A Semi-open Source Vendor Discusses Market Trends


TEC's continuing question and answer (Q&A) series, in which we solicit vendors' responses to our questions and observations on market trends (see previous articles in this series: Two Stalwart Vendors Discuss Market Trends and A Partner-friendly Platform Provider Discusses Market Trends), has become quite popular with readers and vendors alike.

Another market player that has voiced its opinions is Norfolk, Virginia, (US)-based xTuple (formerly OpenMFG). Privately held xTuple is a self-financed developer of enterprise-class business process applications powered by open source software and infrastructure such as Linux operating system (OS), PostgreSQL database, and Qt, a C++ graphical user interface (GUI) development framework from Trolltech, a Norwegian software company.

Before delving into xTuple's answers to our questions, some background on xTuple and OpenMFG is in order. The vendor is a relative newcomer in the enterprise resource planning (ERP) arena; it was founded in 2002 (as OpenMFG) by its current president and chief executive officer (CEO) Ned Lilly, a former executive at Landmark Communications.

The first we learned of OpenMFG was at an industry event in early 2003, where the company had a booth and was able to brag about only a few pilot ("beta," or test) customers. Our impression at that time, given the depressed economy worldwide and the demise of so many vendors, was that a brand new ERP provider was not badly needed in the market. However, xTuple has, to a degree, proved us wrong: its current roster of xTuple ERP commercial customers to date is about 100, and it has 20 partner resellers. The company's focus on the smaller enterprises, with up to $100 million (USD) in revenues, must have played a great part in its current success. The free version of xTuple ERP has also been downloaded over 150,000 times from SourceForge.net, the world's largest development and download repository of open source code and applications.

Expanding Product Offerings

Though it might not sound like marketing wizardry to some, the xTuple name was picked to denote the company's diversification in terms of product offering: OpenMFG, a manufacturing-oriented ERP product; OpenRPT, an open source report writer; and the most recent PostBooks open source accounting/ERP application. The name xTuple, therefore, speaks to the exponential growth possible with open source solutions.

The vendor continues to develop and market the xTuple ERP OpenMFG Edition, its commercially licensed solution for small to midsized manufacturers. The maturing manufacturing-focused ERP product will continue to be available under the hybrid community source code license that the company has employed for the past six years. In this arrangement, partners and customers get full source code, and any subsequent enhancements flow into the base product to which xTuple maintains and claims the intellectual property rights.

Having been referred to on occasion as a quasi–open source provider, xTuple points out it has never claimed that the OpenMFG Edition is fully open source (the vendor knows enough about that "clique-y" and somewhat snobbish world to be very careful with its choice of words and definitions). Yet the company believes that the hybrid approach has offered the best of both worlds to its users for the past five years in terms of a solid, professionally supported ERP solution built on a fully open source infrastructure, and licensed under a community source license through which community members can be actively involved in the product's ongoing development.

What's new is the PostBooks product: the company has carved off a new, entry-level, fully open source product that shares the same code base as the commercially licensed Editions. In fact, the client binaries are identical for both products, so an upgrade from PostBooks to the Standard or OpenMFG Editions involves running a simple database script. The only difference is that OpenMFG offers more advanced functionality in manufacturing and distribution, which non-manufacturing enterprises probably do not need. For more details about the commercially licensed xTuple ERP Standard Edition product (targeted at distributors and retail), and the free PostBooks Edition and the OpenMFG Edition, see http://www.xtuple.com/comparison for a chart of all three products.

To be more precise, OpenMFG-specific features cover approximately 20 percent of the highest-value functionality, such as multi-warehouse inventory, warehouse transfer orders, lot/serial control, manufacturing resource planning (MRP), master production schedule (MPS), bills of operations (BOOs)/routings, breeder bills of material (BOMs), item transformations, infinite capacity planning, lean/buffer management, returns/service, and batch manager/electronic data interchange (EDI).

