But screw your courage to the sticking-place, and we’ll not fail.
—William Shakespeare, Macbeth

In my last post, I described how, at Opsware, we would step back from the fray about once a year to go through a strategic review of our situation with the board. In November 2005, our conclusion was to keep on building an independent company, but in early 2007 another review led us to a different conclusion. Our business was strong, but meeting quarterly expectations was difficult and the market wasn’t rewarding us for the growth and leading market position we had achieved. We were at a decision point: Plow even more money into R&D and sales expansion or explore a sale. We decided to explore a sale, but only if we could achieve a top-dollar price for our shareholders.

The dance begins: May 2007

Our February update meetings with the major enterprise IT players had piqued their interest and led to several follow-ups.  The key was to get someone to set an initial price. On May 22, we went back to the Opsware board with an update. A company we’ll call Company 1 had just offered to buy the company for $11/share. That was a 38% premium—good, but not exceptional. Once again, we discussed the opportunities, risks and shareholder value implications of continuing in the business. We debated long and hard: What was our number?

We decided to send a strong message to anyone considering us as a target: We told Company 1 that the Board had rejected its $11/share offer and would not give serious consideration to any new offer below $14/share. For good measure, we took the opportunity to notify the other companies who’d followed up with us (HP and several others) that we’d received an offer but were not interested in discussing any deal below $14. Since that was a 75% premium over our current price, Company 1, HP and everyone else told us they were out. We sent all of them letters requesting the return or destruction of any information we had shared. Those brief dreams of exit seemed to be just that—dreams. We went back to the slog of making yet another quarter.

Then, almost a month later, the CEO of Company 1 came back offering $13.25/share. Yes! We were getting close to our magic number! We reconvened the board for a serious, fact-based discussion of staying the course versus selling in this price range. Our analysis suggested we would have to grow revenues at least 75% annually to exceed $15/share. That seemed impossible—Wall Street consensus was 28%.

In addition to upside opportunities, we discussed the risks of staying standalone. (Note #6, macroeconomic slowdown: We had no idea how real a risk that would turn out to be just months later.)

It was clear: We should exit. We decided to stick to our guns and drive for a deal at $14/share, leveraging competition and urgency to get there. The first step was to get Company 1 to $14, which we did two days later. Then we notified the others, notably HP, that we had an offer at $14. A whirlwind of meetings and calls followed with the various contenders.

While we had discussions with 10 companies, by July 18 it became a two-horse race between Company 1 at $14.05/share and HP at $14.25/share, each acutely aware of the other’s interest. To keep up the pressure, we negotiated detailed agreements with both in parallel right up to the end, with a real prospect of getting to $15 or more. Although exhausted, we were on a massive emotional high—this was going to be a fittingly impressive end to an eight-year odyssey!

Crisis strikes: July 18

Then, with 24 hours to go, we had an inconceivably ugly crisis. In diligence, Company 1’s auditors, Ernst and Young (EY), challenged the way we had accounted for three customer contracts, despite the fact that EY was also our auditor and had ordered this accounting. Unbelievably, two regional offices of EY were disagreeing with each other. Called to arbitrate, the national office sided with Company 1’s regional office and informed us we would have to restate our publicly-filed financials for the past several years! While the issue was entirely technical and the amounts immaterial, the prospect of a restatement was horrifying. Not only had our glorious deal suddenly vaporized, but by chasing it we’d triggered something that might panic investors and cut our stock price in half. In the blink of an eye, our likely outcome had plummeted from $15/share to maybe $4/share. We were devastated.

Company 1 was seriously spooked and asked for a couple of days delay. We knew we’d have to tell HP. We put the best face we could on it—“technical issue, immaterial amounts, no impact on value, etc.”. Hearts in our mouths, we waited for their reaction. To our intense relief, they reacted with concern but didn’t walk. However, our leverage was gone and our deal seemed on the brink of collapse.

Our only hope was to get the three customers to agree to a minor change in their contract language that would enable the original accounting treatment. Theoretically possible, but we were almost out of time, and these were huge corporations. Working through the night in a hot stuffy conference room, we started emailing and calling our executive contacts. Unbelievably, in a testament to the strength of our customer relationships, all three mustered their attorneys and got it done in 24 hours. Restatement avoided! We were back on track—but with only one bidder currently left at the table. Could we still get it done?

