Friday, March 21, 2008

xTuple: a hybrid open-source ERP development model

Open source ERP products have made good progress in the past few years in gaining customers. But their challenge has always been to leverage their "developer community" to fill out the functionality of their products.

Developer community is key
There's a little bit of a chicken-and-egg problem here. To encourage contributor code submissions, the owner's development organization needs to have adequate resources to foster relationships with the community as well as to evaluate and incorporate the contributions. But to afford those resources, the product needs a large enough installed base to support the contributors.

Large open source infrastructure software projects such as Linux and Apache don't have this problem. Those products are used to support all sorts of applications, giving them tremendous scale and major service providers take it upon themselves to hire and subsidize the efforts of the project's key leaders. The scale of these projects also leads to a large pool of contributors eager to see their improvements adopted into the core product.

Open source ERP products, on the other hand, are much more narrow in their market focus and therefore have a more limited pool of contributors available. Which is why we do not see an open source ERP project of the scale of Linux, Apache, or MySQL.

A hybrid model
I've been aware of this problem for some time. So, I took note of an announcement from xTuple (formerly OpenMFG) regarding 13 major enhancements contributed by customers over the past few years (which they dub "greatest hits").

A quick read through of the enhancements shows they are not small changes. They include an EDI engine, a complete project management subsystem, multi-currency functionality, an MPS module, and a shop floor subsystem that employs lean manufacturing and theory of constraints concepts. One contribution is from a consortium of five customers who put together a CRM module.

To understand how xTuple has gotten this level of cooperation from its customers, I called Ned Lilly, the firm's CEO.

According to Lilly, xTuple is an open source hybrid. It develops its product in three versions: a completely free and open source Postbooks edition, which it positions as suitable for small businesses; a Standard edition, which is its intermediate offering; and its OpenMFG edition, which is the premium offering.

The Postbooks version is freely available for download from xTuple, at no charge. The latter two versions are sold under a traditional license, through authorized partners, who pay an annual fee to xTuple. Source code is available to customers, and the partners are encouraged to develop enhancements for specific customer requirements. Under their agreement, the partners are then obligated to return those enhancements to xTuple, which evaluates them for applicability to the core product.

Customer contributions are included in all three editions, even the open source Postbooks version, but the really sophisticated stuff, of course, is saved for the higher end versions. According to Lilly, about one third of the source code across the three editions was contributed directly by customers, and another 60% or so was developed by xTuple under customer sponsorship. The CRM module mentioned earlier was a customer-sponsored development.

Open source purists might balk at this hybrid model--why isn't it all free, why do partners have to pay to be partners, etc.? Nevertheless, it's hard to argue with success. Although all ERP vendors claim that customers drive new features and enhancements, few can actually point to code that their customers have directly contributed. xTuple is making good progress in the key success factor of open source projects: getting users to actively contribute to development.

If we have to compromise a little with the open source model to make that happen, so what?

Related posts
OpenMFG is now xTuple
Open source ERP gaining adherents
The disruptive power of open source
Total cost study for an open source ERP project
ERP Graveyard

Monday, March 17, 2008

Cloud computing: can Microsoft turn from servers to services?

As everyone knows, Microsoft rose to its dominant position as a software provider. Today, however, it faces threats of disruption from players such as Google, which delivers software functionality as a service, over the Internet.

The new approach is dubbed "cloud computing," because the software that provides the service does not reside on a specifically-identified computer but is deployed over the Internet (which is often depicted graphically as a cloud) from any number of networked computers. The user has no idea of where the host computer is or what operating system it is running. The application is just out there "in the cloud," as so to speak.

A threat to Microsoft
Market leaders spend a lot of time thinking about how they can lose their market position. As Andy Grove, founder of Intel said, "Only the paranoid survive." The trend toward cloud computing, or software as a service, certainly is a threat to a company that depends on organizations buying servers and installing software. So Microsoft's current strategy, at least the one it offers publicly, is "software plus services." That is, it continues to sell software, but it is also offering hosted versions of its popular Exchange messaging system and Sharepoint collaboration system.

Which brings us to Phil Wainewright's view of Microsoft's strategy: "bunkum." In a Feb. 1 post, he wrote:
The attempt to define software as somehow separate and parallel to services is what exposes Microsoft’s software-plus-services notion as bunkum. Software is the stuff that powers services. Before the Web existed, the easiest way to get those services was to install a software package. Now that the Web exists, the easiest way to get those services is to click a link and let someone else run the software. That’s what Google does. That’s what Yahoo! has always done. These companies aren't doing away with software; they’re huge consumers of it. When talks about ‘No Software’ what it really means is ‘Lots more software, but we’ll take care of it for you.’ The software becomes part of the service provider’s infrastructure.
Now, in a new post, Wainewright says that Microsoft may already be realizing that software-plus-services will not be enough to counter the threat of cloud computing. He points to recent comments by Microsoft's chief strategy officer Ray Ozzie, which suggest that, "despite the public bluster, Microsoft’s top brass already secretly realize that they must put services, not software, at the center of their world view." Read Wainewright's whole post for Ozzie's comments that reflect his view of Microsoft's need to make a more complete turn to services.

Dynamics a laggard in cloud computing
Where does this leave Microsoft in terms of its enterprise applications, Dynamics? Microsoft is way behind the curve in deploying these products in the cloud. It is just now getting ready to release CRM Live, the software-as-a-service (SaaS) version of its Dynamics CRM product. As far as its other Dynamics products, such as Axapta (AX) and Great Plains (GP), the only SaaS deployments are hosted versions offered by some Microsoft business partners.

