"Cloud computing". You probably couldn't invent a more confusing term. Let's break its meaning down from 3 perspectives.

1) Consumer / Business User:
The value of "the cloud" to consumers and business users is their ability to roam. Today you may be using Outlook on your PC. When you go on vacation you take your laptop with you because your software is installed on that computer. By moving your email to "the cloud", suddenly you're able to access your email from anywhere from any device. Your phone, an Internet cafe... anywhere connected to the Internet.

Do you have a huge collection of MP3s on your personal computer? Then you're probably aware of the overhead to maintain such a collection and the limitations that personal storage imposes (not to mention copyright concerns... shame on you ). Moving to a cloud-based music streaming service lets someone else manage your libraries and enables you to "roam" without constraint.

Pretty much any business software need can be addressed by "the cloud". Just take a look around your office and gradually starting replacing the ball and chain of on-premise apps with Internet based solutions. CRM, email, financials, backups, customer service and support, collaboration software.

Your "Kool-Aid" alert should probably be going off by now. No, not "everything" can be moved into the cloud. But portions of software can be moved to the cloud to give you a roaming experience. Photoshop designers will not find an equivalent Internet based application, but you can host your PSD and graphic assets files on a "cloud storage" service to open up collaboration, provide peace of mind by having off-site backups, and facilitating access to those files from your work/home/netbook computers.

Does "the cloud" seem less confusing now? Sure it does. It's a no brainer. It just makes sense that this trend will continue. Next...

2) Financial Users:
If you bought an ERP/CRM solution in the 90's, it was likely capitalized as an asset to the company and depreciated over time. But "cloud" software is just another operating expense, like electricity or water. You control the throttle of how much of the technology to use and pay a monthly or annual subscription.

Is "the cloud" even less confusing now? Sure. Who wants to own and maintain business software as an asset? Nobody. Next...

3) Technologists:
OK, so here's where the war and source of confusion begins. Hold on to your seat. Older software companies are now recasting themselves in "the cloud". Microsoft is thinking beyond the clouds towards "Azure" blue skies. Software companies that were previously "Software-as-a-Service" are now "Leading Cloud Providers". Amazon (yeah, the folks that sell books online) have an "Elastic Cloud Computing" service. IBM, who is *way* late to the game is suddenly talking about "private clouds" and moving your business into "the cloud". There are multi-tenant, isolated tenancy, and virtual machine architectures.

More money is now being spent on marketing "the cloud" than actual R&D for building "the cloud". The positioning for "what is the cloud?" is reaching a climax as vendors seek to be associated with this new buzzword.

The bottom line? As long as whatever is being called a "cloud" enables consumers and business users to roam and the service is just an expense, then you shouldn't care about "how" a particular cloud is implemented.

April 1, 2009

Cubic Compass today announced today the release of a new ground breaking technology for creating and managing interactive websites. Dubbed "Neural-Interactive Content Creation", or NIC², this patent-pending technology allows content creators to interact with an i-Dialogue Content Management System (CMS) via a neurological interface that instantly converts thought patterns into web content, dialogue script, and workflow rules.

"The pace of change in today's environment is just too fast for simple tactile interaction with a web CMS", said Cubic Compass Founder Mike Leach. "Today's websites and portals must respond at the speed of thought".

Beta test user Samuel Anders (pictured below) successfully utilized NIC² technology to provision, develop, and launch a fully functional website and customer portal in 17 seconds. "I consider myself a 'power user' of technology, so I was eager to plug-in to NIC² and translate my ideas into a working solution. The results were simply amazing."

The following tasks are just a small example of what can be accomplished in milliseconds using NIC²:

• Provision new landing pages
• Change passwords
• Grant permissions to portal users
• Publish a press release
• Add/modify workflow rules
• Update style sheet
  

NIC² is licensed per user and requires a lengthy pre-qualification questionnaire and note from a Doctor before use.

(Actual results may not be identical to those experienced by Mr Anders. NIC² is not FDA approved and is currently not legal in many countries. Please contact sales for more information)

Microsoft's Gen4 data center architecture involves trucks pulling up to an open air park and plugging right in to the Internet. No roof. Looks a little like an electrical utility transformer station.

Check out the video.