Psych test: July 19

Perhaps sensing advantage, HP execs suddenly dropped their offer to $13.75 in a tense 3 pm meeting the next day with Ben and me in their executive offices. We were furious, but we had come this far and we had no other live bidder. Maybe we should just suck it up and accept that EY’s ineptitude had cost us a few hundred million dollars?

Then, thinking on the fly, we realized it wasn’t that simple. How we reacted would be critical, not just to the deal price but to the very survival of the deal itself. Anything less than an unequivocal rejection would signal to them that they were in the driver’s seat. If we were now willing to accept $0.50 less per share, then why not $1 or $2 or $3 less? If there was really no one else at the table, why not drag out the process and wear us down? After all, this was a very expensive deal—why not terminate and restart at a much lower valuation?

We knew HP was the only one currently at the table, but they couldn’t be sure of that. We knew it was time to “screw our courage to the sticking-place”. We informed them that we had scheduled a final board meeting for 6:30 pm and we would not be recommending the deal with HP for anything less than the original price.

With Company 1 still AWOL and no word from HP, we started the board meeting. What the hell were we going to tell the board? Had we killed our deal over 50 cents a share?

We needn’t have worried. By 6:35 pm, my cellphone was lighting up with calls from HP’s M&A guy, worried they were about to lose the deal. We made him wait. When I finally called him back about 7 pm, he told me they were willing to offer $14.05. Ha! Although some board members and fellow execs strongly urged us to take it, we resolved to go back for the final 20 cents as a matter of principle and to preserve the crucial psychology of a hard-won deal. Fifteen minutes later, we had an agreement to sell Opsware to HP for $14.25/share in cash. We were drained, but elated. Our eight-year journey was at an end.

In summary

Although we never built Opsware with the intent of selling it, acquisition ultimately turned out to be the best outcome for our shareholders and our employees. While the company was clearly an attractive target, the significant premium we achieved in the sale was also the result of the company’s ongoing investment in strategic business development and a tightly executed deal process. And in knowing where the exits were.

So you got to let me know
Should I stay or should I go?
—The Clash

To lead you to an overwhelming question….
Oh, do not ask, “What is it?”
Let us go and make our visit.
—T.S. Eliot, The Love Song of J. Alfred Prufrock

Taking the exit

On July 23, 2007, HP announced it was buying Opsware for $1.6 billion in cash. By any measure, it was a very attractive deal for Opsware shareholders. The acquisition price of $14.25 a share represented a 74% premium over the prior six-month average, and a forty-fold return for anyone who had bet on us at our low point in October 2002, when only an intensive investor relations effort had saved us from NASDAQ delisting. Almost five years later, the acquisition multiple of almost 16 times trailing revenue still far exceeds that of any other billion dollar-plus enterprise software acquisition—ever.

The hard work of hundreds of employees contributed to making us the clear high-growth market leader in what was finally an important enterprise software category, but the significant premium we achieved was also the result of a multi-year, strategic business development effort designed to turn a merely good outcome into a truly exceptional one. That business development effort was built around five key principles that you, too, should consider as you build your company:

  1. Always know where the exits are: Take the time to build relationships with potential acquirers. You never know when you may need them.
  2. Step back from the fray occasionally: Review the company’s strategic situation with the board every 12-18 months and evaluate the alternatives with quantitative and qualitative rigor.
  3. In evaluating your alternatives, ask yourself how you would feel if the environment were to change radically: Days before we signed the HP deal, the Dow hit 14,087—its highest level ever. Within months, hedge funds and mortgage companies started imploding and we entered the worst recession since the great depression. Just three months later, a sale at even $8/share would not likely have been achievable.
  4. If you do decide to sell, a meticulously executed competitive process is key to a successful M&A outcome: This SEC proxy statement description of Opsware’s process details discussions with 10 companies, stretching over nine pages. Every one of those contacts was tightly scripted and carefully orchestrated to drive to the ultimate successful outcome.
  5. Creating and managing the acquirer’s psychology is critical: The key is subtly to convince the acquirer that they have no option but to acquire you, but that you have multiple attractive alternatives—and to reinforce that impression in every interaction, no matter how small, until the deal is definitively done.

