Then word came, via AllThingsD, that something big was coming from the board of directors. And as AllThingsD first reported (again), HP directors made one of their own, Meg Whitman, the former eBay CEO who had become a director earlier this year, the new CEO. Léo Apotheker resigned, but don’t cry for him, because according to his contract, he made out rather well. Even before it was made official, investors applauded the move, sending HP shares skyward.

Analysts did what they always do, and, well, analyzed. And though it looked more like drama criticism, it’s not as if HP hasn’t known boardroom dramas before. Finally, the deed was done, meaning it was time to hold a conference call, but not before talking first to Kara Swisher of AllThingsD.

I got to talk to Whitman and HP Chairman Ray Lane yesterday, too, but I had to wait until after the conference call. With so many critics screaming that Whitman has no experience running an enterprise hardware company — and let’s be honest, there aren’t that many who do — I asked her to elaborate on the defense, made on the conference call with analysts, that her experience as a buyer of enterprise technology, during her years as CEO at eBay, provided important experience that will help her be an effective CEO at HP. I also asked about Autonomy, the British software firm that HP is in the process of acquiring for $10 billion, and how it will fit within HP; about the company’s plans for cloud services; and about the state of the HP brand amid all the corporate mishegas that has unfolded in the last several months.

AllThingsD: Meg, the main criticism of you, since you’ve been named CEO of HP, is that your primary experience before was at eBay, which is a consumer-facing company. The response on yesterday’s conference call has been that at eBay you were a purchaser of a lot of enterprise technology and that this gives you some important relevant experience. I get the point, but could you elaborate on it a bit? How does having been an enterprise buyer help you be HP’s CEO?

Whitman: What HP needs now more than anything else is management skills, communication skills and a commitment to executional excellence, all of which I know well, and are sort of core competencies from my 35-year career in business. I know technology because I ran a company whose very existence would not have been possible without it, and was a very significant buyer of technology products. And so that brings me a unique buyer’s perspective. But I have not spent 35 years in the enterprise business. Add so what that means is that I will be relying heavily on Dave Donatelli; on Todd Bradley; on the senior executives at HP; and also, frankly, on Ray Lane, who was at Oracle for many years, and EDS, and who knows this space well. So I think what customers will get is that one plus one equals three.

Lane: I agree with that. What we need here, and what we didn’t have before, is operational execution, communication skills, getting the team on the same page and leading them. The CEOs of $130 billion companies are not leading the technology development of those companies. I think Meg can go into any enterprise and visit with any CIO or CEO and do really well. So whether it is the technology side or the sales side, I don’t think anyone is giving her enough credit on those fronts. She can do just fine. And then on top of that she has strong operating executives under her who do know the enterprise business. But right now it is the need for leadership of the people, a focus on executing and operating. I could point back to Lou Gerstner at IBM, or even my own days at Oracle. When I joined Oracle, people thought the board had lost its mind, because I was a consultant at Booz Allen. People scoffed and said ‘How is a consultant going to lead the worldwide sales force at Oracle, a trained wolf pack?’ And somehow I figured it out. And I knew nothing about software, but I learned, and I learned from Larry Ellison, who is one of the best.

I want to talk a bit about Autonomy, and about unstructured data. You made a comment about that when you talked with Kara Swisher of AllThingsD yesterday. Talk to me about where you see Autonomy fitting within HP. Do you still intend to let it be independent? How do you see the alignment shaping up?

Whitman: It’s a big and fast-growing market. Of all the data out there, about 15 percent of it is structured and 85 percent of it is unstructured. And the unstructured data is growing by leaps and bounds. There are not a lot of good software companies that can help companies manage unstructured data and help companies make business decisions based on what they see in that unstructured data. So what we hope to do with Autonomy, and I’m enthusiastic about this acquisition, is take what is fabulous about Autonomy — they have a leading position in the marketplace — and put it through the very powerful HP distribution system. And I think what Mike Lynch is excited about — he is the founder and CEO of Autonomy — is taking this great product and getting it into more people’s hands. And we just need to grow this company as fast as we can; extend our lead and our accumulated experience in this area. So that’s the plan for Autonomy.

Lane: Yeah, I think the synergies are great, and I think it makes a lot of sense. It will make a lot of sense to customers if HP engages them in a dialogue of managing unstructured data.

You don’t think HP paid too much for Autonomy?

Whitman: You know what? It is what it is.

Lane: We wish we could have bought it for cheaper, but it was the market price. People thought we overpaid for 3Par, and you know what? We’re hitting it out of the park.

Is HP still going to be player in cloud services? That was a big commitment that Léo made in March. How far along is that plan?

Lane: Absolutely. The cloud is way ahead of plan. So our cloud services have gone live. So that is absolutely part of the plan, yes.

Meg, a lot of the same people who applauded your selection to HP’s board of directors are criticizing your selection as CEO. Why do you think there’s a disconnect?

Whitman: I don’t know. There’s always people who have different points of view on things. What I have to do — and I said this on the conference call — is lead this company, make it a great company again and fulfill its destiny as the icon of Silicon Valley and of California, and deliver the results. I will have to prove myself by delivering the results. If we’re going to restore the confidence that investors have in us, and that employees have in us, we have to deliver. We have to mean what we say and say what we mean and deliver the results. And that is what I intend to deliver.

