Gabor hits Send

When you start to build something new ...

2010-02-08T06:20:00.002+01:00

When you start to build something new, think about the what could be, the what may be, and the what will be. Don't settle, don't give up, don't get stuck in a box built by other people's misguided interaction paradigms. The internet is open and free, and that means there are no rules.

-- Dustin Curtis, The Manifesto

reMAP Meetup at SXSW?

2010-02-05T16:36:00.004+01:00

Joshua Baer (founder of email startup OtherInbox) and me are thinking about putting together a meetup for people interested in replacing IMAP.

We would meet sometime around the SXSW Interactive festival in Austin, Texas (March 12-16). Exact time is still unclear, but Joshua has kindly offered up his company's offices, which are just a few blocks from the event.

Who's interested?

The IMAP Compatibility Matrix

2010-02-05T16:25:00.003+01:00

There's a great discussion going on in the comments of my post about replacing IMAP, and also on YCombinator's Hacker News.

One frequent point made about my proposal is that IMAP already has a number of extensions that do things I suggested, such as the THREADS, SEARCH, and IDLE commands for conversations, search, and long polling ("push").

Commenter Bron (who works on a popular IMAP server) writes:

Also, IMAP plus a random mix of extentions becomes a HUGE compatibility matrix, and you have to offer fallback from all the nice stuff to really shitty workarounds just in case the server doesn't support a feature that makes your life bearable.

So - I'd love to see a decent competitor to IMAP as a protocol. If nothing else because it would reduce the compatibility matrix. Alternatively, I'd be happy for an IMAP5 to be defined which included all the good extentions - so anything which advertised support for IMAP5 had to support them.

Defining an IMAP5 standard might be a good alternative than something like reMAP, and adoption might be much faster. However, a potential IMAP5 would still remain a protocol inaccessible to many developers so it might not unleash the storm of creativity I'm hoping for.

Keep those comments coming.

How to Replace IMAP

2010-02-03T04:08:00.010+01:00

IMAP is a complex and difficult protocol. Its original design, RFC 1730, dates back to 1994. Here's a sample IMAP session that shows you all of IMAP's quirks: It's a stateful protocol where you login and select a mailbox to operate on, it has an unconventional parenthesis-heavy representation for structured data, and fetching messages is a multi-step operation.

Design Outline

One of these days, I'll write up a spec for the mail access protocol of my dreams. I'm planning to call it the "reimagined Mail Access Protocol" (short: reMAP) which is handy because reMail already owns the domain name :-)

I would aim for a RESTful design with the following properties:

All communication over HTTP / HTTPS: Pure TCP connections are great, but for transferring large amounts of email, HTTP is the way to go. Problems like security, parallel downloads, persistent connections, caching, compression, download continuation via ranges, and so on have already been solved. There is no reason to solve them again.

Stateless: There's no reason to introduce state like IMAP does with its selected mailboxes. All you need is a HTTP session cookie used for authentication purposes. Session cookie also allow for things like OAuth. OAuth would let third parties get your permission to access your email without having to give them your username and password.

JSON and UTF-8: All data that's ever sent to or received from the server would be in JSON format. JSON is much more human-readable than XML. UTF-8 would be the only encoding allowed, since it is able to represent any character in the Unicode standard.

Conversations as first-order objects: Gmail, the iPhone SMS app, and Facebook's messaging system have shown the value of viewing messages not individually but in the context of conversations. In reMAP, the server would be responsible for grouping together messages. While you could still access individual emails, the first-order unit of data would be a conversation.

Labels, not folders: Labels are much for flexible than folders. Each conversation should have multiple labels, and the labels would be included when you request the message, rather than having to scan all folders for the message via IMAP.
Stable and unique IDs: IMAP has a UID for each message, but it changes the moment you move the message into a different folder. An IMAP server can also declare all UIDs to be invalid at any moment throughout the session. No more! reMAP would have stable and unique IDs for all conversations, emails, and attachments.

The beginning of the end for MIME: Yes, you could still get the MIME representation of each message that is sent. But MIME is a messy and complex beast. Instead of requesting the MIME-encoded message parts, you could just ask the server to give you the message as represented in plain text or HTML. Attachments can be downloaded in separate HTTP calls.