“Your best exit may be behind you”

As I’ve discussed in previous posts, the best CEOs leverage strategic BD to cover critical flanks not well covered by other functions—helping to navigate crises, driving acquisitions and strategic partnerships, leading international expansion. No great entrepreneur sets out to build a company to be acquired, but as the company grows, it’s important to have options and to understand them well. Perhaps the most important role of a strategic BD exec is proactively to make sure the company always has options and that the CEO and board always know where the exits are. They tell you on an airplane that “your best exit may be behind you”, but that’s not something you ever want to hear in business.

Stepping back from the fray: November 2005

At Opsware, 90% of our board discussions were naturally about the day-by-day work of building a great company: quarterly bookings, wins and losses, product plans and challenges, financials and so on. However, about once a year we would make a point of stepping back from the daily battle for a more strategic review of our situation. As part of this review, we would address more fundamental questions.  For example:

  • How do we feel about our market?
  • How is our competitive position?
  • Who are potential acquirers of the company, and what’s their current state of mind?
  • What are the major opportunities and risks facing us?
  • Bottom line: What’s the likely value of staying the course versus exiting?

Prior to these board discussions, we would make a round of senior visits to HP, BMC, Oracle, EMC and other behemoths to give them an update on the business.  While our ostensible purpose was to explore the potential for a partnership, our real objective was to understand (and pique) their interest in our space and to make sure we’d have an open door if we should ever decide to explore selling.

I’ve linked our board presentation from November 2005 here. As you can see, it’s a pretty thorough analysis. Here’s the executive summary:

In other words, let’s keep marching!

Stepping back from the fray: February 2007

In early 2007, Marc Andreessen, Ben Horowitz and I made another set of visits to the usual suspects in preparation for a new strategic review with the board. We had made major progress since we had seen them last: exceeded $100M in annual revenue, fired up a productive distribution deal with Cisco, added storage automation, expanded internationally. As before, we went through a pretty slide deck that showed a winning company in a strategic category, with a final slide that mused vaguely about partnership. Left hanging in the air at the end of some of the meetings, we could sense the “overwhelming question” in the mind of the big company CEO: “Should we buy these guys before someone else does?” Several of them requested follow-up meetings.

For us, too, the possibility of actually selling the company was of more than theoretical interest this time around. Meeting Wall Street’s expectations was still a challenge every quarter. Our one serious competitor, BladeLogic, was about to go public. Despite five years of strong execution, our own stock price seemed stuck in a narrow band around $8/share due to continued heavy investment in R&D and sales expansion. Strategically, we were now at a critical juncture: Invest even more heavily in new capabilities like monitoring and enhanced support for virtualization—or capitalize on our current position and the strong interest from acquirers. That exit looked pretty tempting.

That said, selling for a typical enterprise M&A premium of 25-30% didn’t feel compelling to us or the board. We hadn’t escaped the jaws of death and worked this hard, for this long, to build a great company, only to sell it for 10 bucks a share. We were resolved: The only way we’d consider selling was for a price much higher than what we felt we could achieve in the standalone case.

Here’s the recommendation from our February 2007 board update:

That led us into a dramatic three-month dance with a series of suitors that would test our negotiating and business skills, as well as our resolve.

Next up: The dance begins

In part I, I talked about how up-front investment in a high-caliber strategic business development function helped to save Loudcloud when an existential crisis hit in 2002.  But the value of strategic business development didn’t end there.  In fact, it was only beginning.

Opsware V1: Single product, single customer

The EDS transaction closed in mid-August, and we started into the fourth quarter of 2002 in much better shape than we’d entered the first quarter.  That said, we still had plenty of challenges.  We were a single-customer, single-product company.  Moreover, the product—software that automated the administration of servers—was powerful but far from ready for broad commercial adoption.  It had been developed solely for internal use by skilled (and somewhat forgiving) Loudcloud system administrators in a controlled environment.  It was difficult to sell and hard to implement, requiring significant process change on the part of the customer’s IT staff.  It wasn’t easily consumable in small chunks, leading to protracted sales cycles and long implementation periods.

Of course, we had 60 great engineers and plenty of cash, and they lost no time starting the work needed to make the software easier for our small but talented sales team to sell.  While they did that, the BD team led the development of what would become a selective but highly productive M&A strategy to turn Opsware into a broad data center automation platform.