PostBooks is available now as free and open source software (FOSS) on SourceForge.net, and elsewhere under common public attribution license (CPAL) open source license. Based on over 150,000 downloads to date, and the active community of users and developers at SourceForge.net and xTuple's own xtuple.org website, xTuple thinks the offering will be a big hit. Thus, the two communities will grow in tandem, while xTuple pledges to manage the ongoing development of both products, so that enhancements to one can flow into the other. As mentioned above, it is the same code base, after all; should a PostBooks user enterprise ever wish to upgrade to OpenMFG, it only has to run a short upgrade script on its database. Since the business logic for both applications resides in the PostgreSQL back end (in the procedural language), this is relatively easy to maintain.

PostBooks, xTuple believes, is one of the most advanced open source accounting/ERP solutions out there in the market. The recent 3.0 version of xTuple ERP featured what xTuple billed as "the world's only open source ERP product configurator" as part of the base PostBooks package. Of course, the OpenMFG and Standard Editions will also continue to feature added advanced functionality that will make it "worth the upgrade" from the free "sibling" (related) product, especially for small manufacturers that cannot afford the money and time requirements of traditional ERP deployments. As a company, xTuple pledges to support the entire technology stack its applications employ, notably the PostgreSQL database.

Appealing Pricing Transparency

Where xTuple is indisputably "open" (referring back to the question of OpenMFG's true open source nature) is in the vendor's transparency about its pricing and current product functionalities. For one, product pricing is available online. Despite new product additions, OpenMFG pricing remains unchanged, and the application continues to be offered as either 1) a subscription-style license (which includes software maintenance) at $1,000 (USD) per user, per year; or 2) a perpetual license at $3,000 (USD), plus 18 percent annual maintenance (for a minimum of five users). The pickup among the current customer base is roughly 50-50.

While PostBooks is free software, xTuple offers some commercial support options, such as an annual retainer at roughly half the cost of OpenMFG annual license and incident-based bundles of consulting hours. Following are some of the vendor's selected productized professional service offerings:

  • the QuickStart implementation package (10 days and a project plan)

  • three grades of networked server maintenance: basic, backup, and premium

  • PostgreSQL database support, tuning, replication—a number of tiered offerings with details available on the company site

Last but not least, there are some xTuple ERP server appliances, such as three configuration options based on the user count and level of support.

No-spin Product Functionality

As for the product's available out-of-the-box functionality, it is also a "what-you-see-is-what-you-get" approach. There are a dozen or so xTuple ERP modules with functionalities that flow in a cohesive and logical manner. Prospective users can "test drive" xTuple ERP functionality at http://www.xtuple.com/demo/video.

The company's employees are seasoned ERP executives, sales, or presales personnel from former or current competitors of xTuple. This has played a major role in designing xTuple's capabilities from scratch, as it was essential to provide only the necessary, nifty product capabilities first. It was also necessary to stay away from the "functional bloat" (extraneous functionality; see A New Platform to Battle Software Bloat?) which all too often becomes overkill for smaller enterprises.

To that end, currently available xTuple ERP functionality revolves around the following modules, capabilities, or processes:

  • Inventory management—enabling multi-warehouse and multilocation functionality, cycle counting, lot and serial tracking, and traceability, etc.

  • Product definition—enabling BOMs, routings, standard costing, actual costing, etc.

  • Manufacturing—enabling MRP, MPS, capacity requirements planning (CRP), shop floor control (SFC), labor entry, material variances, etc.

  • Purchasing and accounts payable (A/P)—enabling segmented requisitions and purchase orders, accounting controls, vendor performance reports, etc.

  • Sales, customer relationship management (CRM) and accounts receivable (A/R)—enabling 360-degree customer views, sales quotes, multidimensional pricing schedules, sales contacts, incidents, to-do lists, opportunities, etc.

  • Shipping and receiving—integrated with UPS, FedEx, and other leading shippers, and fully bar code–enabled

  • General ledger (G/L)—enabling detailed G/L and journals, bank reconciliation, budgeting, customizable report engine, etc.

  • Many system-wide utilities—fine-grained user privileges and security, batch manager, events engine, hotkeys, calendars, EDI, international locales and complex tax structures, multicurrency, project management, etc.