Push built in: The two prevailing methods for implementing push email are the IMAP IDLE command (not widely available in IMAP servers) and Microsoft's ActiveSync, which requires developers to purchase a license from Microsoft. In reMap, clients could just call an HTTP endpoint on the server which returns as soon as new messages are arrive.

Full-text index on the server: reMAP servers would need to maintain a full-text index of the contents of all messages. There's no reason clients should be required to download and index everything in order to do an exhaustive full-text search of your email.

Enormous Potential

I believe that one of the reasons for the lack of innovation in the email space is the lack of a simple yet powerful email access protocol. Every developer that wants to try something with email needs to first jump through the hoops of IMAP and MIME, or worse, the Outlook Object Model and MAPI. A new protocol like reMAP would lift this burden off their shoulders. We've seen what open, simple standards can do for innovation with Twitter's and Flickr's API. Now imagine unleashing the same sort of creativity to the vast ocean of data that is email.

Let me know what you think. My goal with this post was to encourage a discussion about this topic, and your comments are much appreciated.

If you liked this, consider following me on Twitter where I often tweet about email-related topics!

"A good email client, please."

2010-02-02T20:46:00.001+01:00

My friend Patrick Collison writes on Facebook:

"Gmail proved that, despite the apparently high switching costs, a new webmail client can quickly get a lot of traction. There’s room now to do to Gmail what Gmail did to everything else. The replacement should have some concept of workflow (”archive, but remind me to respond tomorrow”, “send, and alert me if I don’t get a response within a few days”), some concept of teams and colleagues (allow threads to be shared as a first-class object, rather than flailing around with forwards and CC lines), be some way smart about mining the semi-structured mails going through the system (flight booking emails should be automatically annotated with .ics files), know something about prioritization (I like Twitter DMs because of the assumption that they’ll go to a mobile device. If I’ve sent more than 20 emails to someone, they should have the option of copying their mail to me as an SMS).

Potentially most powerfully of all, developers should be able create their own plug-ins that run on the server. There should be an agreement between plug-in developers and the webmail provider that creating a plug-in automatically grants a royalty-free perpetual irrevocable worldwide (etc.) license to the provider, and that the source code to any plug-in may be merged into the main product. Though plug-ins have niche appeal, this could be a good source of new features, and a strong competitive advantage. I’d just fix Gmail if I could.

I’d happily pay for any service that got this stuff right."

Some good ideas in here. Server plug-ins could work in a manner similar to Google App Engine. A lot has been tried around better triage and data mining, with no clear winners yet. Also, there's a danger of feature overload somewhere in here. I'm not sure if we should replace one variant of feature overload (Microsoft Outlook) with another (integrated SMS / Twitter / Facebook / email). Maybe SMS, Twitter, Facebook, and email should instead all just be replaced with one thing that is the combination of all lessons learned with those protocols.

Mobile Advertising needs Rich Media and Better Targeting

2010-02-02T19:46:00.004+01:00

Mobile advertising is a $2.2 billion business. Advertisers put up mobile ads for two reasons: (1) Building brand awareness and (2) generating leads, whether to download their iPhone Apps or sign up for their service.

Two possible improvements:

Two possible improvements:

Rich Media

All ads you see today on your phone are static text or images. Brand advertising works by playing to emotions, and to do that, you need animations and sound. I know you dislike those Flash ads on websites, but they're how the content gets paid for.

Those ads are static for a reason: iPhone doesn't support Flash, and the ads need to be small in size due to limited bandwidth. I'm sure one of the players in this space is already working on building technology to show great-looking, animated, interactive ads - whether with proprietary technology or some sort of browser plugin. Rich ads, rich CPMs. Even the bandwidth problem can be solved: There are plenty of ways to preload ads and cache them for later display.

Better Targeting

If you've tried putting up an AdMob ad, you know that the targeting options are pretty limited: You can narrow down by country, device (iPod Touch / iPhone), and connectivity (Wifi/3G).

There's one type of ads that could unleash dramatic growth in mobile advertising: Imagine you're on the way to lunch at Pete's Burgers, and your iPhone shows you a coupon for Joe's Burger Shack. You go to Joe's and show them the coupon you just clicked on. The ad just generated a verifiable lead, and the advertising network just made a bunch of money.

