BenefitsEmployee Engagement

Next-gen HRMS Upgrades Pose Thorny Dilemma for BPO Vendors

As SAP, Oracle, and Lawson roll out their next-generation architecture, will BPO and ITO providers migrate their clients? It’s a tall order, but it could bring a tidal wave of BPO converts.

by Naomi Lee Bloom

Just when you thought you might have a few minutes to improve your overall HRM delivery system to include consideration of further outsourcing, we find ourselves in the middle of yet another paradigm shift in HRM applications software architecture.

I count eight generations of application software architectures so far on my watch (see Software Architecture v8.0 sidebar), therefore complete redesigns and rewrites are needed to achieve a true implementation, as opposed to a “marketing” story of the noted architecture. However, we all know that there have always been workarounds for vendors who needed the “marketing” story of the next generation without doing the heavy lifting of application redesign and rewrite.

We won’t discuss here (or probably ever!) the not-quite parallel generations in software engineering life cycles and design through programming methodologies. We also won’t discuss here (but will cover in my next column) the many innovations in HRM software payment and deployment options that we also face.

What’s important to us here is that this current paradigm shift in software architecture to multi-tenant web services/SOA applications—i.e., those that are truly able to be delivered via SaaS, which all of the major and many minor HRM software vendors espouse—is happening at the same time as the explosion in software payment and deployment options. Software architecture, deployment options, business models, etc., are deeply intertwined, but their reality is frequently obfuscated by marketing, sales, and PR folks putting lipstick on whatever pig of a solution they are selling.

The bottom line is that we are facing the release of a whole new generation of application software, and we need to plan now (as do our BPO and ITO partners) for if, how, and when to migrate to this new generation in search of the presumed benefits. Keep in mind that we’re speaking here about all new architectures and all new code and object models, except where a specific vendor chooses to maintain backward compatibility by re-incarnating their dated or never-correct HRM object models in shiny new web services. Show me an SOA/web services HRM application that still has first name/middle name/last name under the covers along with the “solve all data design problems” employee status code (or worker status code), and I’ll show you a vendor that has not rethought its HRM object models for the 21st century.

There will certainly be some vendors who take this approach, and some of the blame for them hobbling their next generation in pursuit of backward
compatibility lies squarely with the installed base, whom they dare not push too far. For vendors that are sticking to the familiar, I conjure up the ghosts of MSA past.

I’ve said more than once that an implementation of Oracle’s Fusion applications will be a new implementation, not what we think of as an upgrade for Oracle HRMS, PeopleSoft HRMS, or JDEdwards HR/payroll. A new object model means, at a minimum, data discontinuities. And a new presentation of functionality means that the entire user experience must be re-implemented. Furthermore, many Fusion applications will be new code, with an inherent lack of full functionality and the likelihood in Release 1 of “quirks.”

The same could be said for SAP’s next-generation applications for its traditional target market, which have been spoken of for 2010 and are obviously going to be some evolution/variation of its just-released Business ByDesign. The good news is that by the time SAP launches its next generation for larger organizations, it will have had time to mature that software.

Lawson is also on a next-generation path, releasing in late 2007 a suite of strategic HRM applications (something truly new to its product line) built on its new architecture. It’s now clear that these new applications are more interfaced than integrated with the company’s current generation of core payroll, benefits, etc., but that’s a different issue.

Many, if not most, of the established, so-called talent management software vendors are also in various stages of evolving if not rethinking their own object models and architectures as they attempt to deliver complete and integrated “talent management suites” that are totally SOA/web services and exclusively offered on a SaaS basis. And you’re all familiar with Workday, which is SOA/web services and sufficiently multi-tenant to support SaaS delivery, but may well lack some of the BPO-related multi-tenant features that move beyond reducing total cost of ownership (TCO) to total cost of service delivery (TCSD).

SAP, Oracle, and Lawson are being used today as the core platform of HRM BPO delivery systems (albeit without being entirely suited to that purpose), and many of these customers signed up with the expectation that their BPO provider would take them to the next generation (next upgrade?) of their preferred platform. It will be interesting to see whether the underlying software licenses—whether held by the customer or by the BPO provider—entitle the licensee to a full replacement generation of software at no cost other than being current with ongoing maintenance payments. My guess here is that many of those license agreements weren’t written tightly enough to protect the licensee when the upgrade is really an entirely new product that just happens to address the same parts of the HRM domain.

