progress that shows great promise.
Have you tried reloading map data and zooming in and out on maps in Google Maps vs. Apple Maps in an area with EDGE or otherwise bad internet coverage? Apple's app wins hands-down versus reloading non-vector map tiles.
http://www.bangkokpost.com/tech/gadget/314283/apple-maps-bet...
You didn't have a perfectly working solution before, btw; Google Maps still has a lot of errors.
And what you have now is both Google Maps (via Safari), and also a new Maps app that is internally superior and relies on a less-mature dataset which will mature quickly and give Google some much-needed motivation to improve its data as well. I'm not sure why this is such a big disaster.
Edit: By the way, Apple Maps finds Amsterdam Centraal (not "Central") Station just fine, if you spell it correctly. I just checked; it finds the exact same place as Google Maps.
I don't know if I'd make this comparison. The level of erroneous data in Apple Maps is just insane compared to Google Maps.
I was under impression that this all similar to "antena gate" (kinda not big deal), but iOS maps are really really bad even in Bay Area.
EDIT: Here is an example: Search for "428 University Ave, Palo Alto": you will end up in Los Gatos.
I'm working on a routing/location based app right now. Out of curiosity could you give examples of what you mean by "really bad"?
Still though, "really really bad" makes me think you have experienced many other issues. I'm genuinely curious about how Maps performs in the places it is "supposed" to be at least passable. Got any other examples?
Meta: please don't edit your comments to respond to sub-comments. It makes the flow of the conversation very difficult to follow and makes my previous comment seem irrelevant and out of place.
I'm just one user, but if a navigation app takes you the right address in the wrong location in a place like the Bay area or NYC, that's a "really bad" problem.
(I've never used Apple Maps.)
The app, as I said, is superior in my opinion to the Google Maps app. It zooms better, loads faster, looks better, has better layout, is prettier, and has turn-by-turn directions and 3D features that were absent in the Google Maps app. Remember that the app and its functionality is completely separate from the database it points to. None of the map data is stored in the app.
"Having a medical emergency? Then he might not want to use your iPhone five to search for an emergency room."
> Maps is likely to improve, and fast.
This is going to be the real test for Apple. They need to fix stuff, they know that, it will be interesting to see when improvements will be noticible.
Cook spoke of 500 million searches in two weeks. If only 1 promille triggered a feedback or "report problem", that would be 500.000 change requests. How can you brute-force this massive amount of data?
A quick Google search will tell you that there are non-trivial problems all over the world in major cities at major landmarks with Apple maps.
It's nice that you agree with the parent poster that Apple Maps are fantastic, but let me say that I don't think so. There are far too many obvious and MAJOR cockups (eg. Hong Kong is misplaced - it's not even on the island, Tokyo's main railway is missing, etc.) ... but you think that's Fantastic??
I will point you to Garmin and TomTom as two examples of maps that FAR exceed Apple's "fantastic" offering.
In 6 months time, the landscape will (hopefully) be different ... but to call this offering fantastic is ignorant.
Perhaps Scott realized that and decided to take it back from Google? I don't know. I do know it is pointless to try to second guess the executive staff of a company like Apple (or Google, or Groupon, or Zynga, Etc) even though it is fun.
I'm sure they are keenly aware of the shortfalls of their own software, as much as they are of the hardware they make. That's the whole point of QA. What pushing out products early does is let you start getting feedback on the known unknowns - problems you'd expect to have but didn't know exactly where they might be, and the unknown unknowns - problems you didn't see coming at all.
http://nerdtwilight.wordpress.com/2009/06/25/apples-steve-jo...
Gargade in, garbage out.