Meg, you have a lot of history managing brands. I’m thinking of the job you had managing brands for Procter & Gamble. What’s wrong and what’s right about HP’s brand right now?

Whitman: I think HP is known as the world’s largest provider of information technology, and we are a trusted brand. We are a worldwide brand that touches both consumers and businesses. If you’re an enterprise, we have full suite of solutions. I know that when I bought enterprise hardware and software at eBay, I wanted one person to choke when something went wrong. I wanted one supplier to go to and say ‘Hey, this is not working.’ And so I think we have a fabulous brand in a world where technology is increasingly fundamental. I will say — and Ray would say this as well — I think we need crisper communications with all the constituencies. I think on Aug. 18 we confused people. We didn’t mean to do that, but we did. And so I think we’ve got some work to do around communicating crisply and cleanly about what we’re about — the moves that we’re making — to employees, customers, shareholders and, frankly, to the press.

It all started at 12:47 am PT on April 21 in Amazon’s Elastic Block Storage operation, which is essentially the storage used by Amazon’s EC2 cloud compute service, so EBS and EC2 go hand in hand. During normal scaling activities, a network change was underway. It was performed incorrectly. Not by a machine, but by a human. As Amazon puts it:

The configuration change was to upgrade the capacity of the primary network. During the change, one of the standard steps is to shift traffic off of one of the redundant routers in the primary EBS network to allow the upgrade to happen. The traffic shift was executed incorrectly and rather than routing the traffic to the other router on the primary network, the traffic was routed onto the lower capacity redundant EBS network.

EBS works using a peer-to-peer technology that keeps data in sync across several nodes, and using two networks–one fast, used for normal operation, and one slower, used as a backup when the primary one fails. Each node uses the network to create multiple copies of the data being used as needed, and when one node stops talking to another mid-stream, the first one assumes the second one failed and looks for another to replicate to. This normally happens so fast that humans aren’t even involved.

When the network change didn’t work properly, one group of EBS nodes lost contact with their replicas. When their connection was restored, so many had gone down that when they started up replicating again, the available space ran out. That left several nodes in a bad loop looking over and over again for space on other nodes when there was none. New requests to create new EBS nodes piled up, overwhelming everything else. At 2:40 am Amazon disabled the ability of customers to create new volumes. Once new requests stopped piling up, it seemed Amazon had turned a corner, but those hopes were short-lived.

The EBS volumes kept looking for new nodes to replicate to, putting continued strain on the system. By 11:30 am, technicians had figured out a way to quiet things down without affecting other communications between nodes. Once this was done, 13 percent of EBS volumes were still “stuck.” By noon, attention shifted to finding new capacity for the stuck volumes to replicate to. Not easy. That meant undergoing a time-consuming process of physically moving servers and installing them into the degraded EBS cluster. Naturally, it took a lot longer than expected. The process didn’t start in earnest until 2 am the following morning. By 12:30 pm April 22, only 2.2 percent of affected volumes were still “stuck.”

With the new capacity installed, Amazon started work on letting each node communicate normally again. This had to be done gradually, and work to dial it up just right went on well until the early morning hours of April 23. By 6:15 pm that day, operations were almost back to normal–except for the 2.2 percent of volumes that had remained “stuck.” It turned out they would have to be recovered manually. Their data was backed up to Amazon S3, its general storage service. By noon the next day, all but 1.04 percent of that data was recovered.

A more intensive recovery process was tried, and in the end 0.07 percent of the data involved in the crash could not be recovered, Amazon says.

The company says it is auditing its process for carrying out changes in its network, which is where the problem started, and that it will “increase automation” to prevent a similar mistake from happening again. From that statement I gather that it was a human-caused mistake that was then exacerbated by the way the cloud system was designed to work. Customers who used the affected services, whether or not their services were interrupted, are getting a 10-day credit. The list of other changes it is promising is long and detailed, ranging from having more capacity on hand for use in a recovery to making it easier for customers to access more than one availability zone (those who had done prior to the outage fared better than those who hadn’t) to improvements to its status dashboard.

Finally, Amazon apologized:

Last, but certainly not least, we want to apologize. We know how critical our services are to our customers’ businesses and we will do everything we can to learn from this event and use it to drive improvement across our services. As with any significant operational issue, we will spend many hours over the coming days and weeks improving our understanding of the details of the various parts of this event and determining how to make changes to improve our services and processes.

What’s not clear is the effect not only on Amazon and its reputation, but on the planning of customers who rely on its cloud services or are thinking about using them. A failure as widespread and as widely publicized as this may be forgiven but it won’t be forgotten. Mike Rowan, CTO of RatePoint, a reputation management service for small businesses, indicated on Twitter that he was happy with the billing credit. But time will tell if Amazon loses customers over this.

A failure in the cloud is of course one of the fundamental problems that its critics always point to. Yes, you can save money and time and effort by farming your IT services and infrastructure out to someone else. But when those services crash unexpectedly, you–and scores of others that rely on the same infrastructure–are left to wonder what’s going on and when it’s going to be fixed.

As of now, it seems like Amazon is getting the situation under control, it seems to be getting worse, as other parts of the Amazon service that are tied to EC2 are reporting various failures via Amazon’s status dashboard. Failures are showing up Elastic Beanstalk, and the relational database, and Cloudwatch among others.

