The real story behind a potential open-source disaster.

Its CEO walked the plank. Employees are running for the lifeboats. And
an initial public offering has been put on ice. What went wrong at
Linuxcare?

Just about everything, as this exhaustive investigative report by
Sm@rt Reseller reveals. Interviews with more than one dozen current
and former Linuxcare employees — most of whom spoke under the
condition of anonymity — paint a dramatic picture of a company that
was doomed from the moment it left port.

Indeed, sources say Linuxcare was undermined by warring internal
factions, out-of-control spending and a management team hellbent on a
speedy IPO. The result: Linuxcare’s ship is taking on water, with
surviving executives scrambling to keep the company afloat.

Like Red Hat and Caldera, Linuxcare was positioned as a big ship on
the open-source seas. But instead of hawking commoditised software,
Linuxcare focused on higher-margin services and support gigs from its
inception. The 21-month-old company provides services, support and
training for open-source products. Proponents predicted that Linuxcare
would thrive on booming demand for the Linux operating system.

But Linuxcare’s strategy sprung several major leaks, and the fledgling
company has been taking on water in recent weeks. Among the
casualties: Linuxcare’s board has given CEO Fernand Sarrat and other
top management the boot, pink-slipped 25 percent of its workforce and
scrapped its IPO plans.

“Linuxcare’s management screwed up big-time,” says Eric Raymond,
open-source luminary and a member of VA Linux Systems’ board of
directors. “They had a lot of good people at the line level, too. This
was definitely a case of top-down incompetence.”

Other critics question Linuxcare’s strategy. “There are not enough
companies who have enough Linux deployed to justify support contracts
on just Linux,” says Ransom Love, CEO of Caldera.

As Linuxcare’s remaining managers race to plug holes in the company’s
internal infrastructure and business processes, its board is searching
for a new CEO.

The position won’t be easy to fill. “There’s not a lot of executives
in the world with open-source experience,” laments Linuxcare chairman
Ted Schlein, who is expected to look at some corner-office candidates
in the coming week.

Sell the ship?

Hiring a new CEO isn’t Linuxcare’s only option. According to a
well-placed source close to Linuxcare, the services company has
received acquisition offers in the past from Red Hat and hardware
provider VA Linux. “Either company would do well to acquire Linuxcare
right now,” says the source. Schlein declined to comment on the
possibility of a merger, saying only that additional funding is
imminent.

Meanwhile, many remaining Linuxcare execs blame the stock-market
downturn for the canceled IPO. Most execs, however, decline to discuss
the murky circumstances surrounding last month’s departures of CEO
Sarrat and CIO Douglas Nassaur. “The stock-market instability got to
us,” says Linuxcare’s interim CEO Pat Lambs, who usually runs the
firm’s services unit. “[Plus], you really need a CEO to do the [IPO]
road show.”

Part 2: Where’s the cargo?

Still, there were bigger problems.

From the get-go, Linuxcare boasted a swagger that fostered big
thinking and even bigger spending. The company was conceived in August
1998 by a trio of high-tech consultants with a penchant for
open-source software and deep connections within the Linux community.
“Two years ago, the question [regarding Linux] was, ‘Where am I going
to get support? Who can I call? Who can I sue if it breaks?'” says
Linuxcare co-founder and CTO David Sifry. “People loved [the idea of
Linuxcare] … They went gaga over it.”

The feeling was reciprocal. In March 1999, Linuxcare took out a
full-page ad in The Wall Street Journal declaring Linux inventor Linus
Torvalds the next “leader of the free world.”

But for Linuxcare to succeed, it needed a seasoned captain. Enter
Fernand Sarrat, a 23-year IBM veteran with an amiable style and a
solid record of team building. From his May 1999 hire, Sarrat drew
high marks from employees for facilitating teamwork and raising
Linuxcare’s profile. However, Sarrat also floored the IPO gas pedal,
accelerating the half-built company toward an IPO date it couldn’t
handle, according to current and former company insiders.

Simply put, an IPO became Sarrat’s white whale. The smash success of
Red Hat’s IPO last August and VA Linux’s December offering only served
to quell the voices of opposition inside the company.

“There was always a big rush on Fernand’s part to go public … He was
very concerned about looking good to investors,” says one former
employee who lost his job in the recent downsising.

Reached at his home by Sm@rt Reseller, Sarrat declined to comment on
his track record.