For this type of application to work, targeting needs to get better: We need street-block level targeting, time of day ("only run this ad at lunchtime"), and likely user activity ("user likely walking to lunch").

I imagine that local ads like that could be bigger than the $20 billion / year search advertising business (e.g. Google AdWords), which makes money by generating leads for people to buy stuff online. The number of real-world transactions and the amounts I spend on them are much higher than what I spend online. Done right, mobile advertising could be much bigger than that.

One of these Days

2010-02-02T04:12:00.004+01:00

One of these days, I'll go through my Blogger console and just hit "Publish" on every one of those unpublished drafts I've been nursing. Many of them lack just one final insight, a few just lack a little bit of polish. No reason to pull a Hank Moody [*] on you guys.

Just post it or nurse it to quality? That is here the question.

[*] And by Hank Moody, I'm referring to writer's block, not bad parenting or tooling around in a Porsche.

A Logical Fallacy

2010-01-29T20:51:00.003+01:00

"Every product will have positive and negative criticism. It's fallacious to then leverage one [piece] of commentary to suggest a completely unrelated product 10 years later will be successful because people were saying negative things and that was wrong before."

Nicely put by Anthony Licari on Garry Tan's blog.

Will there be an iPad App Gold Rush?

2010-01-27T23:33:00.007+01:00

In the iPad promo video, Apple VP Scott Forstall predicts a whole new gold rush for iPad apps.

For games, this is true. People will buy and enjoy games on the big screen. But what about other apps?

Excluding reMail, the apps I use most on my iPhone are the New York Times app, Foursquare, Facebook, Tweetie, Yelp, Kayak, Skype, Snow Report, and AT&T myWireless. Almost all of these are there to display content and do social networking.

But why do I use the New York Times app instead of the website?

The NYT app lets me get to the news faster: On my iPhone 3G on Wifi, Safari loads nytimes.com in 55 seconds. The NYT starts up and shows me the latest news in 18 seconds.

The NYT app is optimized for the small screen - the web version isn't.

The iPad loads nytimes.com instantly. Daring Fireball says the iPad is "fast, fast, fast". That makes it much more attractive to skip writing an app and just use HTML 5:

Web apps are an order of magnitude easier to develop than iPad apps.

Apps are substantially slower to update given the iPhone approval process and waiting for users to get the update. [*]

It's easy to make a version of the page optimized for iPad's screen and the size of your finger.

HTML 5 will let you offer offline functionality.

Sounds familiar? Yes, Google's Vic Gundotra has been saying that web apps are the way:

"We believe the web has won and over the next several years, the browser, for economic reasons almost, will become the platform that matters and certainly that’s where Google is investing."

For the iPhone, this hasn't come true yet. But the iPhone's browser are too slow. On the iPad this might change - instead of native apps, we might see a thousand web apps bloom.

[*] Thanks to commenter Tom Pickney for pointing this out.

More on iPad: What the iPad means for Developers

What the iPad means for developers: Spending quality time with UISplitViewController

2010-01-27T22:42:00.008+01:00

"The last time there was this much excitement about a tablet, it had some commandments written on it." - Wall Street Journal.

In the days before the iPad release, a bunch of app developers will be rushing to make their apps iPad-compatible. In the iPad promo video, Scott Forstall predicts a whole new gold rush.

What do most non-game apps have in common? A main list view and a detail view. Whether it's the New York Times app, Facebook, Snow Report, they all follow the main view - detail view pattern.

This is the natural organization scheme for iPhone apps: UINavigationController and UITableViewController make this easy, and there is no shortage of documentation on how to build apps like this.

We've all seen the screenshots of the new iPad email app. List on side, main view on the other. Portrait mode switches the navigation view on the left to a button on the top left that pops over the content.

All those main view - detail view apps? They'll be redone in this style. It looks like we'll all be spending quality time with the new UISplitViewController, which manages the presentation of the side-by-side panes:

The NYT and Facebook apps should be available in this style on iPad launch day. But they might be separate from their respective iPhone apps: Apple has also announced support for "Universal Applications" that run on both iPhone and iPad, but this is not yet in the 3.2 SDK - I guess until then, there be a separate part of the App Store with "Made for iPad" applications.