Amazon’s status messages are below.

1:41 AM PT We are currently investigating latency and error rates with EBS volumes and connectivity issues reaching EC2 instances in the US-EAST-1 region.

2:18 AM PT We can confirm connectivity errors impacting EC2 instances and increased latencies impacting EBS volumes in multiple availability zones in the US-EAST-1 region. Increased error rates are affecting EBS CreateVolume API calls. We continue to work towards resolution.

2:49 AM PT We are continuing to see connectivity errors impacting EC2 instances, increased latencies impacting EBS volumes in multiple availability zones in the US-EAST-1 region, and increased error rates affecting EBS CreateVolume API calls. We are also experiencing delayed launches for EBS backed EC2 instances in affected availability zones in the US-EAST-1 region. We continue to work towards resolution.

3:20 AM PT Delayed EC2 instance launches and EBS API error rates are recovering. We’re continuing to work towards full resolution.

4:09 AM PT EBS volume latency and API errors have recovered in one of the two impacted Availability Zones in US-EAST-1. We are continuing to work to resolve the issues in the second impacted Availability Zone. The errors, which started at 12:55AM PDT, began recovering at 2:55am PDT

5:02 AM PT Latency has recovered for a portion of the impacted EBS volumes. We are continuing to work to resolve the remaining issues with EBS volume latency and error rates in a single Availability Zone.

Update: Here’s more companies that are affected by the outage, according to status updates on Twitter.
Hootsuite, the cloud-based Twitter client, is down because of the outage too. Here’s what the site looks like right now. Also down is the Hootsuite URL shortener ow.ly.

SCVNGR is reporting that it is down too as a result of the outage.

Discovr, an iPad music app, reported that it went down, but shortly reported its service was restored.

Wildfire, a social media app, reports that it is down.

Livefyre is down.

Here’s an interesting one. CampgroundManager.com, apparently a software-as-service application used to manage campgrounds, say it is down.

A service called Totango, which appears to do something with managing customer relations and subscriptions, had some issues, but moved some things around, and got things mostly working again.

ESchedule, a Canada-based employee scheduling service, reports its service is down.

ZeHosting, a Web host, says it is experiencing slowdowns.

Recorded Future, which bills itself as a “temporal analytics engine” is reporting an outage.

PercentMobile, a mobile analytics firm, say its service is down.

The Cydia Store, which hosts applications available for jailbroken iPhones, reports it is down.

Here’s the latest update from Amazon:

6:09 AM PT EBS API errors and volume latencies in the affected availability zone remain. We are continuing to work towards resolution.

Here’s another update from Amazon

6:59 AM PDT There has been a moderate increase in error rates for CreateVolume. This may impact the launch of new EBS-backed EC2 instances in multiple availability zones in the US-EAST-1 region. Launches of instance store AMIs are currently unaffected. We are continuing to work on resolving this issue.

Another Amazon update:

7:40 AM PDT In addition to the EBS volume latencies, EBS-backed instances in the US-EAST-1 region are failing at a high rate. This is due to a high error rate for creating new volumes in this region.

Updating again at 9:10 AM PDT

Amazon continues to post regular updates on its multi-faceted cloud services outage this morning. The latest update message came in about 15 minutes ago, and is reprinted below.

I’ve heard about three other sites that are affected by the outage. Radarsync, a cloud-based service that updates drivers for Microsoft Windows users tells me via Twitter that its service is down. I’ve also seen that Thrillist is having some troubles sending emails. Venmo, an iPhone-based payment service is also down.

This is an update on Amazon’s Relational Database service.

8:12 AM PDT Despite the continued effort from the team to resolve the issue we have not made any meaningful progress for the affected database instances since the last update. Create and Restore requests for RDS database instances are not succeeding in US-EAST-1 region.

And this is one from the EC2 team.

8:54 AM PDT We’d like to provide additional color on what were working on right now (please note that we always know more and understand issues better after we fully recover and dive deep into the post mortem). A networking event early this morning triggered a large amount of re-mirroring of EBS volumes in US-EAST-1. This re-mirroring created a shortage of capacity in one of the US-EAST-1 Availability Zones, which impacted new EBS volume creation as well as the pace with which we could re-mirror and recover affected EBS volumes. Additionally, one of our internal control planes for EBS has become inundated such that it’s difficult to create new EBS volumes and EBS backed instances. We are working as quickly as possible to add capacity to that one Availability Zone to speed up the re-mirroring, and working to restore the control plane issue. We’re starting to see progress on these efforts, but are not there yet. We will continue to provide updates when we have them.

Update at 10:17 AM PDT: Here’s more companies affected by this outage, as offered by contributors to the comments below: ELog.com, a sort of all purpose-notepad in the cloud is down (link goes to a Google cache).

About.me, which AOL acquired last last year is down, and is currently displaying a message saying “We are currently experiencing an outage.”

ECairn, a social media marketing app says it’s down.

Travelmuse, a vacation-planning site is down, though I can’t find an official update on Twitter confirming that it’s connected to Amazon’s troubles.

Web host and design firm Drupal Gardens has an blog entry on its partial outage.

PeekYou, a search company that specializes information about people, tells me it has experienced some trouble, but has shifted its hosting to compensate.