Despite all of the hype, Linuxcare never proved that its business
model was seaworthy. Revenue for 1999 was a meager US$1.5 million,
coupled with US$21 million in losses. Sources say Linuxcare didn’t
have proper cost controls in place. For starters, Sarrat set out to
build an internal technology infrastructure to facilitate customer
support, best practices and its nebulous “knowledge center,” which was
later heralded in its S-1 filing (despite being little more than a
simple search engine, according to two former employees).

Linuxcare hired former E*Trade technology chief Douglas Nassaur in
September 1999 to spearhead the tech-support centerpiece, code-named
Sorcerer. Not long after, many Linuxcare techies became confused as to
what exactly the new CIO was trying to accomplish. Much to the San
Francisco office’s chagrin, Nassaur — who was based in Atlanta with
most of his team — began outsourcing big chunks of work at “hundreds
of thousands of dollars” a pop, says a former employee. “It didn’t
make sense, because we had all this great technical expertise in
house,” says the ex-employee.

With the price tag reaching several million dollars (estimates from
several sources range from US$5 million to US$12 million), it became
quite clear the project wasn’t delivering enough bang for the buck. By
April, the project was scaled back and Nassaur was squeezed out. The
former CIO did not return Sm@rt Reseller’s calls seeking an interview.

“Sorcerer is a perfect, textbook case study for the failure of a
project,” says a former Linuxcare engineer, who says the company has
since scrapped most of the multimillion-dollar project. “[It] was
started without any higher-level architecture or engineering reviews
to prove whether it could really be useful for Linuxcare, or whether
it could be developed at all.”

Linuxcare officials declined to put a price tag on the Sorcerer
project but claim the company is still utilising much of its
technology.

Part 3: Extra ballast

Infrastructure wasn’t the only drain on cash. With the full
realisation that a services business needs quality people, Linuxcare
kick-started a hiring binge that more than doubled the size of the
company between September 1999 and April 2000. At its peak, the
company approached 280 employees, bankrolled in large part by a US$32
million round of venture funding, closed in December 1999.

“We hired and expanded pre-IPO in a very, very aggressive manner,”
anticipating that funds would come from an imminent public offering,
said interim CEO Lambs. When that money didn’t come, she said,
Linuxcare was forced to trim the workforce by one-quarter and curtail
spending across the company. “That’s always a very hard thing to do,”
said Lambs, who doesn’t anticipate further reductions in the
workforce.

To many employees, the downsizing was the inevitable result of a
reckless spending binge. “Almost every employee I knew had a
company-paid cell phone … and Palm [Pilot]. While those aren’t huge
costs, I view it as one sign of the out-of-control spending,” said one
former Linuxcare employee.

While management gambled big on an IPO that didn’t materialise, a
tense civil war seethed below Linuxcare’s placid surface, pitting
open-source purists against much of the CIO’s staff. Only now, after
the ousting of Sarrat and Nassaur, has the internal technology jihad
subsided.

In the eye of the hurricane, again, was Nassaur’s Sorcerer project.
The CIO increasingly backed the use of proprietary software in the
tech-support centerpiece. But open-source believers, led by Linuxcare
co-founders Sifry and Arthur Tyde, were irate, according to former
employees.

“It’s simply a question of eating our own dog food. How could we
maintain our integrity if we bypassed open-source products when it
came to running our own business?” said one former employee who was
laid off in the recent downsising.

But as opposition buzzed to a crescendo, the CIO’s team continued to
champion big pieces of proprietary software.

“Doug, especially, didn’t care what software was used,” another
ex-employee said. “At one point (his staff was) thinking about
deploying Lotus Notes … which isn’t the most Linux-friendly
application.”

At least one member of Nassaur’s camp concedes the IT unit shied away
from using some open-source apps, because they weren’t convinced that
the software was capable of running a world-class services
organisation. “When we got here, (the tech infrastructure) was a total
mess. All this spaghetti code patching everything together … It’s
all we saw,” the source said.

E-mail blockade

Sarrat, meanwhile, scrambled to contain the conflict. At one point,
the 49-year-old CEO disallowed e-mail communication between the
open-source founders and the Sorcerer team. “The founders were trying
to exercise control where they had no power,” a source close to Sarrat
said. “They are good, smart, very bright people, but also very young
and headstrong.”

