Marissa Mayer, head of Google Local gave a generally neglected overview of Google Maps at SXSW.  She reported some new stats (150M users of mobile app, 2B driving miles on Google Navigation), but this was mostly lost in the chatter about new group chat apps or where to get good breakfast tacos.

Google maps has been raising the bar so often that their latest innovations rarely make big news. That’s a shame. Since 2004, Google has consistently, relentlessly raised the bar on what a map platform can and should do. Starting with JavaScript “slippy maps” to one-line search to satellite imagery to draggable routes to StreetView to traffic conditions to building their own map data to 3-D representations (and many more), Google has introduced (or more accurately popularized) map features that quickly become the baseline against which other map providers have to compete.

That’s a tall order.  Google has been at it since 2004.  And being “at it” has meant employing several hundreds of people all through that time. And what looks like an open checkbook, at least by most companies standards.

The conclusion may be obvious but, in my conversations, frequently overlooked: building a modern map platform is really, really hard. There is open-source software that can get a small team to a basic tile server pretty fast, but to compete in the big leagues takes several hundred man-years of effort by very smart people. That’s been true for Google, for Nokia, for Microsoft. Many others who have tried to get there cheaply have failed, or at best, are yet to succeed.

So why this statement of what might seem obvious?  Because the conventional wisdom among many  is that maps are commoditized. Commodities are undifferentiated goods and services that are widely available from multiple sources with little or no switching costs. Maps are commoditized only in the narrow sense that they are freely available and easy to add to your application through API’s.

Available from multiple sources?  Not really. There are only a handful of mapping platforms that give the level of functionality now expected (thanks to our friends in Mountain View).

Low switching costs? Yeah, if you’re a developer, switching out map platforms is some adjustment of code.  A good weekend’s work. But the companies building maps aren’t investing in all that development for nothing. Maps are a key component of their plans to own location context and by extension, ownership of the core elements of the customer value proposition. Said plainly, they give you the maps for free because they want to drive the business model. And if they drive the business model, guess who doesn’t?

The map isn’t dead as has been widely theorized.  Map-less places databases are enjoying popularity but mostly in applications for high density search targets (restaurants in downtown Manhattan). As the density of search targets decreases, maps become more important ways to visualize place.

The strategic question for big players who don’t want to rely on a likely competitor for maps is what to do.  It is not an easy task to catch up, and Google shows no signs of slackening the pace.