Community-led Product Development

Certainly, one wonders how a company of only about 15 employees could have developed so much functionality from the ground up. Well, if one counts the help from the entire development community (including customers and resellers), the answer to this becomes clear. For instance, the current xTuple director of product development, John Rogelstad, is an experienced ERP expert and consultant who implemented OpenMFG as a customer during his prior employment at Marena Group, and who has written major new areas of functionality.

One of these new functionalities is the recently released constraint management capability. This new functionality follows along the lines of the Theory of Constraints (TOC) and simplified Drum-Buffer-Rope (S-DBR) concepts (see The Theory of Constraints Enters the Lean Manufacturing Arena). The developed code was footnoted with references to the "fathers of TOC," Goldratt and Schragenheim. The solution looks deceptively simple, but many people actually take a while to "get" it (learn and understand) because it appears to be counterintuitive at first. Basically, the target customer for this solution is someone who has pulled out half his or her hair trying to program the shop with a complex finite scheduler, only to find that the system recommendations conflict with reality.

In addition to constraint management, new major enhancements in version 2.x were made to CRM, MPS and forecasting, and multicurrency. The CRM module introduced the concept of "accounts;" the idea is that users can have various types of relationships with companies. Conversely, in many other systems, users are prohibited from having a vendor that is also customer. Thus, in OpenMFG, the "C" in CRM really ought to stand for "corporate." An account can be a prospect (one that might become a customer), vendor, partner, competitor, or even a tax authority. Thus, users can have as many individual contacts and addresses assigned to the account as they like. So, the module is really a fully integrated contact manager and CRM system that is also a starting point for deeper ERP functionality in sales, purchasing, etc.

Version 3.x, new this year, adds a user-friendly screen builder and JavaScript support for designing customized system dashboards; major enhancements to returns processing; advanced warranty tracking; integration with external shopping carts such as Yahoo Stores; and the aforementioned assemble-to-order (ATO) product configurator for streamlining non-inventory order processing.

Diverse xTuple Customers …

As stated above, xTuple's install base is nearing 100. Target customer situations include both new, "green field" ERP implementations (with no ERP system ever having been on-site) for smaller companies, and replacement systems for unhappy customers stuck with a casualty of the ERP Graveyard. Based on the functionalities discussed above, the OpenMFG Edition seems a good fit for discrete manufacturers in the make-to-order (MTO), make-to-stock (MTS), and especially mixed-mode environments. There is also a fair amount of support for batch process manufacturing.

In fact, a wide variety of industries is represented in the customer base: industrial machinery (oil pipelines and valves, water purification, hydraulic and pneumatic tools, etc.); transportation manufacturers (engines, bicycle components, aerospace, automotive parts, etc.); semiconductor and electronics (imaging sensors, magnetic encoders, storage, backup hardware, etc.); and consumer goods or retail manufacturers (dental, beauty epoxies, post-surgical garments, etc.).

One particular functionality that makes xTuple applicable to various, seemingly unrelated industries is breeder BOM (also known as inverted BOM, or reverse BOM elsewhere in the industry), which manages coproducts and byproducts. While this need is obvious in food industries (see Fatal Flaws and Technology Choices), it is also applicable for some electronics manufacturers and metal centers. For instance, printed circuit board (PCB) makers can have one breeder item (a circuit board) that produces multiple coproducts (individual chips). The same process would be true for cut-to-size industries (see Cut-to-size/shape Industries).

… Creates Diverse xTuple Competitors

Competition is everywhere. Often, xTuple's competition comes from some Microsoft Dynamics ERP products (Microsoft Dynamics GP or Microsoft Dynamics NAV) or from Sage MAS 90/Sage MAS 200. Sometimes it is SAP Business One, sometimes an Infor ERP product (Infor VISUAL or Infor SyteLine), a Consona ERP product (Made2Manage or Intuitive), or an Epicor ERP product (Epicor Vista or Epicor Vantage). And still, a month does not go by where xTuple does not face competition from some obscure product it has never heard of before. On the other hand, the free PostBooks product, which is quite light on the manufacturing side, will compete directly with open source brethren like Compiere, OpenBravo, ERP5, Open for Business (OFbiz), and with likes of QuickBooks and Sage Peachtree.

