HP's $1.6 billion acquisition of Opsware: the view from the buyer

Executive overview

HP was fifth in the data center automation market despite 200 engineers and $100 million invested. The market had already consolidated — Opsware, BladeLogic, and IBM held 80% of it. In a consolidating market, there is no build path for a channel company like HP; the only options are buy or exit.

HP acquired Opsware in 2007 for $1.6 billion. The thesis: automation and monitoring for IT operations would converge, and owning the market leader would make HP the dominant platform for managing enterprise data centers as virtualisation exploded.

In a consolidating market, profitability only flows to number one or two — if there's no credible path there, the only rational move is acquisition.

The LoudCloud origin: right idea, wrong timing

• LoudCloud, founded 1999 by Mark Andreessen and Ben Horowitz, was designed to be AWS before AWS existed.
• The company raised $346 million before and after IPO-ing in March 2001, then watched its entire customer base — dot-com startups — go out of business.
• In August 2002 it sold the managed services business to EDS for $60 million — the whole revenue-generating business — and pivoted to the internal tooling it had built to run its own data centres.
• That tool, Opsware, was an automated way to provision, deploy, and manage servers at scale.
• The pivot was done as a public company, which made it an almost impossible task: internal ops tooling had to be productised into shrink-wrap enterprise software.
• The core advantage of the pivot: Opsware's team had lived the exact persona they were building for — server admins managing complex environments — years before the market matured to need it.

Why AWS succeeded where LoudCloud failed

• Developer adoption had exploded by the mid-2000s, vastly expanding the potential customer base.
• The cost of starting a company dropped to near zero, multiplying the number of targets.
• Enterprise security and privacy concerns evolved enough to allow cloud adoption.
• Server virtualisation — pioneered by VMware — didn't exist at scale in 1999; without it, every customer needed dedicated physical hardware, making the economics unworkable.

HP's build-versus-buy calculus

• Michelle Feaster inherited an HP data centre automation business generating ~$20 million in revenue at fifth place in the market.
• Two diagnostic questions drove the decision: what stage is the market in (emerging vs. consolidating)? And who are we as a company?
• The market was clearly consolidating — 80% share in three vendors is not an emerging market.
• HP's strength was its sales channel, not product innovation; Sun Tzu's principle applied — know yourself before choosing your strategy.
• Conclusion: in a consolidating market, HP could not build its way to number one. The only options were buy or exit.
• Supporting data: channel fit was high — HP's OpenView product was already selling to the exact IT operations buyers who would purchase Opsware.

Why Opsware over BladeLogic

• BladeLogic had a better product in head-to-head usability tests; Opsware won on enterprise scale.
• Opsware had extended beyond server automation to a full data centre automation suite (adding runbook and network automation via acquisitions); BladeLogic had partnerships but not products.
• HP's strategic thesis was end-to-end service deployment — desktop, server, network, storage — so Opsware's suite aligned directly; buying BladeLogic would have required three additional acquisitions.
• Executing four acquisitions successfully was judged far riskier than one larger deal.
• Larger acquisitions often integrate more successfully than small tuck-ins: there is critical mass of people and knowledge to enable the acquiring sales organisation.

Integration: what worked and what didn't

• Ben Horowitz used his "Freaky Friday" technique — swapping executives across the two organisations — placing Feaster in charge of the combined product organisation.
• Cultural proximity mattered: Mercury Interactive (where Feaster came from) shared DNA with Opsware — aggressive, customer-focused, results-driven — reducing friction.
• HP's IT processes created immediate operational hazards: the licence key generator for all Opsware software was flagged for shutdown at day one of the acquisition; source code migration to HP's standard version control system was near-impossible pre-GitHub.
• Feaster's primary job during integration was absorbing and deflecting the bureaucratic mechanisms of a large company from disrupting the acquired team.
• Sales overlays — using existing Opsware specialists as overlays on HP's broad sales force — were the biggest challenge; enabling HP reps to sell a complex product they hadn't built is a slow, imperfect process.
• The business saw 350% revenue growth in the years post-acquisition.

Why HP didn't become a cloud leader

• Innovation in software is a people problem: the "10x developer" and "10x product person" cannot be replicated by headcount.
• Large companies treat people as interchangeable — a manufacturing mindset that destroys software talent.
• Big companies also require complete data before committing; disruptions (Docker, DevOps, Chef/Puppet) arrive before the PowerPoint has 50 slides.
• The natural successors to Opsware — Puppet, Chef — came from outside HP because they had the right people and the willingness to bet early on a trend.
• Retaining the people who understand both the technology and the threat is the only path; once that talent leaves, the compounding advantage is gone.

Timing, courage, and the role of talent

• LoudCloud was the right product at the wrong time; the pivot to Opsware was the right product at the right time — virtualisation drove exactly the scale problem Opsware had been built to solve.
• Sequoia's research concluded that timing is the single most important factor in investment returns.
• Managing timing is the job of exceptional talent: knowing when to pivot, when to bet, when the market is ready.
• Ben Horowitz's thesis: entrepreneurship is the intersection of intellect and courage — seeing the opportunity is not enough without the grit to act on it.
• The Opsware mafia (Andreessen Horowitz, Benchmark, UserMind, Okta, SignalFX, and others) represents the compounding value of people, not the acquisition itself.