Gamechanger.io, a service that tracks live baseball scoring stats, is down.

I’m beginning to understand what it feels like to be a radio announcer on a snow day reciting school closures after another! The only thing is, there’s no kids cheering.

Update at 10:39 AM: Yet more communication from Amazon, who says it is making progress.

10:26 AM PDT We have made significant progress in stabilizing the affected EBS control plane service. EC2 API calls that do not involve EBS resources in the affected Availability Zone are now seeing significantly reduced failures and latency and are continuing to recover. We have also brought additional capacity online in the affected Availability Zone and stuck EBS volumes (those that were being remirrored) are beginning to recover. We cannot yet estimate when these volumes will be completely recovered, but we will provide an estimate as soon as we have sufficient data to estimate the recovery. We have all available resources working to restore full service functionality as soon as possible. We will continue to provide updates when we have them.

And another concerning the relational database.

10:35 AM PDT We are making progress on restoring access and IO latencies for affected RDS instances. We recommend that you do not attempt to recover using Reboot or Restore database instance APIs or try to create a new user snapshot for your RDS instance – currently those requests are not being processed.

The move comes hot on the heels of Rackspace’s acquisition of OpenStack specialist Anso Labs, which NewEnterprise reported exclusively last month. Rackspace is calling the new unit Rackspace Cloud Builders, and will offer training and certification, deployment and support to companies that want to build and maintain their own cloud running OpenStack. Jesse Andrews, co-founder of Anso Labs, has the title of director of development in the new business unit.

Rackspace is just one of a batch of companies backing the OpenStack movement; Dell and Citrix are also big supporters. Dell has created an OpenStack installer that can be used to get the software up on a set of servers, and once field tests are completed, it’s expected to be offered to the open-source community.

The news here is for companies that have been looking on jealously at all those taking advantage of the public cloud, but that for whatever reason aren’t willing or able to do so themselves. OpenStack offers a way to easily build a private cloud that, depending on its structure, offers at least some of the advantages, like ease of of setup, without some of the perceived drawbacks, like a loss of control over data.

Meanwhile, the speculation around Rackspace and whether or not it’s going to be acquired continues unabated. Its shares have improved by nearly 20 percent since the first of year. It’s certainly not behaving like a company that expects to be acquired. Earlier this week it disclosed in an SEC filing a new 15-year lease on more than 21,000 square feet of data center space capable of maintaining a maximum critical load of nearly four megawatts of power, ready for occupancy by February. Combining that with the cancellation of a lease on a smaller facility, the deal is worth $88 million. Maybe when CEO Lanham Napier said the company is not for sale, he really meant it.

So, it must be doing something right, right?

Hundreds of game developers yesterday showed up to listen to Mark Skaggs, Zynga’s VP of product, who provided pieces of the formula for building FarmVille and CityVille, two of its most popular games.

Skaggs was very informal, down to his slouchy black blazer, and handed out plenty of free advice about the learning process of accumulating one million players on FarmVille and 14 million on CityVille in the first 30 days.

He also mentioned a few little known facts about the games:

• The idea for developing FarmVille came from one of Zynga’s investors, Bing Gordon, a partner at Kleiner Perkins. One day, he put his feet up on the desk and asked, “Why don’t make you a farm game?”
• Farmville peaked at 32.5 million daily users. If you lined them up holding hands across the U.S., they’d cross 6.36 times from New York to San Francisco.
• Production was so quick on FarmVille, the developers stole the avatars from their other game YoVille.
• Zynga was going to retire the old servers that were running Mafia Wars, but it opted for Amazon’s cloud services, which easily handed the traffic explosion.
• FarmVille was going to launch with only a dozen props, but they missed deadlines and went to market with only nine.
• While things have now changed, moms were the killer audience for FarmVille, leading the staff to call Facebook the “matriarch network.”
• Seemed like forever, but there were only 18 months between FarmVille and CityVille’s launches.
• CityVille missed its launch date. Despite stories in USA Today and other publications, Zynga had to hold it for quality reasons. But after launch, it went smoothly and attracted 14 million daily users in 30 days.

And, other advice:

• Learn from success. Play games and look for common threads. If you see someone offering a gifting interstitial, you probably should mimic it. Be cautious about trying to “change the world of social games.” You might get lucky and be really successful, but the reality is you should look at what others are doing and follow their lead.
• It’s no longer a mom network. CityVille was operational in four countries at launch, and has since added three more. “The Germans play a lot and pay a lot.”
• Speed wins. People are coming to you as a Web experience; make it fast.
• Make it fun. You can’t make up for a boring game with volume. At the end of the day, ask yourself, is it fun? If it’s not, who cares that you have a million seeds for a user to pick from.



]]> http://allthingsd.com/20110301/zyngas-farmville-and-cityville-developer-spills-the-beans-on-what-makes-games-great/feed/ 1 http://allthingsd.com/20110114/google-apps-new-promise-no-more-downtime/ http://allthingsd.com/20110114/google-apps-new-promise-no-more-downtime/#comments Fri, 14 Jan 2011 17:20:55 +0000 Arik Hesseldahl http://newenterprise.allthingsd.com/?p=1817 Google is announcing some changes to its service level agreements for its Google Apps customers today. It would seem routine except for what on its face comes across as an extraordinary promise: No more downtime, not even for maintenance.

