Disruptの楽屋裏でGoogleのMarissa Mayerが将来の「ランダムなおすすめ機能」について語る



先週行われたTechCrunch Disruptで、Googleの位置とローカルサービス担当VP Marissa Mayerがステージ上で、本誌編集長Michael Arringtonのインタビューに応じた。そこではGoogleのモバイルの成長からMayerの投資戦略に至るまで、いろんな話題が登場したが、そのインタビューの直後に、Googleのモバイルとローカルに対するアプローチについて、彼女にあと二三尋ねることができた。Googleの、ローカルに対する二本柱方式については、ほんの数週間前にMayerにインタビューしたばかりだが、まだ聞きたいことはいっぱいあった。

最初に質問したのは: AppleがiPhoneとiPod TouchとiPadの上でGoogle Mapsを使うのをやめて、内製のソフトを使うようになったらどうなるか?(Appleはトラフィックデータベースの改良に取り組んでいることを認めたし、秘かに地図関連のスタートアップの買収をすませている)。

当然かもしれないがMayerはこの質問をほとんどはぐらかして、Google Mapsのユーザ数は今や膨大である、同社のモバイル製品上だけでも2億人のアクティブユーザがいる、と言った。また、最近GoogleがMapsに対して行ったイノベーションについても述べて、ベクタグラフィクスのタイルによって、データの所要量が1/100に減った、と言う(MapsのiOSバージョンでは未実装)。

次の質問: モバイル上の位置判断をより精密にする技術は開発されたか? ほとんどの場所でWifi/GPSの組み合わせは良好だが、ニューヨークのような混雑した都市では、精度がががた落ちになる。

Mayer曰く、Googleが今作っているWifiホットスポットのデータベースが、今後の時間をかけた改良とともに役立つようになるだろう。GoogleがユーザにLatitudeによるチェックインを奨励しているのも、そのためだ。それはユーザの参加性を高める新しい手段であると同時に、Googleがデータベースを良くする助けにもなる。ユーザの正確なGPS座標データがシステムに分からなくても、同様の信号の誰かが前にそこでIn-N-Out Burgerにチェックインしていたら、そのユーザもそこに今いる可能性が大きい。Mayerはそういうシステムを、Googleの検索で使われているスペルチェッカーの、ユーザ参加による成長〜質の向上になぞらえた。

最後の質問: 前に彼女が、とても楽しみにしていると言った、携帯上のランダムなおすすめ機能について。Mayerは今回それについて、”2年以内に実現する”と思う、と言った。そういう、意外性のあるおすすめ機能は、まだUIも決まっていないが、知らない間に実装されていた、ではなくて、はっきりと発表が行われるだろう、と彼女は言った。

〔訳注: ランダムなおすすめ機能と訳しているserendipitous suggestions、serendipity機能については、将来は(もしかして今すでに?)「セレンディピティ」が片仮名日本語になってしまうかもしれません。〕


Hey, this is Jason K. with TechCrunch TV.

And I am here with Marissa Mayer.
Again, I actually just got a chance to interview her earlier this month.
But now I have another chance
to ask more questions about all
the cool local and other
products that you guys are working on.

And once again, thank you for coming here to TechCrunch Disrupt.

No problem.
I love TechCrunch Disrupt.
It’s just a really great opportunity for entrepreneurs, and it’s just also amazing to see all these launches.
I just love that there’s this launch vehicle that really helps get companies going.

It’s definitely it’s exciting.

So, here’s a question
that Mike didn’t really touch
on at all, and I’m not
sure how much you’re totally
going to get into this, but there’s
been speculation that Apple is going to do some stuff around maps.

There is a lot of controversy as
far as location and WiFi
data recently, and as part
of that, Apple responded indicating
that they were doing something related to traffic mapping.

And I’m wondering what’s going to

change, if anything, for Google’s strategy
if one day a significant segment
of the mobile market, in other
words all the iPhones out there
and iPod Touches, stopped using
Google Maps powered product, and

transitions over to something that’s controlled by Apple? Is
there any game plan there, I’m sure there is?

Well iIthink I’m
not going to speculate as to
some of our competitors, but I
will say that one thing that

we’re really learning is that
maps is a huge part of the mobile ecosystem.
Today we announced we have
200 million active users
of Google maps for mobile.

And we actually now are

seeing crossover on the weekends.
What I mean by that is that
we see more traffic from mobile
than desktop maps on the weekends.
In June we think that is going to crossover permanently .
We know that’s a really big part of it.

That said, we’re really proud of some
of the innovations that we’ve been able to roll out.
Things like latitude, but also things like Vector maps.
It’s not just about tiled maps anymore.

Which aren’t actually on the iPhone as far as I know.