<a href="http://video.msn.com/?mkt=en-US&playlist=videoByUuids:uuids:b4d189d3-19bd-42b3-85d7-6ca46d97fe40&from=msnvideo" target="_new" title="Microsoft Generation 4.0 Data Center Vision">Video: Microsoft Generation 4.0 Data Center Vision</a>

Update: Here are a couple good online resources for learning more about Azure
MSDN Forums for Azure
http://blog.smarx.com/ (this blog is actually built and hosted on Azure)

Thanks to Steve Marx, Microsoft Azure Program Manager, for pointing these out.

---

I recently received this invitation from Microsoft (emphasis mine):

"With the recent unveiling of Windows Azure, we’re very interested in your thoughts about cloud computing. Take this short survey and tell us what you think about cloud computing. Your opinion counts. "

However, about half way through the survey, they apparently decided my opinion doesn't count (see screen shot below).

Ouch. Talk about a negative customer experience! So exactly how does one participate in this dialogue? Should I just engage with Ray Ozzie directly via email?

(With respect to Aerosmith)

Amazon just announced CloudFront; an Akamai-like service for hosting web resources closer to end-consumers of websites and portals.

How can CloudFront improve my i-Dialogue hosted solution?
i-Dialogue solutions are hosted in Plano, Texas and the Pacific NorthWest. When a visitor to your site in Europe requests a web page or document, that file must travel half-way around the world to the requesters browser (where it is typically cached for future requests).

CloudFront allows commonly accessed files, such as images, cascading style sheets, and javascript libraries, to be hosted and cached at various locations around the world to minimize the distance travelled to serve these resources.

How do I get started?
This service is currently in beta. We will initially pass through Amazon's CloudFront subscription costs directly to subscribers (setup and configuration fees may apply). Contact info@cubiccompass.com to learn more and get started.

How does CloudFront impact the content publishing process?
We are still researching how CloudFront impacts a typical content publishing workflow and what extra steps will be required by content publishers.

The long term goal is to manage CloudFront as a black box and any files/images copied to a specified folder on the i-Dialogue CMS are immediately replicated to CloudFront.

Dynamic pages will still be hosted on i-Dialogue and real-time data will still be retrieved from their content source (such as Salesforce, Google, or Microsoft services).

Where are the edge network locations?
Here's the info from Amazon's website on available network locations:

The Amazon CloudFront Network

To deliver content to end users with lower latency, Amazon CloudFront uses a network of edge locations world-wide. Amazon CloudFront uses the following edge locations:

United States

• Ashburn, VA 
  
• Dallas/Fort Worth, TX
  
• Los Angeles, CA
  
• Miami, FL
  
• Newark, NJ
  
• Palo Alto, CA
  
• Seattle, WA
  
• St. Louis, MO
  

Europe

• Amsterdam
  
• Dublin
  
• Frankfurt
  
• London
  

Asia

• Hong Kong
  
• Tokyo

I applaud Microsoft's foray into cloud computing with the Azure server, but it still leaves me puzzled. Having been a part of the Microsoft culture for so many years, I can see what is going on. There is this perpetual nagging drive to deliver more features and create new waves, when what is needed now is to leverage existing assets and ride the waves of consumer demand.

Microsoft is sitting atop a cloud computing gold mine today with the ultimate software-plus-service "Windows Update".

The old-school mentality is to deliver shrink wrapped software as a one-time transaction for a low price. They then give away monthly security updates, ongoing patches, and lot's of free development tools.

Because of the rate of change in todays networked world, there is now more value in subscribing to an evolving software service than in a one-time "buy and forget" transaction.

Most households would probably think nothing of paying $4.95 per month for Windows update if there were no upfront costs. That may seem like small change, but over 5 years that subscription model yields almost $300.

Plus consider how many bootlegged copies of Windows are running out there (50%+). By minimizing entry costs, demonstrating continual value, and associating key features with genuine installations will encourage more consumers to subscribe.

Businesses, likewise, would find a $10 subscription for Windows and Office much more attractive than a one-time purchase.

But here's the thing that really has got me puzzled. Why is Microsoft building it's own datacenters and talking about becoming a centralized service when it has built an army of loyal and eager partners that will install, deploy, and manage Internet nodes for them?

Google is a fantastic company, but their ability to scale is limited by their ability to raise capital and dedicate internal resources to building datacenters.

If all Microsoft Internet Node Servers (MINS) shared a common API, if all applications written for MINS were portable, and if all existing Microsoft datacenters and partners filled their racks with these servers, Microsoft would not have to spend a dime building datacenters. They could then return to their core competency of writing OS software and applications and leverage the power of their channel partners to scale.