The slow progress in moving its Dynamics applications to the cloud is not a big deal right now, as most small and mid-size businesses are still thinking in terms of on-premise deployments of software. (CRM is the one area where the SaaS model has taken hold, where its ubiquitous access characteristics are an advantage in serving sales users that have no fixed work location.) But as cloud computing becomes a more accepted way of delivering enterprise system functionality, providers that do not have a credible solution are going to be at a disadvantage.

Both Oracle and SAP are making major investments in the new model. SAP in particular is developing its new SMB offering, BusinessByDesign, from scratch to run in the cloud. As for Microsoft, it's no doubt difficult to embrace a business model that undermines the way it got to be an industry leader. But if it does not do so, it's likely to lose its position, not just in enterprise applications but the whole nine yards.

Update, Mar. 18. In the comments section, Michael Sheehan notes that cloud computing and SaaS are really not the same thing. I agree, and I was a bit sloppy in using the terms somewhat interchangeably. In my mind, SaaS is any software functionality delivered as a service, whether it be a dedicated system hosted by the service provider, a multi-tenant system providing services to multiple customers, or a system deployed in the cloud. I suppose in some organizations, the cloud could be used to deploy a system that is used by the organization itself. So, although cloud computing is often used as a way of deploying SaaS, they are not synonyms.

On his own blog, Sheehan points to this post by Paul Wallis that goes into much more detail on the evolution and relationship of supercomputing to grid computing to utility computing to cloud computing.

Related posts
Microsoft Dynamics revenue growth less than stellar
Microsoft Dynamics: management changes spell lack of direction

Friday, March 14, 2008

Microsoft Dynamics revenue growth less than stellar

Josh Greenbaum has put together some revenue numbers for Microsoft Business Solutions, the group that develops Microsoft's enterprise applications, and finds that the group has been growing at a rate less than half of that shown by SAP and Oracle.

The numbers are not easy to get at, since Microsoft stopped reporting revenue for the group in 2006. But Josh made note of a mention by Kirill Tatarinov, head of Dynamics, that revenue was now at $1 billion. Comparing that number to the $919 million reported in 2006, it appears that Dynamics has only been growing at an annual rate of 8.8%.

Josh writes:
But instead of something to crow about, $1 billion – a growth rate of 8.8 percent – represents a pretty lousy number for a group that was growing in double digits only a year earlier. It’s even worse when you consider that during roughly the same period SAP grew 18 percent and Oracle grew 30 percent if you include its acquisitions, and an estimated 8.8 percent if you look at pure organic, non-acquisition-based growth. And don’t forget, that 8.8 percent growth includes what everyone at Dynamics calls the hockey puck growth curve for CRM, which, I was told, has been growing at 100 percent per year for the last two years. Which means if you want to understand how non-CRM growth is at Dynamics, it’s safe to either knock a few points off the overall number (7 percent? 6 percent?) or discount the heavy growth in CRM as a largely revenue neutral.
As Josh points out, this less-than-stellar revenue picture could explain the revolving door of top management personnel at Microsoft Dynamics.

So, what to make of the slow growth of Microsoft's enterprise system products? I don't think the blame can be placed on the products themselves, which comprise the Axapta (AX), Great Plains (GP), Navision (NAV), and Solomon (SL) ERP systems. Microsoft obtained these products in its acquisition of Great Plains and Navision in 2001 and 2002 respectively. These were good products prior to Microsoft's acquisition, and Microsoft has continued to develop them. Customers I speak to are generally happy with them. (Microsoft developed its CRM product from scratch, subsequent to those acquisitions.)

I think the problem is that Microsoft is just getting outsold, especially by Oracle and SAP. Microsoft's sales model is 100% through business partners, a model that is more difficult to manage than that of Oracle and SAP, which rely primarily on their direct sales force and supplement it with resellers. Resellers tend to get busy in implementation with a few good accounts and are apt to slack off new sales efforts. Microsoft can't lean on resellers at the end of the quarter in the same way that Oracle, for example, can lean on its own direct sales force.

Furthermore, few resellers have the scale to compete effectively in larger deals. With the rise of globalization, many small and mid-size companies today are truly international businesses, with production and distribution operations throughout the world. A small reseller in Chicago, for example, cannot deploy resources as easily as SAP or Oracle in supporting a customer with operations in Europe and the Far East. If they try to compete, they simply get outsold by the big guys.

Microsoft is apparently having some success in signing up larger organizations as resellers. Witness the announcement this week that Microsoft is expanding its relationship with EDS to sell Dynamics CRM. Such relationships certainly address the scale issue, but I still question how much EDS is going to jump when Microsoft starts pushing at the end of the quarter.

Now we have Steve Ballmer claiming that Microsoft is now the "leading provider of enterprise software" in terms of "total dollar volume" -- a claim that is ludicrous on its face. Apparently, Ballmer is lumping in Microsoft's database (SQL Server) and its collaboration system (Sharepoint) with Dynamics -- clearly apples and oranges.

Microsoft has ambitions to compete in business applications with Oracle and SAP, but it is trying to do so with a sales model inherited with their acquisitions. The revenue picture reflects the reality that in some fashion Microsoft needs to think about having some role in direct sales for larger deals.

Update, 5:00 p.m. A source at a Microsoft reseller points out that Microsoft's direct enterprise sales people do in fact get recognized when a Dynamics sale goes through for one of their customers. However, the sale must be done through a reseller, and the Dynamics products are not included in Microsoft's volume license agreements. But since Microsoft's direct sales people are measured on enterprise license sales revenue, Dynamics is not top of mind for these folks. In other words, I don't believe they have a quota for Dynamics.

Related posts
Microsoft Dynamics: management changes spell lack of direction