The promise comes in a blog entry posted by Matthew Glotzbach, product management director for Google Enterprise. “Unlike most providers, we don’t plan for our users to be down, even when we’re upgrading our services or maintaining our systems,” writes Glotzbach. From now on, all downtime that does occur will be counted and applied toward the customer’s service level agreement. In fact, the entire section of its SLA that covers scheduled downtime is being removed. This includes periods of 10 minutes or less, which under the terms of its old agreement didn’t count as downtime.

Google also released some data about the availability of Gmail, both the consumer and enterprise versions. It says that in 2010 it managed to maintain uptime 99.984 percent of the time. This, Glotzbach says, works out to about seven minutes of downtime per month. Citing data from the Radicati Group, he says that makes Gmail 32 times more reliable than the average on-premise email system and 46 times more reliable than Microsoft Exchange.

Over the years, Google Apps has taken some criticism for downtime issues and for not meeting the level of availability spelled out in its agreement. Shortly after it introduced its premier version in 2007, there were reports of availability problems. And there have been occasional Gmail outages like this one in 2009. As services we use daily migrate to the cloud, downtime seems an unavoidable by-product, as this list of incidents in 2010 from Data Center Knowledge suggests. It may seem like a small thing, but Google is making a strong statement here. It will be interesting to see if any of the other cloud providers respond in kind.

]]> http://allthingsd.com/20110114/google-apps-new-promise-no-more-downtime/feed/ 2 http://allthingsd.com/20100317/the-case-for-the-fat-startup/ http://allthingsd.com/20100317/the-case-for-the-fat-startup/#comments Wed, 17 Mar 2010 19:00:09 +0000 Ben Horowitz http://voices.allthingsd.com/?p=22721 Much has been written and said about the current economic downturn and the resulting lessons on how to run high-technology companies. Quite famously, Sequoia Capital, the premier venture capital firm in Silicon Valley, held a mandatory all-CEO meeting in fall 2008 during which it advised them to “Cut spending. Cut fat. Preserve capital.” (You can see the presentation here.)

The presentation catalyzed a movement. Start-ups everywhere adopted a lean, low-burn, low-investment model. To this day, companies seeking funding at our venture firm, Andreessen Horowitz, proudly proclaim in their pitch decks that they are raising tiny amounts of capital so they can run lean.

On the one hand, it is a fact that capital invested is negatively correlated with returns in the venture capital industry. Pumping too much money into a small start-up is unhealthy for both the company and the investor. On the other hand, Facebook has raised several hundred million dollars and is on track to produce fantastic returns for all of its investors.

So what’s a start-up to do? Much of what has been written and said about lean start-ups makes good sense. However, that advice is often incomplete, and some of the things left unsaid are the least intuitive. In this article, I will articulate some of those things left unsaid in arguing the case for the Fat Start-up.

Here is my central argument. There are only two priorities for a start-up:
Winning the market and not running out of cash. Running lean is not an end. For that matter, neither is running fat. Both are tactics that you use to win the market and not run out of cash before you do so. By making “running lean” an end, you may lose your opportunity to win the market, either because you fail to fund the R&D necessary to find product/market fit or you let a competitor out-execute you in taking the market. Sometimes running fat is the right thing to do.

What the hell do I know?

“Al Pacino couldn’t be no gangsta, DeNiro in ‘Casino’ he no gangsta
Wanna be, wanna see, wan’ get a shovel
dig Tookie up n*&%^!, cause he know gangstas”

–The Game

At this point, some of you are asking yourselves, “What the hell does Ben know? If he were really smart, then he’d know that thin is in.” It turns out that I have some experience in managing a fat start-up through the dot-com implosion of the early 2000s. This chart offers a brief summary of equity market history when I was CEO of Loudcloud and Opsware (click to enlarge):



Note that the Nasdaq index is very highly correlated to the start-up funding environment. During the two years I was CEO of Opsware, the Nasdaq fell 80 percent, far more than it has fallen during the current 2008-10 downturn. So the 2000-02 environment was at least as traumatic as this one for Silicon Valley companies–and arguably much worse.

Here is a brief summary of Loudcloud/Opsware’s fund-raising history during that time:

• September 1999: Loudcloud founded
• November 1999: Loudcloud raises $21 million at a $45 million pre-money valuation (Benchmark Capital is the lead investor)
• January 2000: Loudcloud borrows $45 million from Morgan Stanley (MS)
• June 2000: Loudcloud raises $120M at a $700M pre-money valuation
• March 2001: Loudcloud goes public on Nasdaq, raises $160 million and is valued in the public markets at approximately $480 million. Total funds raised to this point: $346 million.
• August 2002: Loudcloud sells the managed services business to EDS (this was the only actual business we had at the time) for $63.5 million and becomes a software company (and changes its name to Opsware).
• September 2002: Opsware trades for 35 cents per share or approximately a $28 million market cap.
• September 2007: Hewlett-Packard (HPQ) acquires Opsware for $1.6 billion

During this period, Loudcloud/Opsware had over 20 direct competitors. Almost all the competitors from the Loudcloud era went bankrupt, including MFN/SiteSmith, Exodus, LogicTier, Williams Communication, Global Crossing, WorldCom/Digex and Storage Networks. Those that survived got bought with valuations of less than $100 million (e.g., Totality) or still have very low valuations (e.g., Navisite).