Of the above products, the vendor touts xTuple ERP as the best combination of low cost and solid functionality and product extensibility, and for which manufacturing features are the core of the system rather than add-ons or afterthoughts. While PostBooks is quite competitive in terms of features amid open source offerings, the OpenMFG Edition often trounces the mid-tier commercial ERP offerings mentioned above when it comes to value. This is due to many free infrastructure components and less need for consulting, since the system is quite straightforward and simple to use, having a clean GUI (which is not exactly a typical feature of open source products).

Thus, an average total cost of ownership (TCO) of $50,000 (USD) for the first year is much lower than its proprietary ERP peers named above. On the other hand, xTuple ERP runs on many operating systems (OSs) and client platforms, such as Linux/UNIX, Windows, and Mac—the latter situation unheard of in the traditional ERP world.

Expansion Yet to Happen in Earnest

To be fair, xTuple's customer base is still primarily in North America, with only a few customers in Europe, Australia, and India. However, the company hopes to see a lot more international pickup with the PostBooks open source strategy. As for language support, it currently offers English, French, and Spanish; German, simplified Chinese, Russian, Portuguese, and Turkish are underway.

In terms of localization, the Qt Linguist feature is basically just a glorified Extensible Markup Language (XML) editor. The local xTuple ERP client loads a binary version of an XML translation file at runtime so that each user can see the application in his or her preferred language. xTuple has an open source project underway to allow for parallel XML translation efforts (in a database) through a web site. The project is to roll out later in 2008 and is expected to drive up translation and internationalization efforts dramatically.

In summary, the dual license approach is expected to have a positive impact. For one, the strategy does not cannibalize existing revenues or upset existing customers, as they all use and need manufacturing capabilities. On the other hand, the free accounting product will compete directly with the affordable open source "lite" (not very functional) applications, like Compiere or OpenBravo. This competition is hoped to provide a partnering opportunity with other open source projects and independent software vendors (ISVs). The company has a four-level partner program. The entry level of the program is aimed at consultants who need training on a free product, while a development partner level aims to grow third-party ecosystems, and a "gold" level adds the ability to sell a hosted, software-as-a-service (SaaS) offering.

Most xTuple partners participate at the mid-tier solution partner level—about 20 value added resellers (VARs), and ERP experts in manufacturing, having a technical focus and interest in working on source code. These are small firms, each with a local or regional footprint and defined territories. Canadian partners have helped with French translation and with goods and services tax (GST) and provincial sales tax (PST) regulatory requirements, while the UK partner has helped with value added tax (VAT) requirements. The revenue split is about 50-50 for channel, but xTuple would like to move that channel percentage up, obviously. The vendor claims to have piqued the interest of a substantial number of potential partners worldwide, with over 2,000 VAR registrations so far.

xTuple Architecture

xTuple ERP architecture is of the two-tiered client-server variety. Though this type can sound deceptively outdated, it is certainly worth taking a look at the product. The entire business logic is written in the PostgreSQL open source database; it currently consists of over 270 tables and over 1,000 business logic functions. On the other hand, the graphical client and report writer are built with the open source Qt framework. Furthermore, there is an XML-based application programming interface (API) for integration with external systems, as well as open database connectivity (ODBC) and Java database connectivity (JDBC)—both of which are compliant interfaces to other databases.

Regarding the notion that a two-tier client-server architecture might sound "passé" (outdated) compared with n-tier client-server and service-oriented architectures (SOAs)—see Architecture Evolution: From Mainframes to Service-oriented Architecture—xTuple's retort is that since the business logic is addressable in the database, one could make a pretty good argument that xTuple has been SOA-based from the get-go. It is certainly not like "traditional" fat client-dumb servers (lots of processing done by the client versus not much processing) at all. In the recent API work, xTuple has been extending this concept: making smart database views addressable for outside programs, with built-in business logic, error checking, and referential integrity enforcement. To xTuple's way of thinking, SOA is just old gruel warmed over to benefit legacy systems (see SOA: The little TLA that could by Ned Lilly of xTuple).