World of Warcraft has just 32 developers for 5.5M Lines of Code

2010-01-26T19:46:00.003+01:00

I don't play it, but World of Warcraft is the world's most popular massively multiplayer online role-playing game with more than 11.5 million users. It clocks in at 5.5 million lines of code. According to this blog post, there are only 32 developers working on this code, an impressive ratio of code to coders.

"Population [of development team]: 32. [...] Whole team maintains 5.5 million lines of code. [...] 20,000 computer systems, 1.3 petabytes, more than 4600 people. Operating an online game requires more than just game development!"

Discovered on Sanjeev's FriendFeed.

"After a few pitches, entrepreneurs realize that the distant future is safer territory than the immediate."

2010-01-21T22:09:00.001+01:00

Check out "How to raise money without lying to investors" on VentureHacks.

How Startup CEOs Need to Evolve over Time

2010-01-21T18:42:00.003+01:00

Here's a great post by startup CEO Steve Blank on how entrepreneurs need to adapt to changed requirements over the course of a startup's life:

There's a great discussion about the perceived benefits of sticking with the founders or firing them from the board's / VCs perspective:

"Looking at the abrupt change in skills needed in the transition [from early stage to large company], it’s tempting for a board to say: Maybe it’s time to get more experienced executives. If the founders and early executives leave, that’s OK; we don’t need them anymore. The learning and discovery phase is over. Founders are too individualistic and cantankerous, and the company would be much easier to run and calmer without them. All of this is often true." [...]

"Time after time, startups that have grown into adolescence stumble and succumb to voracious competitors large and small because they have lost the corporate DNA for innovation and learning and discovery. The reason? The new management team brought in to build the company into a profitable business could not see the value of founders who kept talking about the next new thing and could not adapt to a process-driven organization. So they tossed them out and paid the price later."

Why Cars in the US Should Come with Aspheric Rear View Mirrors

2010-01-20T21:13:00.005+01:00

If you've ever driven a car in Europe, you've probably noticed that the rear view mirrors are curved to eliminate blind spots. This is called "aspheric". Meanwhile in the US, all cars are shipped with flat rear view mirrors. Even European-made cars like my 2008 Volkswagen Rabbit (assembled in Wolfsburg, Germany) are sold with flat rear view mirrors like this one:

I mentioned this to my brother in Germany. He walked into a VW dealership and found out that given your VIN, any dealer in Germany can track down the exact specs of your US-equipped VW. I found the global trackability of my car pretty surprising. As a Christmas present, I got this:

See how the curved glass lets you see my hand with the camera? Having aspheric rear view mirrors decreases the size of your blind spot, as illustrated here:

Someone mentioned to me aspheric rear view mirrors are illegal in the US, but I haven't been able to confirm this. It wouldn't make sense: The US must be the country with the highest number of lane changes per capita per year. I haven't been able to find a vendor of these mirrors in the US, but next time you fly to Europe remember to bring your car's VIN.

Simplicity Wins in the Top 25

2010-01-20T03:50:00.005+01:00

Most of the Top 25 free apps on the App Store are games. But what kind of games are the winners?

You'd think that the Top 25 free Games would be high quality, high production value, immersive, made with love. Things like the airplane fighting game Skies of Glory:

Or how about Skater Nation, where you get to skateboard, grind, and jump like the pros?

You'd think these games would be the winners. But you'd be wrong. It turns out the games in the Top 25 is mostly simple, OK quality stuff. FallDown (at number 9), makes you move the ball down the screen faster than it scrolls the background. A great iPhone developer could write this game in a few days:

Another incredibly simple game in the Top 25 (at number 20) is Traffic Rush. All you need to do is keep cars from crashing on an intersection. Once again, super easy to write - a good programmer and a decent graphic designer could crank this out in half a week.

So why is it that these simple games win, and the high-dollar productions lose, at least in terms of number of downloads? I believe it's the iPhone audience and how these devices are used:

iPhone users are casual gamers: this is no PSP Portable. They just want to waste some time on a subway ride, and don't care for learning the ins and outs of an airplane fighter game.

Anything that takes more than 30 seconds to learn falls by the wayside.

Each level should take no longer than a subway ride. Gameplay that develops over blocks of 10 minutes and longer is a non-starter.

