A look at the Rational game plan

Chairman and CEO Paul Levy co-founded the Cupertino, Calif.-based Rational in 1981. Coming out of the Air Force and Stanford, Levy and company built a methodologies company, which they have, of late, ably recast as a first-class development environment provider. Levy's vision is to build a tools house that can stand alongside players like SAP, Intel and Microsoft. He recently spoke with Editor Michael Bucken.

When and why did you and Mike Devlin found Rational?

No business in the world in the late 1970s was dedicated to helping companies that depended on developing software turn that dependence into a strategic competitive advantage. It was just about this time that Mike and I both went back to graduate school at Stanford.

Pretty much near the end of our tenure at Stanford, we decided to start the company -- we recognized that software engineering technology hadn't been developed. We wrote the business plan for the company in 1980. Mike and I, as Air Force officers, didn't have a whole lot of money. But we did invest our life savings at that point in time to try and get the company started just a little bit in order to refine some of our technical and product ideas. Once we expended all of our personal financial reserves, only at that point in time did we go out to raise capital in 1981.

Our mission was to ensure the success of customers whose businesses depended on developing mission-critical software. And we may have wordsmithed it a little bit, but that mission hasn't changed in 18 years.

What are the key elements of Rational's strategy?

With the standardization of the Unified Modeling Language [UML], Rational is now able to put itself on the radar screen as the company that has probably had more to do with -- and has more technical insight and experience -- pioneering best practices for developing mission-critical software, evolving or developing those best practices, and then helping customers exploit those best practices.

The next key element or attribute of our strategy has to do with first identifying if a customer -- like Chase Manhattan Bank, Fidelity, Boeing or Lucent -- wants to put into practice the software engineering principals that Rational pioneered and developed. What tools do they need to automate all of the critical aspects of using these best practices? Over the last 36 months, working closely with customers, we tried to refine a view of what you really need to provide from the customer's standpoint -- not Rational's standpoint -- and what would approximate the complete solution for automation of these best practices that customers require.

First and foremost is requirements management. Customers need to be able to define, track, manage and relate the requirements motivating the development of mission-critical software applications to everything from architecture models, test cases [and] different versions, [in order] to incorporate the customer feedback [needed] to refine and change requirements over the entire developmental life cycle of that software application. That led us to focus a lot of attention on requirements management and ultimately to us bringing the RequisitePro product to market.

The second round of Rational's strategy was to identify the categories in which the customer needs to have a product to provide a complete solution. And then [we want to] make sure that we have the best-in-class product in each of those categories -- not a second- or third-tier product, and not a check-in-the-box product. I think Rational's deep heritage has always been that we believe that the best product ultimately wins. You have to do a lot [of other things] right, but ultimately the best product wins in the markets we serve.

When did you decide to fill out the product line?

It became apparent to us about 24 months ago that this whole notion of visually modeling an architecture was going to become, maybe, the next revolutionary technology that could broadly affect the way software was developed. Not as a standalone product category, because the power, we came to recognize -- once you had a semantically rich graphical model of your architecture -- was to be able to relate to everything you needed to do with that architecture to develop an application. This means that you have to be able to relate it to requirements, you have to be able to test, and you have to have some kind of configuration or change management infrastructure to enable process and scalability.

What became apparent was that as important as modeling would be, we were on to something that was a lot bigger than just visual modeling. Our feeling was that we had two choices: Either try to exploit the modeling franchise as a point product or to be sort of hyper-aggressive and, ultimately, try to build a suite of products around modeling. [This would let] customers use the power of the graphical model as a means to take their software development capability, based on these practices, to a level that I think was unimaginable even 18 months ago.

Was the plan right then and there to buy or were you going to build first?

Our aim was to find the best technology, the best people and the best products in the world in these market areas, and then try to combine them in a unified company, with a unified product architecture -- the emphasis being that they had to be the best product in each category and we had to do it fast.

Fast forward to where things might be five years from now. You have people like Intel providing the underlying hardware technology in the industry. You've got people like IBM Global Solutions providing massive system integration services for developing general-purpose business applications. You've got people like SAP eliminating the tedium of developing from scratch every time. You've got people like Microsoft providing the basic hard software infrastructure and platform technology. And there's one seat left at that table. That's Rational's seat.