Another possible "catch-22" (problematic or illogical situation) here could be the possible esoteric nature of xTuple technologies—in other words, Qt, and PostgreSQL, however inexpensive (or free) and simple to learn, are not exactly as popular as Microsoft .NET or Java, and where does one find programmers for that skill? To that end, the vendor is dismissive and believes that there is nothing esoteric about PostgresSQL—it is more standard structured query language (SQL)–compliant than any commercial database. Likewise, Qt is really just C++, which is still the most widely used programming language out there.

The vendor also claims that far and away the most useful technology to know in the world of enterprise applications is plain old SQL, which is a point xTuple hammers home with its VAR partners and with the more technical users at customer sites. The vendor admits to database snobbery and to trying to capture everything over the course of processing ERP transactions. That means that, with very few exceptions, it should allow users to generate reports. But xTuple is trying its best to keep everyone current—especially the customers that buy its managed services option, which includes automatic installation of all upgrades, along with nightly off-site backups and database optimization. The company might have initially resisted the urge to call it xTuple Network (XTN), but that is certainly no longer the case, given that XTN options are now part of the price list.

It's Saas-y Too

This little-known fact may surprise some, but xTuple can support either a single-tenant, multi-instance, on-premise architecture or one that is multi-tenant, single instance, and on-demand. If a customer wants a dedicated server, then it is a pretty straightforward single-tenant case. If the hosting partner wants to go multi-tenant, then xTuple can support that option a number of ways as well (for example, one database instance or multiple discrete databases and user roles, or multiple database instances and dedicated ports). SAP and Glovia have both apparently claimed that their products, SAP Business ByDesign and GSinnovate (available since late 2006; see A Veteran Enterprise Resource Planning Vendor Makes a
SaaS-y Statement
) respectively, were the first to offer on-demand SaaS solutions "from the ground up" (starting with the basics). However, xTuple scoffs at both vendors; OpenMFG was, according to xTuple's CEO, SaaS-compliant back in 2004.

As for the dilemma about multi-tenant, SaaS, on-demand versus single-tenant, on-premise deployment (and whether any model will ultimately win, or if there will be some coexistence down the track), xTuple thinks it should all come down to customer choice. New technologies and paradigms rarely replace the old order completely, and the market usually just ends up with more fragmentation. The vendor intends to support both models until customers stop asking for one of them; xTuple has deliberately architected its system so that it is able to accommodate the choice.

SaaS, to xTuple, is really just another way of delivering the product, and the vendor and some of the partners that have been providing SaaS for years are amused by the perception that it's something new. Certainly there are some customers that are interested in more of a subscription-style model (which xTuple now supports; see www.xtuple.com/pricing ) or in having the server hosted or managed off-site, per the application service provider (ASP) mode (which both xTuple and its partners support); and for them it makes the option available. The deployment choice is just another element of xTuple's overall flexibility story (as is its ability to run on multiple platforms, discussed earlier).

Yet, No One Is Perfect

On a somewhat down note, the system is currently not readily customizable other than via the aforementioned system-wide utilities and user-based personalization, authorizations, alerts, and so on (in other words, customization is basically limited, as in the case of SaaS deployments). Namely, for major customizations, customers can get source code and develop whatever they want, which xTuple can then roll into a standard OpenMFG function. In other words, to make meaningful changes to the application, users need to modify database functions and triggers (which is reasonably easy), and possibly the client-side, which is much less easy, in C++.

Nonetheless, xTuple is talking about some additional architectural changes coming down the pike that will make the application infinitely more customizable, using the Qt scripting language (which is very "JavaScript-y"). The basic idea is that, just as the report writer currently stores report definitions in XML inside the database, the application will also store XML display layouts in the database as well. Users will then be able to customize these layouts or forms (the Qt screens are already defined with XML).