Startup time is a huge issue: It easily takes 20 seconds just to load Skies of Glory. Unacceptable to most users, since they know that their gameplay could be interrupted at any time by a call or text message.

The take-away? Simplicity wins. Dumb it down, make it load fast, gratify immediately, keep the gameplay fast-paced, keep each level short. And they will come, download, and play.

Wow

2010-01-13T02:21:00.005+01:00

I nearly fell out of my chair when I read about Google's new approach to China. It's hard to contain my excitement about what I feel is doing the right thing. Go Google!

I'm sure in the next days we'll see a lot of speculation. Google has already been accused of doing this for business reasons, not noble intentions. Yet, any way you slice it, this is as hard a stance on free speech as a company could possibly take. I'm happy to see that Google decided to put its foot down. Very proud of my former employer right now.

2009 in Review

2010-01-04T16:48:00.003+01:00

It's a "Gabor hits Send" tradition to write New Year's wrapup post. Here's what I wrote about 2008 and 2007.

I feel like the last year was a lot of work, and very little fun, but my travel map proves otherwise: I went on trips to Tahoe, LA, Las Vegas, Utah, Alaska, Istanbul, and Rome, not to mention seeing my parents in Switzerland. Not quite like 2008's trip around the world, but certainly not the celibate life I remember living.

At reMail, we built reMail 1.0, which was a failure, but a taught me many lessons, including: (1) Users won't give a startup their email credentials, and (2) we should've launched even earlier. While waiting for the App Store approval, I built reBoxed, a project that was fun to build but wasn't ultimately useful enough. I then built reMail 2.0 with some contractors, and it was a success. I spent the rest of 2009 refining it, releasing a new version of the software every 2-3 weeks.

When Paul Graham asked me about surprises I encountered building a YC startup, I said that the biggest lesson learned was "Fast iteration is the key to success." I think that sentence was my theme for 2009.

Walt Disney

2010-01-03T22:18:00.004+01:00

I visited the Walt Disney Family Museum in the Presidio yesterday. The museum recounts Walt Disney's life and how he built the Disney empire.

One thing that really stood out is just how many failures he had to endure. Before the Disney Company, Walt had started another business, Laugh-O-Gram. It went bankrupt. In the late 1920s, Disney's hit character was Oswald the Rabbit. But the rights to the character were owned by Universal, not Disney, and they lost the rights to their money maker. After losing Oswald, Disney had to come up with something new and ended up creating a character called Mickey Mouse. Similarly, Disneyland's opening day was a huge disaster. Disney's story is a story of how persistence pays off.

The Mother Lode of Data on the Mobile Internet

2009-12-30T12:50:00.004+01:00

Browsing the Internet this morning, I found this Morgan Stanley report on the Mobile Internet. Depending on which version you look at, you'll find hundreds of slides packed with data and insights on mobile internet usage. I recommend you at least flip through the this short 92-slide version.

Here are some of the highlights - this is the data I found most interesting.

iPhone Growth

The growth of iPhone + iTouch outpaces that of Netscape, i-mode, and AOL. It's more explosive than anything we've seen so far.

iPhone and iPod Touch are growing at the same rate This slide was meant to demonstrate the explosiveness of the iPhone platform, but another thing it demonstrates is how iPhone and iPod Touch sell around the same number of units, and have done so consistently even through the introduction of the 3G and 3GS.

Web Usage

Unproductive sites are increasing their addiction levels. Online global time spent is trending heavily toward Facebook and YouTube. MSN and Yahoo are shrinking away while Google (probably the most work-related of all these sites) is holding steady. I wonder about the effect of all this on global GDP.

Google now makes $20 per user per year in ads. I still remember when my friends were asking "Who clicks on all those ads anyway?" Somebody does. Google's annualized revenues per user have increased from $10.22 / year in 2005 to $20.06 in 2009. That's a large chunk of the total ad revenue per user on the Internet, which is $46.41. Wow.

PS: Morgan Stanley has done an outstanding job in assembling all this data. I just wish they'd hired a graphic designer for their slides - they do look a bit busy, especially pasted at small sizes.

I'm off to Europe

2009-12-22T02:07:00.003+01:00

I'm off to Europe for the holidays to spend some time with my family.

