Name: BC17D2T2F WiFi User Experience Updates
Uploaded: 2025-06-26T10:27:01.000Z
Duration: 17 min 32 s
Description: Thousands of YouTube videos with English-Chinese subtitles! Now you can learn to understand native speakers, expand your vocabulary, and improve your pronunciation...

Users understand what's going on for the system and can understand the signals that we're trying to send, as well as when the system does something automatically for them, they understand why it made that decision.

And then most of the time, we want Android to do the right thing on behalf of the user when it's possible to do so.

So let's go through a few of these problems.

Auto-connecting can be really frustrating.

So if you are, how many of you guys have Xfinity Wi-Fi saved on your phone?

OK, so when you saved Xfinity Wi-Fi, you were making a commitment to auto-connect to 20 million hotspots within the US.

And not all of them are high quality at a given time.

So what that means for a typical user, when you're in range of these Wi-Fi networks, and some of them are good and some of them are bad, is you will be connected to them, and often your connectivity will stop.

On the backside of that, we have something called Wi-Fi stickiness, which is once you are connected to Wi-Fi, you can be connected for too long.

And that can cause your connectivity to drop as well, even as you're walking out of range of that access point.

So we know that today, about 40% of our users turn off Wi-Fi, primarily because they're connected to bad networks.

And we think this is something we can start to tackle in Android.

So we did a survey of some of our users, and we put up just this little icon here on the right.

You can see Coffee Shop Wi-Fi, and said, what do you think about Coffee Shop Wi-Fi?

Many of our users picked up that it's not secure, but many of them also picked up other signals that aren't actually there.

So some of them said it's fast and reliable.

Some said that it is unreliable and slow, just the opposite of that.

The truth is, they're starting to draw conclusions based on signals that they wish were there, things that they would like to know about Wi-Fi, and that they've inferred over a period of time, but that we don't actually tell them with the system.

So we also did a survey and said, what do you actually want to know about Wi-Fi networks?

What are some of the things that users most desire to know about Wi-Fi?

And it's things like, does the network have connectivity?

Is it protected by a captive portal, or can I just connect to it?

And then, is this something that is fast and reliable, or is it going to be slow for me?

So those are the types of things that users actually want to know about the networks they're connected to, and today, we're not telling them that, so they're starting to guess, right?

Third major problem, whenever you ask a user to do something manually, right, and specifically in this case, we start to create a little bit of user debt.

They have to remember to undo that action.

When they turn Wi-Fi off, because they're connected to a bad hotspot, then when they get home, they have to remember to turn it back on again.

We find about, on a monthly basis, about 60% of our users forget to turn on Wi-Fi at least one time, and many of them are doing this on a weekly basis, at least, and then we also find that, on average, when they forget to turn Wi-Fi back on, they can burn through as much as 300 megabytes of data in one episode, right?

So you think about streaming videos at home, forget to turn Wi-Fi back on, that's extracellular data use that amounts to a few dollars worth of data for a typical user.

We also know that, as a result of some of the UI that's in there, the main interaction mode that users have today is to turn Wi-Fi off, right?

So that can hamper kind of getting to a good state for the user when they get back home.

So today's connectivity manager can contribute to some of these things.

First and foremost, the main interaction mode, as I mentioned, is the toggle.

There's not a lot of data that the user is given to help them decide between networks other than signal strength, and it is relatively frustrating from an automated perspective because it can, in some instances, connect you to poor networks.

Okay, so those are some of the broader user experience problems that we're looking to solve.

And now I'll talk about just some of the initial solutions that we're looking at to help address some of those.

The first of these is called intelligent network selection.

And this is really a foundational piece that we will build upon.

That's something that we want to build other features on, some of the other ones that you'll see in this presentation use this as well.

So to set this up, today, most of the connectivity decisions that are made on the phone, especially regarding Wi-Fi, tend to be based on data that's gathered locally on the device, right?

So signal strength, information on whether the network is secure or not, information on whether it's saved or not, that all influences connectivity decisions.

So I mentioned the Xfinity Wi-Fi example earlier.

Wouldn't it be great if you could tell which of those 20 million hotspots are good and which are bad before you connect it, right?

If the system could tell, it can make more intelligent decisions.

And so we want to start to bring to bear some of the historical network quality information that Google's gathered on all these different networks over time to make better connectivity decisions from an automated system perspective, as well as to inform better user decisions on making network selection decisions.

So that's one important data source that we want to bring to bear across a lot of the connectivity decision-making that you'll see in the next few slides as well.

So in the initial case, what we want to do is primarily apply this to public networks.

As many of you know, most of the growth in Wi-Fi over the past few years has been wide area public networks like Xfinity, Time Warner Cable, AT&T, et cetera.

And so we want to make better use of this technology to make those experiences better.

So typical use case would be, if you're walking around and you have Xfinity Wi-Fi saved, normally you'd be auto-connected to any of them.

This will help you determine which ones are high and low quality and potentially avoid the lowest quality ones of those.

So what would you need to do to enable this for your users?

We'll have more details on this coming up, but for now, what we would like to request is that folks insert this snippet into their config.xml file for their overlay.

And then that more information about kind of the tests and how to implement all those things would be coming potentially in a future integration doc.

But that's kind of the basic level of enabling that score.

I'll talk about some of the features that are built using this data and this information in the next few slides.

So let's talk about the new Wi-Fi settings and Wi-Fi picker.

As part of the broader settings revamp, you know, we're really cleaning up some of the UI UX around settings.

I'm sure you guys have heard about some of this already.

If not, it'll come in some of the subsequent talks.

But one of the things we want to do with Wi-Fi is align with the broader settings revamp and the broader UI UX improvements, which result in, you know, a little bit less information overload for the user.

One of the other concepts we want to introduce is the idea of a friction badge.

And what a friction badge tells you, and you can see we've pulled a lot of the locks and things into the side on the right, is it quickly gives you some information about what it would take you to connect to that network.

Those are things that we will start to add in in the friction call.

And so users can make better decisions before they connect about what kind of, what they'll have to do to actually get access to a network.

So they don't end up connecting and then being surprised when they have to pay $10 just to get access, right?

So those are the types of things that users have asked us for, for information on.

And we want to start to give users better information on how to make those decisions.

Over time, we will improve this and try to add more information on other elements that will help drive this user decision-making.

Okay, another major feature that we want to introduce is something called auto Wi-Fi.

So I mentioned earlier the scenario where, you might get home and forget to turn Wi-Fi back on.

How many of you have, in the past month, forgotten to turn Wi-Fi on at least once?

Okay, all right, so it's a real problem, right?

And if you have kids, they may forget on your behalf too.

So this is something that costs users a lot of money from a cellular perspective.

What we want to do is just make it very easy for users to get Wi-Fi back on when they're near a saved network.

So we know saved networks is a strong intention from the user that they want to be connected to this when it's available.

And so one of the things we can take advantage of is that there are ambient scans already occurring for use with location services that allow us to determine when one of those networks is in range.

So we can just compare the list of networks in those ambient scans with the saved networks list and use that to reactivate Wi-Fi on behalf of the user when those are in range.

Of course, we can also do things like quality checking, things like that.

But a typical use case would be, if I'm out and about and I turn off Wi-Fi because it's not working, sometime later I'll get home and then this will just activate automatically.

So this is a fairly easy fix for an important user problem that, believe it or not, causes a lot of pain for many of our users.

Okay, so one other thing, open Wi-Fi networks notification.

So I have a before and after here on the right.

Many of you probably see this one up top, right?

And what that does today is that will pop for any open network.