Will client/server go the way of dinosaurs?
The pundits who predict the end of software as we know it are probably being a little premature, despite all the progress in hosted applications
Under the powder-blue skies of San Francisco this month, Salesforce.com employees and disciples milled around sporting “No Software” lapel badges with the distinctive Ghostbusters-style red-bar logo. In a keynote speech, CEO Marc Benioff celebrated the “end of software” and analysts flocked to back his Orwellian “hosting good, on-premises applications bad” mantra subtext.
It must be something in the air. From the hippies in Haight-Ashbury to the yippies in UCL Berkeley and the beatniks in North Beach, the city by the sea has seen much revolutionary fervour in the past. Benioff is not a man lacking charisma – much less, confidence – but his election to figure-head of a radical business change merits qualification.
This is not to knock Salesforce and its hosted CRM tools. Benioff is a seer and has vaulted past every objection in his way. Is Salesforce only for small companies? Try ADP and Merrill Lynch for 5,000-seat sites. Enthusiastic partners and customers? I’ve never seen a vendor so keen to put users in the witness stand. Real profits and revenue growth? In shovels. Resilience and reliability? Seemingly. Company stability? I wouldn’t be surprised if Microsoft came bearing billions but this is a firm that would fall in value if acquired, such is its culture.
Salesforce is the toastmaster for a bunch of smart firms that are changing the rules of business applications, from NetSuite and Intacct in ERP, to RightNow in marketing campaign software, and Ketera in spend management. Where I differ from the most evangelical supporters of the trend is in the assumption that it will crush the client/server model. The software-as-a-service crew has a powerful story to tell in terms of low capital expenditure, predictable costs, spread payment, mobile access and incremental functionality gains but I don’t think it’s enough to displace on-premises software in one fell swoop.
If it were, then Oracle must be a lot stupider than I give it credit for. It didn’t agree to spend $5.85bn – over double the Salesforce market cap – on Siebel because it sees a spectacular growth engine but because it knows most customers will continue to pay maintenance revenue on a critical application for years and even decades to come.
The situation is analogous to the transition from the mainframe to client/server. In the early 1990s, anybody who said that big iron would continue to prosper would have been regarded as an ostrich with head in sand. Supporters of Salesforce and its ilk see Oracle as a new Computer Associates, mopping up client/server apps and companies in the same way CA did with mainframe software firms. The comparison is legitimate but fails to take into account the fact that CA prospered on the back of its spending spree and that mainframes continue to appeal today albeit at a flat growth rate.
When pundits compare the hyper-growth of a Salesforce against a client/server stick-in-the-mud they are contrasting apples with pears. Mature companies grow slowly but as SAP, Oracle and several others show, this is far from being an indicator of imminent demise.
As defenders of the mainframe computing model used to quip, dinosaurs were dumb but that didn’t stop them ruling the world for millions of years. Salesforce is a great company and the application service provider model is proving its value but firms won’t shift away from proven processes unless there is a hugely compelling reason.