Gmail: Expect bigger changes in next 5 years

Five years ago, Gmail launched with a splash big enough that many thought it was an April 1 joke: an entire gigabyte of online storage.

Larger online e-mail rivals Hotmail and Yahoo Mail quickly matched that advantage, but in the meantime, Gmail has grown to become a force to be reckoned with. It's got tens of millions of users, Google said, though it won't pin down a precise number. And its growth today, in terms of new users joining the service, is faster than it was four or five years ago, said Todd Jackson, product manager for Gmail.

In a chat Monday, Jackson offered an assessment of what Google has accomplished with Gmail thus far and what it expects in the future.

Todd Jackson, Gmail product manager

(Credit: Stephen Shankland/CNET)

CNET News: First of all, you launched Gmail on April 1. Are you going to have an April Fools' stunt this year?
Jackson: Keep your eyes peeled. I can't tell too much. I can't spoil the joke.

The 1GB in-box was pretty surprising when Gmail arrived. Do you expect anything new from Gmail that's that shocking or paradigm-shifting?
Jackson: We've been working on innovating Gmail over the last five years. It's our goal to stay constantly on the leading edge of what users want--particularly the most demanding users. When we added chat to Gmail, I considered that a big milestone. Similarly when we added video chat. I thought these were really important in expanding the scope of communications that Gmail makes fun and possible and easy and fast.

Communication is more than just mail. It was a really good ground for us to start on. We want to stay on that bleeding edge. Gmail Labs is a good testing ground to be trying new things and getting stuff out there to the public fast even at the scale we're at. It becomes difficult at the scale we're at, with a large user base, to launch things at the same speed as when you were small. We want to think of ourselves as the start-up that happens to have tens of millions of users.

I think we got a lot of that big bang at the original launch because of the gigabyte of storage. That was the hook that got a lot of people interesting in checking Gmail out, but then what got a lot of them sticking with the product were things about the UI (user interface)--conversation view and search and the quality of the spam filter. All those things that don't add up to the same headline, but they're the things that really make the product great. We're going to be going for more of that.

Do you think the difference between Gmail at launch and today is going to be less or greater than the difference between Gmail today and where it's going to be in five years?
Jackson: Many of the things we've been working over the past five years were under-the-hood things. Things that don't dramatically change the visual look of the product but really people expect to have in a mail product. Things like POP support, IMAP support, a mobile UI. Little things like save draft or rich-text editing. We didn't have any of that at launch. You couldn't boldface, you couldn't italicize. We've been adding these things over the years that people just expect to have.

We've also been investing. The big change to our front-end infrastructure that we launched the year before last allowed us to have a larger number of engineers contributing to the code simultaneously and allowed front-end development to go a lot faster because of the new modular JavaScript architecture. It also made things like Labs possible because it allowed us to serve different modules of the code to different users. I view that as a huge enabling technology.

I think over the next five years you will probably see a large amount of visible change, maybe more so than in the past five years. That's because for the first five years we had to focus on all the nuts-and-bolts things people want. We did some very innovative thing in terms of chat and video chat and expanding the number of ways to communicate in Gmail. I think you're going to see more things in that direction, and things that directly impact the way the product looks and feels.

You mentioned communication. Gmail has added some instant messaging, but at the same time we have Facebook messaging and chat, Flickr Mail, Friendfeed. We have all this communication going on outside our in-box. Will Gmail expand to accommodate all that so we get back more of a unified communications hub?
Jackson: I don't want to speculate on actual features. But I will say it's a core mission of Gmail to be a powerful tool for all the ways a user wants to communicate. We know that goes beyond mail. This is something that's very much on our minds. We started with e-mail because it's something everybody used and we saw a lot of room for improvement. We're going to be similarly looking for other things like that as new communication technologies emerge.

What do you think about a tie-in with VOIP providers?
Jackson: I think that would be very interesting--potentially something with Google Voice. I don't want to strongly commit to anything, but (Google Voice integration) is something that strongly interests us.

IM was one of the components of the partnership between Yahoo and Google announced last year. Is that going to happen, or did that fall off that map when the partnership with ad sharing fell apart?
Jackson: I don't want to go into details there. We're happy with the partnership we launched with AOL Instant Messenger. We think that IM integration and federation is a good thing. I'm very happy with the way our chat implementation uses XMPP. iChat and various other clients talk right with Google Talk, which is a direction we like.

