Saturday, April 09, 2005

Google, Maps, Mail and Microsoft

Coloured pools near Napa, CA. I wasn't the only one impressed by Google Maps' new satellite imagery; it's developed quite a buzz.

Matt Haughey talks about soil chemistry and why the south Bay is bordered by coloured pools. (I'd always wondered about these pools near Napa, which you see from the plane on the way into SFO.) Over at Flickr, they're using Google images to make memory maps. And Google Sightseeing collects landmarks, under the tagline "why bother seeing the world for real?"

Robert Scoble, typically, gives us a Microsoft tease:

Now, I've seen Microsoft's future mapping strategy and I've been sworn to secrecy. Don't count us out yet. After all, we have TerraServer and a few other things that work well on maps.

For instance, here's the front of the building I work at on TerraServer. Here's approximately the same thing on Google Maps.
Well, OK, Terraserver. Microsoft got there first. But who cares? Google Maps is attractive not because it was first (there's been mapping on the web for years) but because it's the easiest to use. And Google Maps is here now.

Google's secrets: it's very good at reliability, speed, simplicity, and cleanliness.

This has always been true of the Google home page, which is spare, attractive, and fast to load: compare Google with, for example, the cluttered Yahoo! home page. When you go to Google, you go there to search, and you want to search right now and as easily as possible. Google understands this. The page is simple, the focus is the search box, and yet the branding is strongly and immediately recognisable.

A past project of mine was a driver layer for a web-browsing set-top-box. We always demoed Google, because Google always loaded fast (especially once you've got the logo in the cache); it always looked good (even in 8-bit-per-pixel graphics); and it's obviously useful (even though for demo you pick the search term carefully to lead to results which you know you can render well). It turns out the simple design, white background, and strong colours of the Google home page are ideal for display on a television: "good enough to lick" was how we described it.

Interestingly, while the MSN homepage is cluttered, the MSN Search homepage has learnt the Google lesson. It's simple, colourful, search-focussed, and branded (the MSN colours and butterfly, and with a hint of the default Windows XP theme).

Anyway, Google's taken "fast, simple, clean" and applied it to maps. Google Maps is exceptionally easy to use. The map is very immediate, and encourages exploration and zooming. And the search is simple and broad: where other mapping sites insist on street, city, state, zip being entered in separate boxes, Google provides a single search box and does its best to understand whatever you put in it.

Google understand that if you're looking for a map, the map is the most important thing. Their maps are exceptionally clear. They occupy most of the browser. And they scale with the browser, so if you make the window bigger you get more map. Other mapping sites give you a fixed-size square of map and surround it with navigation and advertising, which is frustrating: exploring the map is like looking through a porthole.

And Google's no-reload trickery — drag the map and it'll fetch the newly-exposed times without reloading the page — is a masterstroke. Maps are about exploration, and Google's technology allows for seamless and uninterrupted navigation. Most significantly, the continuous scrolling means I can easily follow a highway without losing my place on the map.

This isn't wholly new, of course; Java-based maps have been doing this sort of thing for ages. In the UK, I was always impressed by the RAC's dynamic route-planner map (example), provided by Map24; and by's street maps (very good for getting around Paris; and guess what, they can also overlay aerial images). But again, Google have done it very well: they've made it easy and fun to use.

So, Terraserver: it's OK. It does have more detailed images. But it makes me go through a multistage search to find my home address; it shows me images through a porthole (although to be fair it does let me make the porthole pretty big); and it forces me to jump-scroll the map. Yes, it was there before Google. But in terms of usability, fun — lickability — it's now playing catchup.

Scoble further speculates:

Now, what's possibly next from Google (or MSN or Yahoo)? Use your imagination. What would you like to put on top of that map? Er, image? I got a few things. How about a sushi icon? Huh? Click the sushi icon and it takes you to all the nearest sushi restaurants. How about a camera icon? Click on that and it takes you to all the nearby photo opportunities. How about a Hospital icon? Click that and it takes you to the closest hospital. How about a blog icon?

How about a Flickr icon? A McDonalds icon? A Starbucks icon? A Sears icon? A Scobleizer icon?
How about clutter? How many different icons can you fit on that page? Google have already got it right: if you want it, search for it, and they'll put pushpins on the map for you.

Microsoft, along with everyone else with a webmail platform, is also playing catchup with Google's GMail platform. Not just in terms of space (where they're chasing GMail's 2-Gigs-and-counting allowance; good PR for Google, who know that most people use nothing like this much space) but in terms of usability. GMail is just very, very, slick.

Scoble asks a leading question:

The real question isn't who'll be first to offer two gigs of space (or the coolest AJAX UI implementation). The real question is: who'll do it for more than 150 million users first?
As a user, I disagree. How many users GMail has matters not a jot to me: what matters is that it's usable and available, and so far it's done very well on both.