It also will be interesting to see just how quickly those BPO providers implement next-gen software as it becomes available. My guess would be sometime after the second coming, if the direct costs of these new implementations are expected to be paid by the BPO provider. Can you even imagine Convergys moving DuPont, or ADP moving all of its GlobalView customers to an early release of SAP’s next gen? Or Accenture, IBM, or Hewitt moving any of their lift-and-shift BPO customers from whatever release of PeopleSoft, Oracle, or SAP they brought with them to Fusion or SAP’s next generation? A far more likely scenario is that these and many other BPO providers will use the occasion of legacy ERP overhauls to introduce semi-bespoke core platforms, but that’s also a great topic for another column.

For end-users not yet committed to comprehensive HRM BPO, I believe that when Oracle/PeopleSoft/SAP/Lawson’s HRMS on-premise installed base realizes the cost and workload for getting to their next generation, and when senior managements realize what this means in terms of levels of investment, there will be a tidal wave of interest in comprehensive HRM BPO. But that interest won’t translate into new BPO business unless there are suppliers with not only capacity but also a believable software platform story. When SaaS is done properly, it does greatly reduce the upfront costs of using richly functional HRM software, but it doesn’t materially reduce the level of effort needed on someone’s part to configure that software to reflect the user’s business rules and processes, to load the user’s data from a differently modeled data design, or to provide ongoing service delivery.

Adoption of comprehensive HRM BPO, at least in larger organizations, has moved much more slowly than I predicted just a few years ago, but one reason for the delay has been the slow pace of delivery of disruptive next-generation software from today’s primary ERP/HRMS vendors. If Workday picks up steam, as PeopleSoft did in the late 1980s, or something else pushes HR leaders to take an interest in BPO, hang on tight as Oracle, SAP, and Lawson rush to prove their bona fides in truly multi-tenant, BPO-ready, SOA/web services application software architecture. Right now, the high end of the global market is still insisting on big-brand ERPs as their core HRMS, even with comprehensive HRM BPO, but that will change abruptly when far better service levels and costs are possible, either from the next generation of these big brands, from newer HRMS packages, or from BPO provider-owned platforms.

Software Architecture v8.0

In rough sequence but with the recognition that there have been chronological overlaps and remnants of every generation still running, the eight generations of HRM software applications architecture are:

  • No user terminal with batch-only processing;
  • Dumb user terminal (data entry only) with some real-time processing. To achieve real-time processing required a complete rethinking of the application architecture, as well as development of enabling operating systems and messaging software, hence the first real generation of the technology stack;
  • Client/server (essentially real-time cooperative processing) with a very fat client. Here we began the “discussion” about how many tiers (two-tier, three-tier, n-tier) were best and under what circumstances—a discussion that has receded into the background because n-tier quietly won;
  • Client/server (mostly real-time cooperative processing) with a thin, HTML-only client (at least for infrequent users). Here we have another “discussion” that is somewhat but not entirely separate from the previous generation about whether a pure, HTML thin client was suitable for frequent/power users;
  • Web-enabled (a.k.a. server-based computing). This is an approach for taking essentially client/server applications and running them entirely on a server with some additional technology (Citrix comes immediately to mind) that allows the server to publish or serve up these applications over a corporate network or the Internet; while there’s no shame in doing this, it’s a far cry from taking full advantage of what the Web really offers;
  • Native Web is an applications architecture that not only takes full advantage of the distributed processing capabilities of the Web, but also allows use of the full range of Web 2.0 facilities, so that ever-more-interactive and rich-user experiences are possible without the return of a heavy client;
  • SOA/web services as a traditional on-premise application; and
  • SOA/web services built to be SaaS, with the “discussion” already at full steam about whether applications and databases should be redesigned and rebuilt for full multi-tenancy (I’m a very strong proponent of multi-tenancy, especially for BPO providers that must bear all of the costs of service delivery, not just the data center operating costs) or whether some of the same benefits in reduction of operational complexity and cost can be achieved through various methods of server virtualization.
Tags: Benefits, Employee Engagement

Related Articles