Have you moved out of the early-adopter category into the mainstream? Would you characterize your average user as a techno-savvy person or a regular person?
Jackson: That's a good question. We started with the early-adopter crowd. That was on purpose. We wanted to build a product for people who were getting hundreds of e-mails a day, because we believe by focusing on the power user, you're designing the product the rest of the market will want in a couple years when everyone's usage habits catch up to the most active users. We pay most attention to seven-day active users (those who use Gmail at least once every seven days) and usage--the amount of actual engagement with the product. Something that Larry and Sergey (Larry Page and Sergey Brin, Google's co-founders) are always, always telling us is to focus on usage rather than users. That's what matters more. You get better feedback and you are properly kept more on the leading edge if you're focusing on the people who are using the product all the time, using the product all day, than just the casual users.

We've reached a lot of the everyday casual users, and that's great. We're happy to have as many people as possible use it. But we're really focused on the power users and the early adopters.

So do you think you've escaped the tech-savvy niche?
Jackson: I think so. We're available in over 50 languages. More usage comes from outside the U.S. than inside the U.S. Our growth today is faster than it was four or five years ago...in terms of actual number of users using the service...and we're growing in all countries.

I haven't noticed any Gmail Labs features make it into production yet. Is that going to happen at some point with Superstars or something?
Jackson: Yes. We definitely want to graduate some features soon. We look at a combination of things. We look at install counts and uninstall counts and the qualitative feedback we're getting. One of our big goals is to graduate offline (in which Gmail messages can be read, searched, labeled, and read even with no network connection). There are other things I think would be strong candidates, like undo send and they YouTube, Picasa, Flickr, and Yelp previews. Superstars are a pretty fundamental tweak to the UI. I imagine some users could be confused by the red exclamation mark. That's a tradeoff.

It's great to have a place like Labs where you can experiment with these things and throw them out with reckless abandon. We have a very high quality bar for things we are going to launch to all users and promote to all users.

You changed labels without going through Labs.
Jackson: Sometimes we'll do that. For something like labels, we've tested that so much internally with 20,000 Googlers, it's so core to the workflow, we thought it was sufficient, and we were ready to launch it. We did that with themes also. We did that with video chat.

Gmail has a different look with labels and conversation view instead of folders and a traditional in-box. Do you ever think about doing something that looks like what people are used to with Outlook or Yahoo Mail?
Jackson: The angle we take is what are the problems users have, and what is the most efficient way of solving that for users. That's how we arrived at labels rather than folders and at conversation view--and search, for that matter. These were tools that were very very good at handling large volumes of mail. For labels, we added the new move-to button. We have more improvements planned for labels. We understand that users very familiar with the folder model take a little onboarding time to get them used to labels, so we have some stuff planned to make labels more accessible to everyday folder users and also more powerful for people who like and understand the model.

We don't set out to copy or pay too much attention to competitors and what they're doing. We focus relentlessly on the user, particularly the power user. I would bet over time our UI will look more and more different than the other guys rather than more and more the same.

Well, you could say here's my flashy new car. It turns out that objectively it's better if you put the clutch pedal on the right, the brake pedal on the left, and the accelerator in the middle. Objectively maybe that's better, but if you're focused relentlessly on the users that might be a very bad idea. There's always a balancing act between a clean slate and fitting in with what people are used to. But it sounds like you're going to diverge more than converge.
Jackson: I agree with the tradeoff you mentioned. We definitely will use familiar paradigms where they make sense and we think they're powerful. But our No. 1 priority is going to be focusing on the needs of the most active users and giving them the tools they need to communicate as the amount of information increases. I tend to think that in a couple years the tools we'll use for that will look pretty different from the tools we have today.

So the Android phone has a Gmail app. Do you think about releasing a browser for Windows or Mac OS X?
Jackson: There's also the J2ME (Java Mobile Edition) app which runs on Blackberry. We try to take a pragmatic approach. Develop the things that have the best user experiences and are available to the most users. So far in terms of the mobile market, we're seeing a proliferation of new platforms, with iPhone, the soon-to-be new Palm platform, I think Blackberry is coming out with a new app platform, and with Android. All these phones have modern browsers, and the browser innovation is increasing. Potentially with things like HTML 5 we can get a very client app-like experience right in the browser, and then it will automatically work on all these platforms. That's what Google's done in the past, to bet on the browser. But we're open-minded about it.