The vendor expects to dramatically reduce the amount of C++ code in the process. It is also hoped that this addition will further simplify the options for Web views of the system and enable multiple types of user interfaces and metaphors, something the vendor has discussed a lot in a thread on the www.xTuple.org site, called internally "OpenMFG Beautification." xTuple sees this as a way to leverage the xTuple applications into a more meaningful, open, complete, and standards-based version of a "business operating system,"—more than a bolt-together platform (e.g., Salesforce.com's Force.com or the new Microsoft stuff along similar lines), but a real leapfrog of the current ERP status quo.

[Editor's note: the vendor states that "lots of these initiatives were done in the 3.0 release."]

Further, and on a somewhat negative note, in addition to a still-fledgling install base, brand recognition, and global presence, xTuple ERP—a relatively "young" product—has a number of functional gaps when compared to its much older ERP peers. For one, MPS functionality, which has been a matter of course for decades within ERP systems for manufacturing, was added only two years ago. The same is true for the CRM capabilities, which are still thin on the sales force automation (SFA) and marketing automation fronts. The vendor openly admits a poor fit for most engineer-to-order (ETO; see ERP Systems and the ETO Manufacturing Market) and job shop environments. Some configure-to-order (CTO) environments too might currently be a "tall order" (too much to expect) for OpenMFG (see The Essential Components of Quote-to-order Application Suites).

Moreover, xTuple believes that, with millions of underserved potential customers (those who likely do not have ERP software yet, and would find the solutions bigger vendors offer as "overkill" or too much), there will be plenty of fish to catch in the small to medium business (SMB) ocean, even with the currently available functionality (not to mention the help from the community of a few hundred developers from both current customers and partners). The vendor states that the easy answer to future enhancements and directions is what the community or customers want; and that answer is found as people voice their opinions with their keyboards, so to speak, and their wallets. The "hit list" (or most popular functionalities) for version 2.3, released at the end of 2007, was

  • returns and service management

  • detailed revision tracking in BOMs and routings/operations

  • integrated XML import tool with predefined maps for Yahoo Stores integration

  • further expansion of the API, including simplified views of item and CRM information

  • overhaul of units of measure (UOM)

  • ability to reserve (allocate) particular stock for a sales orders

  • major enhancements to the UI, including column width and window size/location memory

  • over 100 other improvements and bug fixes

For background on the 13 most popular community-contributed functionalities, see http://www.xtuple.com/news/20080325.php.

The big areas the vendor is looking at for version 3.0 and beyond include

  • more Web-based options (multiple partners are working on this)

  • more ETO and job-shop functionality

  • expanded project capabilities (for manufacturing, distribution, and professional services)

  • retail and point of sale (POS) functionality, already sponsored by xTuple's first Global 1000 customer

  • full distribution requirements planning (DRP), for which many pieces are already in place

  • light product lifecycle management (PLM)

  • more predefined API and XML "maps" to other systems (for example, shopping cart, EDI, computer-aided design [CAD], etc.)

While xTuple ERP deployments feature on average 20 or so users, the question is whether there is a limit to how many users the system can support, or whether there is no limit per se (based on performance, etc.). In fact, xTuple wants to stay away from competing with the "bigger guys." To that end, the vendor has not defined an upper limit yet, as the application scales with PostgreSQL (which has systems with thousands of concurrent users and terabytes of data). However, xTuple is starting to talk with a few bigger accounts, and is working on some advanced load-testing beyond the relatively modest simulation suite that it has today. This will entail hundreds of users "hitting" (using simultaneously) a PostBooks database, and xTuple will be contributing performance benchmarks to the PostgreSQL community as one of many release-to-release application benchmarks they can run (see the PostgreSQL site for a thread on this topic).

Thus, an ideal prospective xTuple customer would be a manufacturer (ideally in the MTO, MTS, mixed-mode, or batch process environment) that has outgrown desktop accounting, wants to replace a customized legacy system, or is starting up operations with the expectation of fast growth. Its typical revenues would range from $5 million (USD) to $50 million (USD)—although currently, both smaller and larger sites are using xTuple—while local or regional coverage today is supported throughout the United States, Canada, the European Union (EU), Australia, and India. Conversely, larger manufacturers with heavy process manufacturing requirements (see What Makes Process Process?) would be a stretch today, as would highly ETO or job shopping requirements, even with the recently added features in the 2.3 release. Such environments are advised to look at other options for now.