How did we do it?

“I had a dream I could buy my way to heaven
When I awoke, I spent that on a necklace”

–Kanye West

So how did we navigate through the great dot-com crash, crush the competition, emerge as the No. 1 company in our space and sell the company to HP for $1.6 billion? Did we “cut spending, cut now, and preserve capital?” Did we make cash preservation our No. 1 priority?

No, we didn’t. To underscore the point, here are Loudcloud’s average monthly cash burn figures for the quarters ending in:

• Apr 2001:  $39 million
• Jul 2001:  $35 million
• Oct 2001:  $29 million
• Jan 2002:  $25 million
• Apr 2002:  $22 million
• Jul 2002:  $19.4 million

As you can see, we were aggressively investing in the business throughout 2001 and 2002. While we did reduce our cash burn, we did not make cash preservation our No. 1 priority. As it was, over the course of the transition from Loudcloud to EDS, we sadly laid off 400 employees and transferred another 150 to EDS. However, we didn’t scrimp and save our way to a $1.6 billion acquisition: Instead, it’s what we chose not to cut that ultimately got us there.

Loudcloud was a Web-hosting business. Today, we’d call it a “cloud services” business, but people weren’t quite ready for the “cloud” in 2001. We supercharged our hosting business with software (called Opsware) that automated our Web-hosting operations. The other cloud services businesses of our day also had software investments. However, as the macroeconomic climate changed, they all “cut deep and cut now.” In the end, they ended up putting their software in maintenance mode and stopped building new features.

As we weighed a decision to make the same deep cuts in our own software R&D efforts (a move advocated by the intelligentsia of the day, as well as nearly every MBA we had working in the company), I faced a hard decision: Cut deep and get to cash flow break-even quickly or continue to invest heavily in software?

In the end, I decided to run fat so that we could continue to invest in the Opsware software. At the end of the day, I realized that much larger companies like IBM (IBM) could hire smart people and train them. But without a lasting technology-based advantage, it would be increasingly hard for us to defeat them and build our customer base despite early wins with Ford (F), Fox Sports, and the U.K. government (to name just three of our early customers).

Running fat meant that I laid off zero software engineers so that we could keep on investing in our technology, find our product/market fit, and build a lasting technological advantage.

Still, we had to reduce costs or we would clearly go bankrupt. With this new view of the world, I decided that rather than divesting our intellectual property, I would divest our business. Now, that may sound logical the way I’ve described it, but consider these facts:

• We were generating $65 million/year from the Web-hosting business.
• We were a publicly traded company with a market capitalization of close to $200 million.
• All of our investors (pubic and private) believed in and invested in the Web-hosting business.
• We had close to 500 employees at the time. Nearly all of them were supporting the Web-hosting business.
• We had no other business. We had software, but we did not have a software product and certainly did not have a software business.

Despite all of this, we sold the Loudcloud hosting business to EDS and became Opsware the software company. It was not clear that this was a good idea at the time. In fact, the market thought it was a terrible idea: Our stock promptly lost 80 percent of its value, putting our market cap at about $28 million. It’s worth pointing out that this was about $40 million less than the cash that we had in the bank.

During the transition, we shrank our payroll from 450 employees to fewer than 100. Even with this massive reduction in expenses, it would take another three quarters to reach cash-flow break-even, a milestone we finally reached in Q2 of 2003.

One could argue–and many did–that we should have cut a lot deeper than we did given that we only had one customer. Although EDS was a very large customer (it generated $20 million/year in revenue), a brand new software company doesn’t need 100 people. We could have taken steps to reach cash-flow break-even immediately (clearly, that might have helped us get above 35 cents per share). In other words, we could have “gone lean” by cutting deep, cutting now, and preserving capital.

But rather than do what seemed obvious, I decided to keep on investing. Here’s why: In an economic boom, cash is great, but not necessarily a meaningful competitive advantage. If every company is well funded, being super-well funded doesn’t help you win. In fact, being super-well funded can actually screw you.

But in a bust (like the one we were in), having a lot of cash can be a huge competitive advantage because you can use that cash to put enormous pressure on your underfunded competitors. And that’s what we did.

We spent aggressively to match our best competitor’s product, feature for feature. And we used our public currency to acquire important adjacent functionality (network, process and storage management) that our competitors did not have and couldn’t acquire because they didn’t have the cash (or the equity).

In doing so, we were able to beat a really high-quality start-up (Bladelogic) that did not have the massive technical and cultural baggage that came from exiting the managed services business. Bladelogic was eventually sold to BMC (BMC) for $800 million. But I’m firmly convinced that had we not spent the money, Bladelogic would have emerged as the No. 1 company in the space and gotten the $1.6 billion exit instead of Opsware.

In the end, by continuing to invest aggressively in our technological advantage despite a hellacious funding environment, we were able to turn a doomed business into a winning one.

That is the very short version of how we won the market during the great tech recession of the early 2000s.

So did we learn?

“Hegel was right when he said that we learn from history that man can never learn anything from history.”

–George Bernard Shaw (1856-1950)

Every start-up is in a furious race against time. The start-up must find the product-market fit that leads to a great business and substantially take the market before running out of cash. As a result, the top two priorities are always to:

1. Find the product that 1,000 enterprise or 50 million consumers want to buy and grab those customers before your competitors do.
2. Raise enough cash and spend it intelligently so that you don’t go broke along the way.