And I strongly suspect that scaling is a strawman here. Why wouldn't GMail be able to serve 150M users? Google are exceptionally good at back-end scaling (taster; detail), if nothing else because they recognise that speed of response is a key attribute of their search-engine offering. It's hard to imagine that they won't be able to bring that expertise to bear on mail. (Rich Skrenta of speculated on this when GMail launched.) Google is, in fact, so highly scaled that cost and availability of power is an issue.

So, I'll stick my neck out and predict: 150M GMail users, no problem. And it'll still be bigger and slicker than Hotmail, which is what Scoble's trying to talk up here.

And if it's GMail vs Hotmail, Microsoft has another problem to overcome. By initially limiting supply and by stimulating word-of-mouth, Google have made gmail addresses cool. Hotmail addresses are not, and never were, cool. Anyone could have one; and worse, any spammer could easily get one. For anyone who's been around a while, mail in the inbox from a hotmail address has one strong immediate smell: spammy. This is a hard reputation to undo.

All the more odd, then, that Google are misstepping so badly with Blogger. Blogger's been essentially static since Google acquired it. And it's increasingly gathering a bad reputation on two fronts. Firstly, for flakiness: lots of anecdotal evidence, to which I can add my own share of timeouts and database errors on post submissions, and recently a Wired News piece skewering them on reliability.

And secondly, as a haven for spammers. This Google search gives a good feel for it. Scott at Feedster nails them. And anyone who's read Usenet recently will recognise the Greatest News Ever spams, typically from Yahoo accounts, pointing back to throwaway Blogspot pages. (This last one also seems to be spilling over into web forums; I'm surprised it hasn't started hitting blog comments also.)

Reputations like this are easy to acquire, hard to shake; Blogger's in danger of becoming marginalised unless it does something about its problems.

[Update: More Talking about Maps and Mail over at Scobelizer.]


Actually, scaling IS a problem on Gmail. Ask Chris Uhlik (I did) about why they limit the number of users on Gmail. Hint: it's cause his system doesn't scale.

Hotmail already has 150 million ACTIVE users. Gmail isn't even in the same ballpark.

Hotmail's problem is gonna be how to scale up its hard drive space and how to improve the user experience so that users don't feel compelled to move to Yahoo or Google.

Also, why is this a race? Because Google messed with the business model and made it all about advertising. Well, sorry, advertisers go to where there are more users. That's why network TV keeps ratings.
Hi Robert — looks like I'm not the only one up late tonight.

That's not how I read the report of your conversation with Chris Uhlik (scroll down to the FooCamp entry; Scobelizer permalinks don't seem to link to the entry within the archive page):

I asked him why they only give out a few Gmail accounts at a time. [...] Turns out they don't have enough server capacity to deal with everyone who wants an account. So, when they buy a new server and get it installed they hand out another round of invites.

Sounds to me like it scales just fine: more users, more servers. The limitation is that they can't, or don't want to, install servers fast enough to satisfy all the immediate demand at once.

The invite rate, and so presumably the growth rate, has grown considerably since that conversation: issuing invites to users in blocks of 50, and inviting random visitors.

To me, this says "GMail is controlling its growth", not "GMail can't grow".

And as for messing with the business model: when was webmail not about advertising? I'm sure Hotmail was displaying banner ads to users long before GMail appeared on the scene...
Scobelizer permalinks don't seem to link to the entry within the archive

Oh yes they do: unless, like me, you use IE6 and shift-click the link to open it in a new window. Very odd.
A well-argued post. Bravo.

And c'mon, Blogger!
James - You forgot to mention Yahoo's search page: You'll notice that it is a clean, simple interface just like Google and MSN Search. They've also added cool features like having little news and mail snippets that you can enable/disable.

By the way, on the relability department, GMail has had its share of problems too (there were days when GMail users couldn't send/recieve messages and some messages were lost).
Forgot? I didn't even know it existed; I'd always visited to search Yahoo.

Nice. Very Googly.
One nice thing about these blog debates is that I almost always end up with another Blog to add to my permanent blog collection (added!). I've been using Hotmail since the beginning and I was part of the first wave of GMail beta-testers. Thoughts? No comparison. The SPAM issue is dead on, and all my SPAM comes from Hotmail accounts. I have yet to see a single one on GMail end up in my Inbox or even Bulk. Not one. And the curious thing is that I have been publishing my GMail address all over the place, just as I had *not* with my Hotmail account address. Only MS [it's my MS Partner email address] knew that one for sure; them and their partners.

As for scaling, and this goes to something that Scoble brought up, when you send or receive mail to/from multiple email accounts in the GMail system, I don't think it is keeping anything more than one copy. Remember, Google is a *search* engine company. Wouldn't it make more sense to keep one copy and issue pointers in the Inboxes of the receipiants? I don't know if they are doing this, but it is not beyond the reasonableness standard. So, if you do want to send that 200 MB video to 20 GMail members, don't be surprised if only one copy exists inside GMail. Not that I would even contemplate doing such a thing. It is far easier to load it up to my private FTP/Website folder and issue direct links, which my family and friends have been doing for quite a while now even from the warzones inside Iraq! This is our fourth or fifth war for the family and we do know tech (grin).