Does the Internet Make Software Better?

The enterprise software market is becoming a battle of “platforms,” big, interwoven networks of applications that form the basis of CIOs’ architectures and attract smaller companies to build secondary applications that work with the big platforms. The leaders in the platform races are, of course, Microsoft (operating systems and desktop software), SAP (and its big suite of enterprise applications) and Oracle (databases and a growing set of enterprise applications). But the battle isn’t just between these vendors, says Accenture in a very interesting report.

The real battle is between vendor platforms and the Internet platform, which threatens to blow them apart by becoming the dominant assembly line for software development and the distribution model for a much more fragmented set of software services that can come from any source. Basically, it’s the question of whether you will continue to install software on the premises or get it all through the Internet. Accenture calls the Internet platform the winner in a decision, but not before examining every possible scenario the authors can think of, which is what makes the report so interesting.

The wild card in the battle seems to be Wall Street expectations. Accenture says that between one-half and two-thirds of a typical software vendor’s value is based on market expectations of growth. “Take away that expectation,” say the authors, “and enterprise software vendors will see their market capitalizations plummet and their CEOs looking for work.” Accenture says 50-80 percent of the typical software vendor’s market value is based on expectations of future growth. Google’s stock price is taken to be Wall Street’s demonstration of support for the Internet platform as the future growth winner.

But the authors seem overly optimistic about the pace of standards and SOA, saying open standards are now the norm and SOAs are the de-facto architecture for new software products. That seems to overstate things quite a bit. They also say utility computing models are disrupting markets like CRM. It’s hard to see companies like Salesforce.com as utility computing. And it’s difficult to see vendors creating a truly standard, vendor-neutral SOA architecture behind their products unless they are forced to do so. The real lure of SOA for big vendors would be if it allowed them to speed up their innovation capabilities. That would free them from having to develop big suites all at once and enable them to sell more incremental products at a faster pace–which would no doubt keep Wall Street happier.

But that leaves their dominant position open to threats from vendors we used to call “best-of-breed” but now call “software as a service” (Saas) providers that use the Internet as a distribution model. To keep smaller, more nimble vendors from creating a disruption in the big platform players’ markets, the big players will have to find a way to co-opt them, as Salesforce is trying to do with its AppExchange, by building a vague approximation of an open source community around its proprietary platform of applications, or as SAP is trying to do with its proprietary NetWeaver integration middleware.

Accenture argues that there is a disruptive trend of “good enough” today that says time to value is more important than fancy features and functions and that CIOs are more open to considering Saas alternatives in this atmosphere. Yet, research, including Accenture’s, has shown that CIOs like Saas in theory but not so much in practice–at least not yet. Most say it just doesn’t meet true enterprise requirements–only parts of them. Where the Saas model is truly disruptive today is among smaller companies that cannot afford the software offered by big, established vendors. Gartner estimates that 80 percent of Salesforce’s customers are 20 users and fewer and that Salesforce has just 8 to 16 live deployments with more than 1000 users.

Another problem with the Internet as dominant platform is that standards are too low in the infrastructure today to affect the business logic of enterprise software applications. The big vendors have plenty of time to adopt an Saas distribution model for their traditional applications–as SAP announced this week. A distribution model is easier to duplicate than intellectual property or a business model. If, as SAP would have everyone believe, Salesforce is just another application vendor who happens to use the Internet as a distribution mechanism, then it may have Salesforce, and the Internet platform threat, licked.

But Accenture hints that there may be more to the Internet platform than a distribution model. Survival for the big vendors may depend on giving up the arm’s length relationship they have with customers–where outside consultants take responsibility for installing the software and satisfying the customer–and begin to do that lower-margin stuff themselves over the Internet. It is the core of the software as a service (Saas) model that Salesforce popularized. Salesforce has a direct conduit into how companies use its software and what they want because customers access Salesforce’s own servers every time they want to use the software. Theoretically anyway, they keep their customers happier with that direct connection. That’s a blessing and a curse, of course, because if Saas vendors try to please everyone, they will fragment their platforms and lose the economies of scale that come from serving everyone from the same code base.

Interestingly, Accenture believes the preference for fewer, bigger vendors among CIOs is skin deep. Only 35 percent of CIOs surveyed by Accenture see software vendor mergers as a positive development. Even fewer–17 percent–see mergers as resulting in better service or more reliable products. The expected benefits of consolidation–better integrated software infrastructure and a lighter vendor management burden–were seen by 65 and 61 percent, respectively, of CIOs. Seems like those numbers should have been higher to constitute a ringing endorsement of consolidation as a true problem solver.

Accenture says that the software market is much more fractured than it seems–the median segment is worth just $1 billion–leaving big vendors with little real economies of scale in R&D. The only real economies are in the ability to consolidate sales and marketing groups–which can be significant, since most companies spend 40-60 percent of revenue on sales, marketing and back-office support. But how efficient can the salespeople be when they have to sell many different acquired or SOA based products? And consolidating the products leaves vendors open to competition from the internet platform and the Saas vendors and smaller upstarts that use it–especially if, as Accenture asserts, the “good enough” spending theory will continue to hold sway for the foreseeable future.

Accenture calls the internet the “ultimate platform,” as more standards continue to emerge. But there is still plenty of room for lock-in on top of the low-level infrastructure standards that have emerged so far, it seems.

All of this also ignores the larger trend in business process outsourcing. Given the choice between buying more software to support a non-differentiating business process or sending it out to someone else, business people will choose getting rid of the entire process–not just the technology–every time. And our recent story on research from MIT shows that BPO customers are much happier than traditional IT outsourcing customers. Customers don’t care what software platform ADP uses for payroll, or UPS and Fedex use for warehouse management, all they know is that these companies take care of something they don’t want to do themselves.

Do you think the internet platform will ultimately smash the big platform vendors’ software suites into smithereens?