By staking out the middle ground between the two camps, Sarrat was
hoping to prod each side for a compromise. Instead, he ended up
alienating both groups, all but sealing his own demise at Linuxcare,
sources said.

In truth, according to several current and past employees, Sarrat
never fully grabbed the company’s helm from the founders. With Sifry,
Tyde and co-founder David Laduke carrying big reps in the open-source
community and doing much of the company’s technology hiring, many
employees still looked to them instead of Sarrat for approval. Without
the founders’ support, Sarrat was essentially rendered a lame-duck
CEO.

“The board was scared of losing support from the founders … (in
part) because employees gave more credence to the founders than the
CEO,” said the source close to Sarrat. “It was just an impossible
situation.”

Aside from hastening the departures of two high-ranking execs, the
internal technology battles also disillusioned several rank-and-file
Linuxcare employees. “We were supposed to be fighting out in the
marketplace, not at home,” an ex-Linuxcare employee said. “How can we
compete when we can’t even get along?”

Part 4: Plugging the leaks

Linuxcare’s leaders are confident they have buried the religious war
in the past, but they still face a bushel of management problems
spurred by the company’s blistering growth rate over the past eight
months. Its biggest challenges: Beefing up its middle management and
business processes to be more responsive to its employees.

After adding some 150 employees in less than nine months, Linuxcare is
still struggling to incorporate these new hires into the fold and
fully tap their expertise. The processes and managers just aren’t
there. According to one former employee, basic necessities such as
intranet access for remote offices only became available in the past
two months.

Additionally, communications channels throughout the company have been
spotty at best-making the water cooler the most reliable spot to catch
up on company news. For instance, most employees were shocked to learn
of Sarrat’s departure on Web news sites, rather than in an internal
memo. “As an employee of this company, you want to be leveled with …
If things aren’t going as well as planned, we can handle it,” says a
former employee. “This was just another indication that something was
screwed up.”

The lack of communication, coupled with little involvement in
decision-making processes, caused some Linuxcare lackeys to view
management with an “us vs. them” mentality. One engineer, for example,
quit when Linuxcare management reclaimed his rooftop parking space to
give it to a newly hired VP. To the surprise of his co-workers, nobody
in management asked him why he was leaving.

“He was a good engineer … I was shocked,” says a Linuxcare veteran,
who was laid off soon afterward. “I don’t think [Linuxcare’s founders]
want to disrespect or mistreat their employees. However, they failed
to set up the correct processes for hiring, firing [and]
decision-making when [they were in charge].”

While its internal challenges remain formidable, Linuxcare is
sharpening its external focus with the help of a rapidly maturing
Linux market. As a result, Linuxcare’s professional-services revenue
surpassed its desktop-support business for the first time a couple of
months ago, according to Sifry.

“The bad news is behind us. We’re focusing on the future,” says Sifry,
whose zest for open-source software is nearly tangible, even over the
telephone.

In addition, the company still boasts several solid partners and
customers, who say they are unconcerned with Linuxcare’s recent
travails. And company officials say at least a half-dozen more
customers will be announced shortly. “I am not bothered at all by
what’s going on over there,” says Kris Rost, global IS manager at OK
International, which spends about US$36,000 per year on its Linuxcare
contract.

Another positive: The services pioneer boasts high morale among the
many employees who fancy themselves as open-source junkies. For them,
the chance to toil shoulder-to-shoulder with several open-source
pioneers on Linuxcare’s payroll carries considerable cache. “Working
with guys like [Samba author Andrew] Tridgell and [Linux guru Paul]
Mackerras is pretty cool … It’s incredible to see these guys work,”
says an employee with Linuxcare’s Italy office.

Analysts say Linuxcare may yet turn a profit, especially if the
company takes time this year to build out its business in a methodical
manner. “We’re going to [pursue] a much more conservative growth,”
promises Lambs. “My No. 1 job is keeping people focused.”

Linuxcare’s No. 2 job should be to finally build the company’s
infrastructure, a task that was neglected during its ill-fated IPO
scamper. According to a former engineer who left Linuxcare late last
year: “The company has always had a lot of promise … But I think
they would have been much better off without all the [venture] money
… It really drew them away from their original vision.”

Now, to get back on even keel.

A version of this story by Ben Elgin & Steven J. Vaughan-Nichols, was first published in Sm@rt Reseller