(Alright, so maybe Microsoft Marketing could come up with a better acronym than MINS... but letting partners be their cloud computing 'minions' should be implicit ).

Revenue from Microsoft Internet Node servers would far surpass consumer windows subscriptions within 10 years if this model were deployed.

As an example, I just signed up for Amazon EC2 today to deploy a new email server. I don't know how much of the $0.125 per hour Amazon charges me gets kicked back to Microsoft, but it makes me wonder.... why is Amazon having to develop new Windows OS extensions to make their provisioning model work and why is Microsoft building datacenters like they're going to compete with Amazon?

It's just all backwards. The obvious win-win scenario is for Microsoft to develop a provisionable server OS and partner with the Amazon's and Rackspace's of the world to do the hosting.

 

Python is here! A couple months ago I mentioned the need for more dynamic language support in SaaS solutions. Last month we announced support for JScript and this month I'm happy to announce support for IronPython in i-Dialogue.

I spent the weekend learning some Python basics and created a simple (and I mean simple) Python demonstration in the i-Dialogue Developers Sandbox. You're welcome to modify this source and experiment on your own.

After years of working with the Salesforce API using statically typed languages; such as C#, Java, and Apex; it is refreshing to re-approach old problems through new eyes. After only a few days of playing with Python, I see now what others have been raving about.

Web developers will really enjoy using Python with Dialogue Script:

• Cleaner code. Easier to read and manage
• No need for thick IDEs. Too often, an IDE like Eclipse or Visual Studio stands between you and the desired solution. Dialogue Script development is 100% browser-based (yes, it even works in Chrome!)
• Agile business rules management. Work side-by-side with business users and apply business rules directly in web pages
• Dynamic typing

Microsoft is embracing Apple's "I'm a Mac/I'm a PC" campaign and just going with the "PC" stereotype. The message? "It's cool to not be cool" (hey... I've been saying that for years!  )

The guy in the opening scene is an actual Microsoft engineer that resembles John Hodgman from the Mac commercials.

Let's face it. Any good Web 2.0 strategy (or web 3.0) needs to have a "dynamic language" story to attract serious web developers. Languages such as Ruby, Python, PHP, and Perl are seeing tremendous gains in adoption by Web Developers.

No two organizations are alike when it comes to customizing their online presence and Internet Marketing campaigns, and these dynamically typed languages are increasingly becoming the language of choice as API extensions (see Google App Engine's use of Python).

Now that Dialogue Script is generally available and actively used in production, we've started turning our attention towards how to provide Developers with more programmatic control over the display and processing of web forms and portal applications, similar to how Salesforce employs Apex Controllers in Visual Force.

Our open source C# and ASP.NET API is an extremely powerful option for those familiar with Visual Studio.NET and managing strongly typed, compiled languages. But we wanted to evolve our platform and embrace the latest trend in dynamically typed languages and to go one step further by keeping the entire web development experience service-based (ie through a web browser or rich client).

Fortunately, the Dynamic Language Runtime (DLR) will be made available to us very soon and i-Dialogue Developers will have their choice of several dynamic languages to choose from when embedding rich programming logic into their web forms.

The DLR will give Developers the "glue" necessary to mash-up Google, Salesforce, Microsoft Live, Fedex, and any other web service using a familiar programming environment. The feedback from making changes will be instantaneous (no recompiling, moving files, or unit tests), making programming an instantly gratifying experience.

The DLR extensions to Dialogue Script will drive the innovation of new development process lifecycles and quality control processes that enable globally distributed teams to iteratively and incrementally evolve complex websites, portals, and campaigns using nothing more than a $599 laptop and browser.

The entire i-Dialogue object model will be made available, so DLR scripts can programmatically control all facets of a well rounded Internet Marketing Suite, including Profiles, Pages, Email, and Web Forms.

If you've ever written a VB macro to customize an Excel spreadsheet or Word, then you'll feel right at home with this new extension. This is probably #3 in the priority queue right now, with a couple tremendous new features taking priority right now for delivery by Dreamforce, so I'll keep you posted on it's evolution.

Suffice to say, we'll seek to leverage the DLR out of the box as much as possible with very few proprietary additions, so any O'Reilly book or MSDN article on the topic will be 99% applicable, if you want to get a head start.

Here it is July and I feel like a kid on Christmas knowing that the next release of SmarterStats includes advanced charts and graphs using Silverlight. Very cool!