TEC Runs xTuple through the Gamut of Questions

In interview with the author, xTuple's CEO, Ned Lilly, offers his company's opinions on the current state of the software market.

TEC: SOA stack approaches (and the wars between the various vendor offerings): are they all converging to virtually the same thing, or might there still be some differentiation?

xTuple: We agree with Infor here—they are most definitely not converging to the same thing. SOA stacks are indisputably a battleground for what we'd call "soft lock-in" (in other words, lock-in by choice of infrastructure tools, not top-level applications). It's a lot like what we see in the open-source market—each vendor's perspective on SOA is driven by the businesses and products it has to defend. Oracle and IBM are keen on open source for operating systems and lightweight application servers, but not so much for enterprise-class databases. SAP is downright dismissive about open source applications.

In the same way, Microsoft and Oracle see SOA as another opportunity to build "stackiness" for their infrastructure businesses, which in both cases bring in more revenue than the application side of the house. Infor, with a giant bag full of disparate technologies and decades-old products, quite naturally sees it as a giant systems integration project. I don't know a great deal about IFS, but it seems to me that their response is not unlike what we'd say: "Um, we designed our product in an open standards, granular/services-friendly manner in the first place."

TEC: Any comment regarding pro-et-contras for both the platform choice and rationalization ("lock-in," and constantly waiting for the moving parts to be in sync)?

xTuple: Again, how you answer the question depends on where you're coming from. For xTuple, being operating system-agnostic has been a major driver from the get-go. Both our GUI client and our server run perfectly well on Microsoft Windows, Linux, and Mac OS X—and we have customers running every conceivable combination of the six. We believe that's a huge part of our appeal in the marketplace: as Apple and the various Linux distributions continue to grow at the expense of Windows, we are better positioned than anyone to give customers the ultimate in platform flexibility.

The place where we don't offer a choice is a little different. We've selected the open-source PostgreSQL database, because we can't conceive of a good reason to use anything else. It's more mature, robust, scalable, and better-supported than any other open source database, and it offers a level of standards-compliance, freedom, and flexibility unmatched by any proprietary commercial database. Also, because we eschew a Java or .NET middle tier, and leverage the database procedural language for all the business logic, adding new front-ends (such as Web dashboards or wireless handheld devices, both of which have been implemented by our VAR partners) is a snap. So there's another element of platform freedom.

TEC: What are your views regarding the "wrap-around versus rewrite" dilemma? Will any products in your family be completely rewritten in managed code?

xTuple: It's self-evident that massive code rewrites are to software companies as big icebergs are to cruise liners. We're in the happy position of being able to update our code in a modular way—and in fact, since customers and VAR partners both have full access to the source code, they don't have to take our word for it—they can do it themselves. Many have.

The challenge for us is proactive management of what is, essentially, a giant, globally-distributed ongoing software development project. Through an open communications process, well-defined standards and practices, and active peer review, we believe that we have a powerful solution to code maintenance that will continue to elude even those vendors with well-architected legacy products.

TEC: Vendors often de-emphasize major upgrades, turning rather to vertically oriented and optional value or service packs. Anything on that matter on your side—that is, what will the "quantum leap" versions of your products be?

xTuple: Since we're a younger company, we still see plenty of room to grow with major releases. Our observation is that the definition (footprint) of ERP continues to expand—now incorporating elements of CRM, PLM, and manufacturing execution system (MES) for starters. Our community-based development approach is well aligned with this market reality. We do believe that there is a continuing opportunity for vertical specialization, but our instinct is to accommodate vertical market distinctions in the core code whenever possible. As an example, it's more interesting, and useful to a broader group of people, to find commonality in how food manufacturers and PCB manufacturers use breeder BOMs, than to architect overly narrow bolt-on solutions for each one separately. This is another area in which having an open, transparent development community is a huge asset.