Kicking off the M&A strategy

Working closely with the product and sales teams to understand needs and build consensus, we started identifying and prioritizing categories adjacent to our core server automation category and tracking the key players in each.  Together, we developed a set of criteria for acquisitions, including technical factors such as architectural compatibility and APIs as well as non-technical factors like location, team size, culture and financial profile.  As a general theme, we wanted products that fit with our core proposition of data center automation and were compatible with our server product, but were easier to sell and implement, opening the customer’s door for a subsequent server automation sale.  We briefed the board repeatedly to prepare them for specific acquisition proposals to come.  Thus armed for battle, we set forth.

Going into battle

Any good strategy allows for opportunism and, as it turned out, our first acquisition was highly opportunistic and somewhat unusual.  Tangram was a 20-year-old company, publicly-traded on the OTC “pink sheets” exchange, with a market cap of about $10 million.  Their product was IT asset management software, and while their revenues were growing slowly, their 200 enterprise customers were very satisfied with the product.  We seized the moment and picked up $10 million in annual revenue for stock valued at less than $10 million—a bargain.  The deal brought us a complementary, consumable product, 200 new prospects for server automation software and a low-cost base in Cary, North Carolina, that soon became our “offshore” location for support and sustaining engineering.  In short, a good start to our M&A plans, just four months into our new life as a software company.

Our next acquisition, a year later, was much more strategic.  Our customers were beginning to like our server automation product and wanted to know if we could solve their next pain-point: automating the configuration of network devices such as routers and switches.  We identified the four leading players in network configuration management, visited and screened them all, and ran a competitive process to get them competing to sell to us.  (It’s always better to have them trying to convince you to buy than for you to be trying to convince them to sell.)  We got the one we wanted, Rendition Networks, for a combination of cash and stock valued at around $33 million.  In addition to 30 new enterprise customers and a strong engineering team in Redmond, Washington, we would now have an attractive starter product and a combined server-network automation offering that put us further ahead of our only credible single-product competitor, Bladelogic.  Within a year, Opsware Network Automation was generating over 20% of our bookings.  Even better, the Rendition purchase positioned us for a major OEM deal that would further accelerate Opsware’s breakout lead.  I’ll talk about that in the next post.

With servers, network devices and asset management in the bag, storage was the missing piece of the puzzle.  That proved to be harder—there were few storage automation startups to begin with and none that looked like a great fit.  We ultimately did acquire a small startup named Creekpath Systems in 2006 for $10 million in cash, closing their small Colorado HQ and moving a few team members to Redmond.  Unlike our other acquisitions, storage automation proved just as hard to integrate and commercialize as it was to find an acquisition target, and this was the least successful of our purchases.  Notwithstanding that, we now had a unique full data center automation story: servers, network and storage.

Our fourth and final acquisition was in an emerging category called Runbook Automation.  Having automated individual functions such as server provisioning or router configuration, our customers now wanted to automate entire IT workflows across servers, networking and storage, integrating with other systems such as monitoring or ticketing.  Here there were just two viable targets: RealOps in Herndon, Virginia, and iConclude in Bellevue, Washington.  We knew we wanted iConclude, but again a competitive process of negotiating with both companies helped to drive a speedy transaction at a fair price of $54 million in cash and stock in March 2007.  (iConclude’s founder and CEO, Sunny Gupta, now runs Apptio, an Andreessen Horowitz portfolio company).

Although we only made four acquisitions in five years, spending less than $100 million in cash and stock, the M&A strategy played a vital role in transforming Opsware from a single-product player, solely dependent on one giant customer, to a full-suite data center automation market leader with hundreds of enterprise customers.  The acquisitions positioned the company for a major strategic partnership with a technology giant in 2006 and a billion-and-a-half dollar sale to another a year later.

Who you gonna call?

Just like the Loudcloud sale to EDS, Opsware’s M&A strategy relied on many essential ingredients and actors: an aggressive but disciplined management team, outstanding engineering, marketing, sales, legal and finance functions, an acquisition currency and a supportive board.  Again, however, a small effective business development team was there to leverage these assets and drive the process to achieve results while the rest of the company continued to fire on all cylinders.  While we acquired four companies, we evaluated hundreds and gained massive industry insights in the process.

What’s your M&A strategy?  Who’s driving it?

Next up: Strategic Distribution