SmarterStats is one of many applications included with every i-Dialogue subscription.

Just when I thought Google Apps subscription price of $50 per user/per year was the ultimate in web productivity tool value, Microsoft comes along and announces $3 per user/per month for hosted email/collaboration tools and simultaneously rolls over their existing hosting partners.

It's been over a year since I took a hard look at offering a hosted i-Dialogue solution for Microsoft CRM. At the time, the Microsoft CRM partner subscription model really wasn't in place (to be fair, the Salesforce Royalty Winter gave partners shivers too ).

But a lot has changed in a year:

• Microsoft launched CRM Online
• Microsoft reached out and gave us a great deal on a MSDN subscription
• Microsoft launched the Service Providers License Program
• The CRM 4.0 API offers a dramatic improvement over previous releases

I continue to be astonished at how many of our Salesforce customers use Microsoft Outlook, and I really see no end in sight for the Exchange beast, so it's worth conceding on this front that any truly effective eMarketing or Sales "dialogue" must integrate well with Outlook... somehow.

More on this later...

Those following my recent video blog entries have probably noticed a pretty concerted effort on our part to improve the user experience (UX) of our CMS.

You can pretty much single-click your way through the most common functions today as we've converted to a 3 tier UI that progressively reveals advanced features and functionality as you descend into the dialog windows.

It's amazing how many experiences we tolerate in Windows on a day to day basis. This Windows UX Ideas site has started a constructive dialogue on the topic.

(No, that's not a Salesforce Ideas site. Even though it looks identical, the author custom developed the site in PHP... very nice)

Just noticed that Microsoft CRM Online is now officially available. Not sure how personally invested I'll get in learning it. I responded to several Beta trial invitations, but never received an invite (and we're a Microsoft partner?!?).

There is a 30 day trial that you can apply for.

Kudos to Microsoft for changing the name from "Live" CRM to simply "CRM Online" (As Simon Cowell might say "The whole Live brand just seems a bit too cabaret-ish for business software applications" ).

If The Big Switch can be said to represent the centralization of computing resources, then Microsoft's Live Mesh may signal the path towards the big switch back to a decentralized model (albeit with some centrally hosted Microsoft infrastructure).

Some ideas for using Mesh with our CEM platform and Salesforce:

• Real-time notifications when Leads/Contacts enter your website
• Sync and offline access to CRM data and documents across several devices
• Rich development of email campaigns and web page content
• Social networking with employees, partners, and customers

 

Hands on with Live Mesh

In classic Sun Tzu fashion, Marc Benioff embraces competitor SAP by stating:

"I want to figure out how to get SAP to build on our platform. SAP needs to write its new apps on our platform."

For Microsoft's part, SAP integration came naturally since SAP is used in-house to run the business. The two giants have silently agreed on the mutual win-win opportunity of using Office tools (Excel and Word) to access SAP apps.

Unfortunately, integrated Microsoft-SAP applications have an air of being written for Microsoft by Microsoft.

Is it likely that SAP will develop their new apps on Force.com? Probably not.

Is it likely that SAP customers will (and already do) develop Force.com apps integrated with SAP? Definitely, yes.

In contrast to the "destroy the enemy" strategy employed against Seibel, I think it is a wise path for Salesforce to acknowledge that on-premise enterprise software will be around for years to come and that the "art" of victory against established ERP players will be one that embraces the revenue streams of existing infrastructure and provides a migration path to SaaS/PaaS alternatives.

Nicholas Carr advises his readers to put their ears to the ground as Microsoft prepares to announce further details about their "computing in the cloud" strategy.

Some facts and observations I've made:

• Ray Ozzie, Bill Gates heir apparent, is a smart guy and has been under the radar for over 2 years working on something (presumably).
• Microsoft has been building, acquiring, and deploying datacenters at an alarming rate. But the utilization math doesn't quite add up. There's a lot of untapped, unused computing power owned by Microsoft.
• In my recent discussions with various Microsoft employees, there is a much more open and humble acknowledgment that they are in the canonical Innovators Dilemma and must branch out into hosted services while still retaining the Windows OS/Office cash cow.

It will be interesting to see how similar this new strategy is to Microsoft's Hailstorm platform, which was announced 7 years ago. Ironically enough, the Architect of Hailstorm, Mark Lukovsky, left Microsoft and joined Google to ultimately implement and fulfill this vision.