TEC: Microsoft Desktop supremacy—solo, Duet, or can many still play at this game?

xTuple: Who cares?! We believe Microsoft Windows is an important platform, and will have a major role to play for the foreseeable future. But the market has a funny way of resisting attempts to extend the platforms that are overly complex, "kludgey" [made up of components that are not well matched], or that raise questions of privacy or anticompetitive behavior. We believe in five years the desktop market will be considerably more diverse than it is today—which is why robust support for Linux and Macintosh is so important to us.

TEC: Incidentally, what about partners—ISVs and VARs? Anything similar (or more impressive) on your side, with regard to the above?

xTuple: VARs are a major component of our strategic plan going forward, and a very strong fit with our community-based development approach. Through the past six releases, we've had numerous examples of a local VAR partner capturing customer requirements, working with our community to design and code the solution, and then us certifying, testing, and including the new functionality in the next supported release of xTuple ERP. We believe that as the consolidation continues in the ERP market (see our associated blog, The ERP Graveyard, for more on this), the value proposition of a vendor who works with VARs in this manner will become ever more apparent. Not only do we not compete with VARs for sales, we embrace them in the product development cycle, and give them additional tools for increasing customer satisfaction and gaining greater revenue.

TEC: Going mid-market versus defending it: what, to your mind, will the key success factors (KSFs) be in this market, and what have you additionally been doing there?

xTuple: We agree that SMB customers need most, if not all, of the same functionality as larger enterprises. What they don't need is the price tag, complexity, lock-in, sales pressure, and implementation headaches. So, for us, the KSFs are simple:

  • Continue to focus on delivering a simple but powerful product that can service everyone from the Intuit QuickBooks upgrade to the SAP replacement (an accurate description of our customer base today); and

  • Continue to build a strong VAR channel, and leverage their expertise through an open, community-based approach to product development, implementation methodologies, and post-sales support.

TEC: SureStep, QuickStep, and Accelerate: breakthrough implementation methodologies, or just "baby steps" in that regard? Are we missing something earth-shattering in your offering there?

xTuple: Yawn. We don't try and convince our customers that there is a whole lot new under the sun in ERP implementation methodologies. Rather, as indicated above, our emphasis is on leveraging an open, active community of users to promote best practices. The emphasis, we believe, should be on accountability and successful delivery of services, rather than on trying to bamboozle customers into thinking professional services can be bought like cans of soup.

At the end of the day, a successful ERP implementation is all about the people involved in the project. We believe that we are attracting a lot of smart, experienced people who understand that—and are eager to partner with a vendor whose product, flexibility, and TCO help to eliminate some of the more obvious impediments to success and put the focus where it needs to be: on following a defined process (whichever one you like best), and on successfully managing the people.

TEC: At the end of the day, who do you think is in a better position to ultimately win in the market? Or maybe none of them will, and all of them will simply remain at the current equidistant positions?

xTuple: No market is static; as our ERP Graveyard blog (www.erpgraveyard.com) bears witness, the vendor picture changes almost weekly. But we believe that the market is so fragmented—largely due to the nature of ERP in the first place—that it's unlikely there will be one dominant "winner." We believe the "tier 1" players will continue to have mixed success moving down-market; the mid-tier will continue to consolidate through mergers and acquisitions (M&A); and Microsoft Dynamics will pick up share just because they're Microsoft.

But we'd be surprised if any one player ended up with more than half of the mid-market, which we'd define as companies with $100 million (USD) to $1 billion (USD) in revenues. We'd be shocked if anyone had more than 10 percent of the under-$100 million market; but of course, that's a group that's notoriously hard to measure. In any case, we believe that the appeal of open (open source, open standards, open communications), multi-platform, community-based solutions will only continue to grow—and that xTuple will continue to grow with it.

Further Readings on xTuple and OpenMFG

For more on xTuple's genesis and thought process in its earlier years, see TEC Talks to OpenMFG—Free and Open Source Software Business Models and Evaluating Alternatives: Key Questions to Ask When Considering an Alternative ERP/MRP System.

comments powered by Disqus