That’s right.

And so…

The iPhone map is not as good.

No, the vector maps, they’re beautiful, they’re fast, there’re one one hunderd of the data.
And the other nice thing, that in
addition to being able to really
quickly drag and drop
and move around, because were transmitting

mass data we can do interesting
things like cashing the map
around you, so if your connection
suddenly fails, you’re not
lost sitting there looking at
the blue dot in the

middle of an area
that isn’t filled in, and so there’s just a lot of potential.

So here’s another question that’s
actually been especially important
to me out here in New York.
I noticed that identifying exactly

where I am is not working
so well in New York City.
I’ve got my wifi
enabled and I know Google
has a very powerful data base
as far as associating were you

are using the wifi hotspots that are available.

Is there some other technology that I’m not aware of?
Or something that is coming
down the line that will be
able to improve on these wifi hotspots?
Is this sort of something that is going

to improve overtime as you get
more hotspots in your database?

I think that the hotspots are something that will help.
New York city’s particularly challenging because
of the tall buildings and
the signal and the surround is

also hard to get a signal
into a really tall
building and it can drain the battery.
things like that, and do they
really understand where a
person is when they say find me and show me on the map?

That’s actually reasonably hard to do.


But one of the things here
is really about inputs, and
so, you know, we really hope that the users check in.
We have loyalty offers in

order to promote more people doing that.
The other value in check-ins is it actually causes person to say.

I’m here Actually, I’m here,
and we understand them, what
that signal looks like.
So then later when someone

else is there, we have a better chance of getting it right.
And so there really is
like a lot of times in
these types of systems where they need to learn.
All we’re talking about is really
the learning problem, they need a lot of input.

I wanna make sure I’m clear on that.
So, these explicit check-ins could
eventually help you who even who aren’t explicitly checking in.

That’s right.

It will just makes the database better.
We’ll say we have a

vague idea and these coordinates seem to match with this person who explicitly checked in there.

That ‘s right.
For example, like, spelling on Google.
Today our spell-check is so good because more people mispelled before you.
They mispelled and they figured out howDid I get it right?

We learned off of that pattern, and now we are really good at spell checking.
Same thing, right?
If we could guess where
you are, but now that you
told us means that we’re more likely to get it right for the next person.

So, one of the things

you talked about on the panel
earlier was about this
notion of having serendipitous suggestions based on where you are.
And I know you’re not going to
be able to go into too much
detail here, but as far as

the timing on when would we might expect this sort of thing?
Are we talking long term, like
it’s going to be years before the
data is there, or is this
something we might see this year
or next year, we’re going to

start having our Android phones say like, “check this out.”

I think it’s probably inside of a two year horizon.


Because, one of the reasons I moved from Search
to what we call “Geo”
is that- I just think there’s something really special happening.

With location, with mobile phones,
we can just do things that
we couldn’t do before, because we
can understand so much more about what’s happening with a person.
Where are they?

Is this some place where they are frequently?

Is it somewhere that they only are very occasionally?
And also it’s getting a little
bit of contacts and what other
things may have been doing.
These are making really good suggestions, understanding social context.
All of these different pieces, and

so a lot of the
building blocks you need are
in place, and I think this is something that is going to happen quickly.

Once it does start making this connections
again this may not be something
you can get into, do you think

they’re going to be more subtile in
the sense that you’ll open a
browser and it will say like,
I don’t know, there’ll be a link
that’s at the top of
the ten blue links, it’s nearby, as

opposed to a popup when
you open your phone that says “check this out now.”

I think it’s so new that
it’s hard to speculate on what the right UI is.
But I do think that a
lot of times with a new

medium you really need to start out somewhere more explicit rather than less explicit.
It’s sort of interesting.
I haven’t gotten to go to a
show this time but I like Broadway musicals.

I mean.
I’m going to see a Broadway musical.

Actually, I don’t really like the revivals.
The only thing that I
don’t like the revivals is I
gonna have to be super explicit.
The thing with the look, like they
have a whole scene change were someone

says like will I ever see him again?

Right, Right.

And I do think the
same is true with technology, when you are designing a new interface.
At least in the beginning so people understand,
what are you doing, why are you doing it, what does it mean for me?

You need to be more explicit.
Then later you can be more subtle.
So exactly the right way to
leave it in I don’t think we quite understand yet.

And we need to still
figure that out, but I do think

that, at least in
the beginning, just to make sure
people really understand what’s going
on, because we want to be
very clear and transparent with users.
It probably will be more explicit.

Great, alright so I think
I’m being that I’m are out
of time, thank you so much for
joining us once again and for coming to TechCrunch Disrupt.

Thank you very much for having.

(翻訳:iwatani(a.k.a. hiwa))