Microsoft clearly has the Engineering resources and talent to compete with Google, Yahoo!, and Salesforce.com in Internet delivered services. It makes one wonder what exactly is preventing them from tapping their true potential? 

Per Nicholas' advice, my ear is firmly planted to the ground...

On-Demand software has revolutionized how enterprise software is deployed. Declarative customization provides line of business managers with an unprecedented number of options from which to "declare" how a CRM record is defined or presented in a page layout. But let's face it... at some point, all enterprise software solutions require some level of programmatic customization. Enterprise software vendors simply cannot anticipate each and every business requirement.

Salesforce customers, having reached the limits of declarative customization, now have domain specific programming languages such as Visual Force and Apex to programmatically customize their CRM system.

Content Management Systems (CMS) and Enterprise Portals already have a strong legacy of supporting programmatic customization, but achieved this flexibility through integration with existing development environments and the modification of physical files and templates.

Many CMS's have attempted to offer purely browser-based declarative customization environments, such as Wiki's and Blogs, but failed to meet the demands of creative and marketing professionals, which require detailed control over the presentation and branding of web pages and emails.

It's time for a new scripting language that meets the unique demands of Customer Experience Management (CEM). It's time to offer an optimal balance between declarative and programmatic configuration options. It's time to recognize that there is a new generation of workers entering the workforce that embrace Domain Specific Languages (see Fowler: DSL) and expect programming the web to be as easy as programming their iPod, Tivo, or DVR. It's time for a language like "Dialogue".

"Dialogue" is a new scripting language designed specifically for helping organizations to interact with their customers and partners online using concepts familiar to both the web (HTML, CSS, Javascript) and CRM (Lead, Contact, Opportunity, Case records).

The key tenants of Dialogue include:

Creativity / User Interface: Dialogue is primarily a presentation language that resembles HTML and is used in the creation of dynamic web pages, landing pages, and emails. Graphic designers and web developers have control over each pixel in the presentation of Dialogue scripts.

Business User Support: Business users can easily modify Dialogue scripts to present information most relevant to customers and partners. Changes can be made to Dialogue scripts without the assistance of a web programmer.

Multiple Development Environment Support: Dialogue scripts can be managed using either a web browser or rich client. Todays content management professionals are accustomed to working with tools such as Dreamweaver and Expression. Dialogue plug-Ins for web development environments (starting with Expression) will be developed to enable the remote persistence and management of Dialogue scripts using a locally installed rich client. But the flexibility of storing and managing content entirely on the web will remain the primary focus.

CRM-Driven: Dialogue scripts provide direct access to CRM records and data for maximum personalization.

Agile: Dialogue scripts support an agile methodology that allows organizations to go from concept to deployment within a matter of hours or days. Campaigns can be modified on the fly and no longer suffer from the constraints of "waterfall" project management methodologies where considerable upfront planning is required to reduce the anticipated cost of change.

The Dialogue scripting language will help organizations transition from static, brochureware web designs common in the 90's, to highly interactive and personalized online experiences. It is our goal that all Salesforce customers will unlock the capabilities of their CRM system to delivering these rich online experiences. Dialogue is the final link in enabling this level of interaction.

Our portal and eMarketing Developers spend a considerable amount of time translating customer requirements into working features, which requires the use of highly abstract tools, languages, and technologies. To this end, we've standardized on the use of Microsoft tools and services, commonly referred to as ".NET".

We were fortunate enough to spend the afternoon with Scott Hanselman, Development Community Liaison for Microsoft. Scott is fairly well known in the .NET development community. In fact, if you've ever Googled any search term remotely related to .NET, chances are you've most likely come across his blog or PodCasts (prompting some to wonder if he ever sleeps ).

I was really impressed with what Scott presented. Not only with the wealth of new features and services in .NET 3.0 and 3.5, but the myriad 3rd party and open source tools available that can be uniquely combined to create an extremely powerful, test-driven development environment.

The development tools space has become very stable and commoditized. Much like the component-based Eclipse development environment used by Salesforce.com, Visual Studio.NET is so extensible and customizable that the only limits are the Developers imagination. Scott maintains a great list of indispensable tools and utilities.

The highlight of the session? Probably LOL CODE. An actual programming language based on the LOL CATS concept. Who knows... maybe next generation CRM applications will be programmed in languages that look like this?