The browser app you get to day when you visit Gmail on your iPhone isn't as good as it could be. We're investing a lot of energy in that. We haven't yet explored some of the faculties that HTML 5 gives us.

It looks like we'll see some JavaScript performance increases in iPhone OS 3.0. Will that help you out?
Jackson: Absolutely. Any browser advance that makes JavaScript execution faster, we love. It's amazing to see the innovation that's happened in the last couple years in Firefox 3, Chrome, and now Safari 4. Gmail is blazing fast in the new Safari 4 beta. It's really enjoyable to use. This was a big bet we made a couple years ago when we redesigned the entire front-end architecture (the Gmail interface). We did it using modern coding practices. It was a modern implementation of a JavaScript app. We were really betting on the fact that browsers would get better and better in JavaScript performance.

At the same time you have to support the browser that has two-thirds of the market (Internet Explorer). Do you have a divergent code base depending on which browser is use, or is it mostly the same application?
Jackson: It's mostly the same app. When we launched the new Gmail JavaScript architecture, it didn't yet work on IE 6. It worked on all versions of Firefox, IE 7, and Safari 3 and above. We've now expanded to cover IE 6 and all the new versions of all the browsers. From a team efficiency point, we want all the engineers working on Gmail to be working on the same code base and a single implementation. We'll still supporting that older version of Gmail, but the number of users using that is pretty low at this point and keeps going down.

Do you expect to be able to do radical changes to the UI because of improving JavaScript?
Jackson: There are a couple things that will let us do radical things with the UI. One is just the new modular infrastructure we have. It makes it a lot faster to develop different UIs and experimental UIs. That's going to drive change fast. And as browsers get better and faster and to have potentially other hooks--extensions into the desktop that browsers haven't traditionally been able to do--that's going to drive a ton of innovation.

Can you be more specific about hooks you're interested in?
Jackson: You can imagine using capabilities of the browser to do things that are more integrated with the desktop. For example, being able to drag photos from your desktop right into your browser. As the browser gets more advanced and becomes the main distribution mechanism for all these services, I think you'll see innovations that lead to a lot more speed and a lot more integration with the desktop.

When I click that contacts tab, I go away and wait for it to load and come back.
Jackson: We are fanatical about performance. If there's anywhere in the app that isn't performing, where pages aren't blazing fast, you can imagine it's something we're working on. I agree the contact manager is slower than it should be, and it's something we're working on.

Are you ever going to limit the number of filters? I'm sorry, you have 50 filters, if you want more you have to pay for the premium version?
Jackson: We want to keep as unrestricted as possible, because we know they're really useful in getting through your mail.

Is spam an increasing or decreasing problem for the service--for what you have to deal with?
Jackson: The amount of spam on the Internet is increasing rapidly, but the amount of spam reaching your in-box is decreasing rapidly because of all the effort we're putting into those systems. Spam is a growing problem on the Internet, but the tools and filters we are developing are getting better and better at weeding it out.

But for you, do you think the tide has turned and you are conquering the spam problem, or do you have to spend ever larger resources every year to stay level?
Jackson: I would never be so presumptuous as to declare victory over the spammers. There are a lot of them, and they're smart. From day one, we've been putting some of our best minds to this, developing highly automated, scalable tools that leverage all the crowdsourcing of our users reporting spam. We're going to keep doing that--investing in tools that have broad effect.

We want the false negatives to be zero and the false positives to be zero. As a user if you have check both your in-box and your spam folder, what's the point?

Do you use AdWords technology underneath that to select ads based on the message content?
Jackson: Yes, mostly we use AdSense technology to extract the topics in your e-mail and give you relevant ads.

What's the click-through rate on the ads?
Jackson: I can't give specifics. It's pretty comparable to what it is on the content network. It's a good source of revenue for us. We do actively work on ads. We have a number of ads working on making ads better on Gmail. We want them to be more useful to users, and we want to make more money on them.

There's this coverage question--where do you show an ad vs. not showing an ad. The search guys are pretty careful about it because they don't want to train people to be ad-blind from ignoring low-quality ads. When I look at Gmail ads, sometimes they're spot on, but sometimes they're somewhere between laughable and inappropriate. One co-worker who does animal rescue gets ads for puppy mills. Is Google looking as carefully at whether ads are improving the customer experience in Gmail as they do on the search side?
Jackson: Yes, we are. I agree. The example you brought up--for a company that prides itself on organizing the world's information and serving you stuff that's really relevant, we should be doing better than that.