Clearly, you can’t succeed if you don’t achieve both priority No. 1 and priority No. 2. So why is taking the market more important than not running out of cash? Because the only thing worse for an entrepreneur than start-up hell (bankruptcy) is start-up purgatory.

What is start-up purgatory, you ask? Start-up purgatory occurs when you don’t go bankrupt, but you fail to build the No. 1 product in the space. You have enough money with your conservative burn rate to last for many years. You may even be cash-flow positive. However, you have zero chance of becoming a high-growth company. You have zero chance of being anything but a very small technology business (see Navisite). From the entrepreneur’s point of view, this can be worse than start-up hell since you are stuck with the small company.

You recruited all the employees, you raised all the money and you made all the promises. You either see it through or leave–without your good reputation. No one wants to work for an entrepreneur who quits his or her own company. This is start-up purgatory, where you work just as hard, reap none of the rewards, and watch all your best people leave you. It sucks to be you.

The Bottom Line

Spending a little or spending a lot is a means, not an end. Choose the right strategy to win the market or you may end up going straight to purgatory.

As you listen to the virtues of the lean start-up–lightweight sales, light engineering, and so on–keep the following in mind:

• If you are a high-tech start-up, your value is in your intellectual property. Don’t stare at your spreadsheets so long that you get confused about that.
• You cannot save your way to winning the market.
• The best companies can raise money even in this market. If you are one of those, you should consider raising enough to wipe out your competition.

Thin is in, but sometimes you gotta eat.

Ben Horowitz is co-founder and general partner of Andreessen Horowitz. He co-founded Loudcloud, later renamed Opsware Inc., in 1999 and served as CEO of the company before it was acquired in 2007 by Hewlett-Packard. He was most recently vice president and general manager of Hewlett-Packard’s Business Technology Organization Unit.

]]> http://allthingsd.com/20100317/the-case-for-the-fat-startup/feed/ 12 http://allthingsd.com/20091217/palm-posts-loss-ships-783000-smartphones/ http://allthingsd.com/20091217/palm-posts-loss-ships-783000-smartphones/#comments Thu, 17 Dec 2009 21:26:22 +0000 John Paczkowski http://digitaldaily.allthingsd.com/?p=31029

The second-quarter loss Palm reported Thursday afternoon was narrower than the one it reported last year, but still fell far short of what Wall Street had been expecting. The smartphone maker lost 37 cents a share for the period on sales of $302 million. Analysts had been expecting a net loss of 32 cents per share on revenue of $266.2 million.

Palm (PALM) did manage to ship a total of 783,000 smartphone units during the quarter, though, a five percent decrease from last quarter, but a year-over-year increase of 41 percent. That said, the company actually sold only 573,000 units, down 29 percent from the previous quarter and down four percent year-over-year. Seems the launch of the Pixi wasn’t quite as successful as Palm had hoped.

“We are continuing to execute strongly against our long-term strategy with the delivery of Palm Pixi, the new carrier launches completed this quarter, and the upcoming opening of Palm’s full developer program,” said Jon Rubinstein, Palm’s chairman and chief executive officer.

“We’re still in the early stages of a long race,” Rubinstein added, “and we’re energized by the opportunity to compete in this exciting market. We remain confident that Palm’s innovative product design capabilities, integrated cloud services and the differentiated and delightful Palm webOS experience will provide the foundation for our sustained success.”

Once again, Palm did not break out unit sales of the Pre or Pixi in its earnings release, below. At $11.26, Palm shares are down 3.92 percent in after-hours trading.

Palm Reports Q2 FY 2010 Results

SUNNYVALE, Calif.– Palm, Inc. (NASDAQ: PALM) today reported that total revenues in the second quarter of fiscal year 2010, ended Nov. 27, 2009, were $78.1 million. Gross profit was $5.5 million, and gross margin was 7.0 percent. These results include the effects of subscription accounting applied to Palm(R) webOS(TM) products as required by GAAP.(1) In accordance with this methodology, revenues and direct cost of revenues for Palm webOS products (currently Palm Pre(TM) and Palm Pixi(TM) smartphones) are deferred and recognized over the products’ estimated economic lives.

To facilitate comparisons to Palm’s historical results, Palm has included non-GAAP adjusted measures, which exclude the impact of subscription accounting, stock-based compensation and other items detailed later in this release. The company believes this information will help investors better evaluate its current period performance and trends in its business.

Non-GAAP Adjusted Revenues in the second quarter totaled $302.0 million, non-GAAP Adjusted Gross Profit was $77.3 million and non-GAAP Adjusted Gross Margin was 25.6 percent.

“We are continuing to execute strongly against our long-term strategy with the delivery of Palm Pixi, the new carrier launches completed this quarter, and the upcoming opening of Palm’s full developer program,” said Jon Rubinstein, Palm’s chairman and chief executive officer. “We’re still in the early stages of a long race, and we’re energized by the opportunity to compete in this exciting market. We remain confident that Palm’s innovative product design capabilities, integrated cloud services and the differentiated and delightful Palm webOS experience will provide the foundation for our sustained success.”