If you talk about the new economy, which is based on knowledge and technology -- and what it's going to take to make that happen -- I think the linchpin between Intel, Microsoft, SAP and IBM is Rational. The linchpin has to figure out how to exploit all of this technology and all of these services capabilities, and that comes from creating a quantum leap improvement in the development of mission-critical apps.

My feeling was that to earn that seat at the table five years from now, we have to be in front because these markets are dynamic, they're moving very quickly. You get out in front and you can leverage a financial performance. You identify yourself as the company providing this solution in the mind of the customer. The opportunity to become the leader in that market space, an important market space, we felt was ours by birthright, if you like, because we've been pioneering these technologies for 15 years.

The third element of our strategy was to have best-in-class products that, taken together, enable the customer to deploy a complete solution.

How important was acceptance of UML to the strategy?

Very important. Before UML, this whole notion that there is a better way, a more rigorous way, to build mission-critical software was viewed as sort of fiction. In the past, most Case technologies failed miserably. Customers were more than skeptical, they were almost depressed. The hardware technology had improved by six orders of magnitude, [but] the software seemed mired down in the swamps. UML provided the first glimmer of hope because it became a cross-platform standard.

With UML, customers began to broadly realize that software engineering might finally be coming of age. Software, all of a sudden, was becoming more amenable to a process that would be developed with the same amount of rigor and the same relatively predictable results as when you build a bridge or a skyscraper or a hardware application. UML, I think, really put Rational on the map.

You talk like it happened overnight. How long was the UML effort? Has its acceptance helped Rational expand its customer base?

UML was almost a two-year campaign. We began UML almost when we began executing the strategy to assemble the full complement of product. There was the work of Grady [Booch] and Jim [Rumbaugh] and Igor [Jacobson]. Getting Microsoft on board. Getting Oracle on board. Getting IBM on board. Getting the Object Management Group [OMG] on board -- opening up the process. And we tried to do the right thing. Our view was that getting the market broadly across the chasm, relative to UML and best practices, would benefit the whole industry. No one would benefit more than Rational.

To make it happen we knew we had to run an open process. We tried to make UML the best it could be without diluting the standard by having it become designed by committee.

You were on a buying binge for about two years. How has the acquisition strategy evolved?

Mike and I have been doing this for 18 years. We're pretty battle-hardened. We made a lot of mistakes. But one thing that we believe from a management standpoint is that you need a shared vision that people can believe in. You need a mission that focuses everything that everyone in the company does and that provides people [with the] context of "doing the right thing" as you empower your frontline people. You need a disciplined financial plan that represents a set of shared commitments that every manager and every individual in the company believes is their responsibility to implement and execute.

Then you needed common and shared core values, whether it was the acquisition of SQA or Performance Awareness or Re-quisite or even Pure Atria, which was very challenging because that [firm] was roughly our size. From Day One we helped people understand: here's the mission, here's the vision.

What we told people was 'Look, I don't care whether your company heritage was from Rational, from Pure, from Atria, from Performance Awareness or from SQA. It doesn't matter. You should all be proud of that. I'm proud that my heritage was in Rational. But on this day, we are founding a new company and we're all equal as founders in this new company.'

How do you stack up versus another aggressive acquirer, Platinum?

The modeling game is over. [Rational] Rose wins. I think the same is true with requirements management. I'm not saying there won't be other companies that might aspire to do what we're doing, but I don't think we're going to be caught from behind. But complacency is dead. I respect all of these competitors. We're going to be focused on continuing to invest in each and every one of our products, and in unifying our products together and growing ourselves.

Can the close relationship with Microsoft cause problems?

The industry is dynamic. Things change. The good news is that we're really, really close to Microsoft. We now have 75 people up at our Redmond [Wash.] site. We're on the canvas with Microsoft so, independent of their intentions, we understand their capabilities.

I think the Microsoft relationship will grow closer. I think we'll expand in areas like testing configuration management, and I think there's a very good chance that two or three years from now, Rational will emerge as Microsoft's most valuable partner in the software industry.

How hard has it been to evolve from Unix-only to multiplatform offerings?

Our aim is to bring out an independent platform to our customers that depends on mission-critical software. We have no bias, no religious attachment to any particular platform. In the cold, hard light of day, when you look at growth rate and you look at Windows and NT on the server side for developing mission-critical applications, that's where the growth is.

I would estimate that a year ago, less than 20% of our total product bookings were products on Windows and NT. In our March quarter, more than 40% of total license bookings are on the NT Windows platform. That's a 'double' on a year-to-year basis.