HAI
CAN HAS STDIO?
I HAS A VAR
GIMMEH VAR
VISIBLE "You said " N VAR N " !!"
KTHXBYE

News is out that Microsoft is investing $240 Million in Facebook for 1.6% in equity, putting a $15B valuation on Facebook. What?! You read correctly. A 1.6% stake for $240M.

Hard to justify by any means, but it reminds me of when the Yankees signed Alex Rodriguez to a 10 year $252M contract. I was among the many Seattle Mariners fans who paid top dollar to see this new Yankee's star play, which in itself probably justified the valuation because the Yankees were selling out just about every game in every town.

A self-fulfilling prophecy perhaps? I would expect Facebook's 50M user community to easily double within the next year as people tune in to see what the commotion is all about.

The Microsoft CRM Live buzz appears to have died down. We started looking at the opportunity to implement an "i-Dialogue for CRM Live" solution back in March, but nothing ever materialized that we could get our hands on.

This delay violates the on-demand credo of "Release early and often". I had really hoped Microsoft would earn their operational stripes by jumping into the market with "good enough" and letting customer and partner feedback dictate the evolution of CRM Live. Now we're left to assume that the project is in a state of "analysis paralysis" with Program Managers bickering with Engineers over how to "get it perfect" for launch (it'll never be "perfect" guys... don't take it personally. It was ready last year).

Dynamics CRM 3.0 is a great CRM solution (see my test drive last year), but I'm beginning to have serious doubts about Microsoft's ability to enter the market behind Salesforce.com. This is not like throwing an XBox up against a PlayStation. There's no Halo 3 equivalent killer app that can make CRM Live an overnight success (or within 5 years) that I can see. But there is one killer infrastructure angle up Microsoft's sleeve that could change how CRM is deployed and hosted in the future.

If Microsoft truly wants to be a CRM contender then, in my opinion, they need to leap frog the whole multi-tenant architecture discussion and focus exclusively on next generation CRM virtualization, which is the ability to pre-configure an entire Windows CRM server with all dependencies, and allow this "image" to be hosted by any 3rd party.

Leverage would then fall back to the consumer who could "shop around" for virtualization hosts and take their entire CRM/Active Directory/SQL Server VM image with them anywhere they want to go. I suppose that would give new meaning to "Where do you want to go today?"

Microsoft's announcement that MS CRM Live will be priced at $45 per user has raised some responses from the competition.

You can buy the on-premise version of MS CRM today for about $1,500, which includes a perpetual use license. Spread that cost out over 3 years and you get $41.66 per month. Add a few bucks for hosting and $45 sounds about right.

I have a feeling that there will be more parity between an organization's size and the number of seats licensed at this price point.

For example, an organization with 20 employees may only subscribe to 5 Salesforce seats at $75 per month and limit access to those in Sales or Support. Whereas the same organization may be more inclined to purchase 10-15 seats at $45 per month.

While Salesforce clearly has more cross functional value than MS CRM, customers often do not want to pay for unused functionality in the early stages of adopting CRM when centralized contact management is the primary goal.

Wow. I anticipated some level of technical difficulty when integrating with Microsoft CRM (support for which we recently announced) but didn't anticipate any significant legal issues. Apparently, Microsoft CRM 3.0 requires the licensing of something called an External Connector for each web server integrated with Microsoft CRM and I'm not the first to be confused about the applicability of this license in a Web-to-Lead, customer portal, or partner portal context.

The connector is about $15,000 USD, and amortized over 5 years comes to $250 per month. I *really* hope Microsoft CRM Live modifies this license to be more inline with current subscription-based expectations. Just getting SMB customers to try on-demand, loss leading features for $95 per month can be difficult enough. Don't make partners start that far in the red. They'll never adopt the platform.

Will the partner/ISV absorb the cost of externally integrating with MS CRM Live? Will the end customer? Answers to these questions are still not clear as we enter the 11th hour prior to Microsoft launching their on-demand CRM Live solution.

My suggestions:
* Implement a variable, value-based external integration licensing model. Not a "one size fits all" license.
* It still remains to be seen if Salesforce.com's 10% referral program is the right model to attract and retain partners, but it is more closely tied to value on a per customer basis.
* License the XML web service gateway separately from the core CRM application per external end-point. Implement a SOA friendly model that encourages Microsoft CRM customers to purchase additional web service connectors for multiple end-points at $750 per year (This is where I think Salesforce has missed a tremendous opportunity to license their API to Professional Edition customers and is the most griped about integration topic).