The company shipped a total of 783,000 smartphone units during the quarter, representing a 5 percent decrease from the first quarter of fiscal year 2010 and a year-over-year increase of 41 percent compared to the second quarter of fiscal year 2009. Smartphone sell-through for the second quarter was 573,000 units, down 29 percent from the first quarter of fiscal year 2010 and down 4 percent year-over-year.

On a GAAP basis, net loss applicable to common stockholders for the second quarter of fiscal year 2010 was $(85.4) million, or $(0.54) per diluted common share. This compares to a net loss applicable to common stockholders for the second quarter of fiscal year 2009 of $(508.6) million or $(4.64) per diluted common share. The company’s second quarter of fiscal year 2009 results included a non-cash charge with a net impact of $396.7 million to the tax provision pertaining to the increase of the valuation allowance for the Company’s U.S. deferred tax assets.

The company’s net loss applicable to common stockholders on a GAAP basis reflects accounting guidance, effective in the first quarter of fiscal year 2010, which requires the anti-dilutive provisions of Palm’s series C preferred shares and related warrants to be treated as derivatives for financial reporting purposes. The fair value of the derivatives were estimated as of the first day of fiscal year 2010 and are marked to market on a quarterly basis, with any change in value reflected in the company’s financial results for the period. The series C derivatives balance was $178.7 million at the end of the second quarter of fiscal year 2010 compared to $235.0 million at the end of the first quarter of fiscal year 2010. This reduction in fair value resulted in a $56.3 million non-cash gain on series C derivatives and was reflected in the company’s second quarter GAAP financial results. With regard to the series C derivatives, any future increases in Palm’s stock price from period to period will be reflected as a non-cash loss on these derivatives in the company’s financial results, and any future decreases will be reflected as a non-cash gain in the company’s financial results.

Non-GAAP Net Loss for the second quarter of fiscal year 2010 was $(59.6) million, or $(0.37) per diluted share. This compares to a non-GAAP Net Loss for the second quarter of fiscal year 2009 of $(80.2) million, or $(0.73) per diluted share.

Earnings before interest, taxes, depreciation and amortization, or EBITDA, for the second quarter of fiscal year 2010 totaled $(70.1) million. EBITDA, adjusted to exclude the impact of subscription accounting, stock-based compensation, net other income (expense), restructuring charges and a gain on series C derivatives, or Adjusted EBITDA, totaled $(48.3) million.

The company’s cash, cash equivalents and short-term investments balance was $590.0 million at the end of the second quarter of fiscal year 2010. This includes net proceeds of approximately $360 million from the company’s public equity offering, which closed on Sept. 23, 2009. Cash from operations for the second quarter of fiscal year 2010 was $16.7 million.

Palm may periodically provide new software features free of charge to customers of its Palm webOS products and currently recognizes Palm webOS product revenues and related standard cost of revenues on a subscription basis based on the applicable product’s estimated economic life, which is currently 24 months. The company records deferred revenues and deferred cost of revenues on its balance sheet, and amortizes them into earnings on a straight-line basis over the estimated economic product life.

Palm announced today that it expects to early adopt two recently released accounting standards related to revenue recognition, Accounting Standards Update (“ASU”) No. 2009-13 and ASU No. 2009-14, effective for its third quarter of fiscal year 2010. These accounting changes will result in a substantial portion of Palm webOS product revenues being recognized upon delivery. The remaining Palm webOS revenues, which are related to future services and deliverables, will be recorded as deferred revenues on the company’s balance sheet, and amortized into earnings on a straight-line basis over the estimated economic product life, which is currently 24 months. Under the new standards, all related cost of revenues will be recognized upon delivery. This change in accounting will reduce the amount of revenues that Palm will defer on its balance sheet but will have no impact on cash flows and does not change how Palm accounts for Palm OS(R) products, like the Centro(TM), or its Treo(TM) line. Consistent with the company’s past practice, Palm will continue to provide non-GAAP, adjusted measures that exclude the impact of deferred revenue accounting, stock-based compensation and other items as appropriate.

]]> http://allthingsd.com/20091217/palm-posts-loss-ships-783000-smartphones/feed/ 0 http://allthingsd.com/20090511/android-on-steroids/ http://allthingsd.com/20090511/android-on-steroids/#comments Mon, 11 May 2009 17:51:09 +0000 John Paczkowski http://digitaldaily.allthingsd.com/?p=17359 2009 is going to be a banner year for Google’s Android mobile operating system. Strategy Analytics estimates shipments of handsets running the OS will grow 900 percent this year as more vendors adopt it. At that rate, it will far outpace the growth of Apple’s iPhone, whose shipments the company expects to increase just 79 percent in 2009.

“Android has fast been winning healthy support among operators, vendors and developers,” said Strategy Analytics director Neil Mawston. “A relatively low-cost licensing model, its semi-open-source structure and Google’s support for cloud services have encouraged companies such as HTC, Motorola, Samsung, T Mobile, Vodafone and others to support the Android operating system. Android is now in a good position to become a top-tier player in smartphones over the next two to three years.”

If that comes it pass, it’s bad news for Palm (PALM) whose Pre handset and WebOS will be facing off against a powerful mobile triumvirate: Apple (AAPL), Research in Motion (RIMM) and Google (GOOG).

]]> http://allthingsd.com/20090511/android-on-steroids/feed/ 3