I heard that Switzerland is totally covered in snow. I heard my brother got snowed in the other day - quite unusual in a country with the level of road services that Switzerland has. I'm certainly looking forward to doing some snowboarding with the Alps!

Mobile Email Usage to Grow from 131M to 434M in 2 Years

2009-12-15T07:31:00.006+01:00

I just found this great report [pdf] by email research group Radicati. It comes with bucketloads of stats and predictions on email usage.

In particular, this one table is a gem. According to Radicati, mobile email usage is poised to grow from 131 million users today to 434 million in 2011.

That's pretty substantial growth - almost 100% growth year-on-year! I'm happy that's the space that reMail is in.

The In-App Freemium Model

2009-12-15T07:26:00.002+01:00

We're now letting users download reMail for free and then upgrade to IMAP support and no ads. Previously we were charging $4.99 upfront, now we're selling features via In-App purchase. I think many useful apps will eventually charge this way. For example, navigation apps might charge you per city or per routing instead of charging $80 upfront. Games have already shifted to selling virtual goods and levels.

Initial sales are encouraging. I'll let you know how it worked out for us once I have more data.

Read more about how we implemented freemium on the reMail blog.

reMail 2.7 is Out

2009-12-11T04:59:00.003+01:00

We released reMail 2.6 late last week and it had some bugs. Bugs we should have caught in our QA department [*]. Sorry about that.

However, we fixed those bugs ASAP and resubmitted a new version. reMail 2.7 is out on the App Store now and you should update.

[*] We don't really have a QA department. Yet.

Shocker: A Drop in Email Overload!?

2009-12-07T20:52:00.004+01:00

The Radicati Group is a research organization owned by HP that surveys email and messaging usage patterns. Here's a shocking new blog post:

Our newly released Business User Survey, 2009 shows that for the first time since we started monitoring email traffic patterns, the amount of email that reaches business user inboxes is actually decreasing. Survey respondents indicated that they sent and received an average of 108 email messages per day in 2009, which is noticeably lower than the average of 140 email messages sent and received in 2008. This is a fairly significant decrease of 23%!

Their explanation for this data is that users are shifting to other means of communication such as IM for certain type of messages: "Wanna grab lunch?"

This data does not fit my observations. For me, 2009 was the year of notification emails, as illustrated here and here.

My alternative explanation: This is a survey-based report. People generally overreport on how busy they are. They've recently become more realistic.

Now if only I had $2500 to buy that research report.

The Vision for reMail

2009-12-06T23:04:00.004+01:00

As an entrepreneur, there's one question you get asked a lot: "What's your vision?"

Unfortunately, I don't have a beautiful answer like Larry & Sergey's "Organize the world's information and make it universally accessible and useful." That's a powerful one. So powerful that Eric Schmidt apparently questioned their sanity when he first heard it.

For reMail, the goal is to solve the two big problems in mobile email:

You're on the way to a meeting and you need to look up where it is, or who you're meeting. This is an email search problem.

You're walking from the meeting room to the bathroom and have 45 seconds to catch up on your new messages. This is a prioritization problem.

These two problems are not orthogonal: You can see prioritization as a special case of search. You're searching for important messages. You can see search as a special case of prioritization: You want to see the important messages related to your current state of mind, as expressed by the query. The solutions aren't as clearly separated as the use cases are. And that's good for reMail.

We started with search because it's easier to solve, and the value is clearer to the user. We've built a pretty successful product - it needs a lot of refinement but it fills a clear need for users. I've built some sketches of prioritization tools in May - reBoxed and something I called "reMail Commander" - but they need a lot more work.

We were so luck to focus entirely on mobile email. Mobile email usage is growing and there's no reason why mobile email usage shouldn't eclipse desktop usage in 5 years or so.

Sounds smart? This decision fell into our lap: In December 2008, we had a meeting with a potential investor (he didn't invest and probably wouldn't like to see his name here). We had plans to do stuff on the web, desktop, and mobile. During that conversation, it became clear to us that on the desktop we'd get killed by Outlook 2010, Postbox, Zimbra, Thunderbird, and many players with deep experience. In webmail, we'd get killed by Gmail Labs. Mobile seemed like the right spot, with lots of whitespace and huge problems. I'm happy that's what we decided to focus our efforts.

Say it with me, all together now: reMail is reimagining mobile email.

Thanks for your time.