They needed it overnight. They wanted the best. In the race to build the first online hardware store, Peter Hunt and Rich Takata put their company in the hands of outright strangers

The Company
Name: Inc.
Founded: Incorporated May 1999; Web site launched early 2000
Location: San Francisco
Cofounders: Chairman and CEO Richard Takata; president and chief operating officer Peter A. Hunt
Employees: 35 full-timers
Mission: Creating an online home-improvement store, magazine, and community for do-it-yourselfers

The Developer
Name: Xuma
Founded: 1998
Location: San Francisco; with offices in New York, Los Angeles, and Las Vegas
Cofounders: CEO Joe Cha; chief technology officer Jamie Lerner
Employees: 250
Mission: Producing built-to-order Web sites for E-businesses

In December 1998, Peter Hunt set out to tackle what should have been a simple, joyous task: building a tree house for his four-year-old son for Christmas. Hunt couldn't wait to start the project. It wasn't just that he welcomed the diversion from his high-powered job as an investment banker. It was more that, ever since he'd been a kid, Hunt had loved working with his hands: making model trains and airplanes, building furniture, fixing things around the house. As an adult he'd dreamed of buying a corner hardware store in some rural New England town, the kind of place with a bell over the door and shelves lined with hinges and screws and doorknobs, where he'd spend his days happily helping customers pick out the right paintbrush or handsaw.

But with two kids and a top job at Montgomery Securities in San Francisco, the lanky, thoughtful Hunt didn't even have time to stop into a local hardware shop, let alone wander through a giant home-improvement warehouse. He assumed he'd save time by buying everything he needed for the tree house -- instructions, lumber, materials, and tools -- online. So he went to the Web and looked for hardware sites. And looked. And looked.

He found nothing except a loose network of like-minded tree-house enthusiasts, many similarly frustrated by their own fruitless online searches for supplies and information. From that experience, Peter Hunt, banker, suddenly figured out how to finally become Peter Hunt, hardware guy.

Hunt's thinking went like this: What if there were lots of little communities out there -- tree-house builders and woodworkers and plumbers and fixer-uppers and even contractors -- all hungry for online hardware and home-improvement advice? And what if somebody could provide it for them in one friendly, convenient location, sort of a virtual corner hardware store?

In early 1999, Hunt, then 35, took a week off to write a business plan for just such a company. But he knew something was missing. He needed a business partner -- a veteran hardware retailer, a real insider. When he asked around, he kept hearing the same name: Rich Takata.

Richard T. Takata, then of Seattle, had been in the hardware business for 24 years. Takata, who'd most recently been president and CEO of Eagle Hardware & Garden Inc., had remained with the 41-store chain after North Carolina­based Lowe's Cos. had bought it for $1.4 billion. (In fact, Hunt's firm had handled the sale.) Although reserved and soft-spoken, Takata, then 49, was hardly averse to risk; in his spare time he occasionally drove race cars. A lifelong do-it-yourselfer, Takata also knew his industry and its customers. Over the years he'd waited on thousands of people, even during store visits when he was the company's CEO. And like Hunt, whose own company had been acquired by North Carolina­based NationsBank, Takata was ready for a change.

Xuma's founders named their Web-development company after an ancient Chinese battle cry.

In April 1999, at the urging of a mutual friend, Hunt and Takata met for dinner to talk about launching a big home-improvement E-commerce site. They discovered they had much in common. Both were quietly intense, articulate, committed. Both were customer-service evangelists, true believers in keeping promises and building long, loyal relationships. Both believed passionately in the Internet's potential for business. And both had high -- some might say almost impossible -- standards for themselves, their companies, and those who worked for or with them. Of the two, Takata was more tactical and analytical, a manager with a keen recall of industry statistics and an almost instinctive understanding of business trends. Hunt, very much the money guy and deal maker, was also more sentimental. (He documents's growth in a scrapbook filled with mementos like copies of the company's incorporation papers and of its early bank deposits, and Polaroid photos of each newly hired employee.)

From the start, the two men agreed they wanted to do more than lead the Web in sales of drill bits and deck stain. They wanted to re-create the old-fashioned corner hardware store of Hunt's dreams and Takata's experience: a place with well-stocked shelves, knowledgeable clerks, lots of how-to information, and most of all, a friendly, collegial, yes-you-can-do-it-yourself atmosphere.

It would, of course, be called

They knew it was a big idea with hefty potential; pulling in even a small fraction of the $400-billion-a-year home-improvement business would yield a fortune. They were amazed that nobody had launched the kind of venture they envisioned, and they knew that before long somebody else certainly would.

What they wanted to do -- build a full-service online hardware store and community -- would cost at least $2.5 million. Takata and Hunt knew they needed to move fast -- and they did. Within three weeks of their first meeting, both men had quit their jobs; put up $250,000 each of their own money; raised about $500,000 more from angel investors, family, friends, and colleagues; and incorporated their business. But for all the partners knew they needed to do, there was still plenty they needed to learn.

Chief among the lessons: just how tough it would be to build a big Web business in a matter of months -- and how much tougher it would get when circumstances forced them to launch the site weeks earlier than planned. They didn't know for sure whether a no-name newcomer like could compete with new and upcoming E-commerce arms from brick-and-mortar brands like Sears (which was already selling parts, tools, and appliances online), Home Depot (which plans to launch a full E-commerce site this summer), and Ace Hardware (which would begin selling merchandise online at late in 1999). In addition, dot-com start-ups were also racing to market. Major projects included, then under development in nearby San Mateo, Calif., and's new tools and hardware store, which would also launch by year's end. And it was entirely likely that some of those in the race would end up as roadkill.

When it came to's technology, Hunt and Takata knew they weren't looking just for a speedy job. Sure, there was no time to waste -- they wanted a beta site before year's end, a quiet launch by March 2000, and a public launch shortly after that. But building their Web site would also be a big, complex, cutting-edge project. They needed transaction processing capability and complete descriptions and images of some 37,000 products, everything from penny nails to power saws to Phillips-head screwdrivers. They would shoot to double their inventory, which is distributed from a warehouse in Kansas, within their first six months online. True to their customer-service mission, Hunt and Takata also wanted, from day one, to offer how-to articles, visitor message boards, animated step-by-step project instructions, a massive glossary of hardware terms, a superb search engine, and live customer service, online and in real time.

That last capability would become, in fact, the real cornerstone, so to speak, of Using interactive windows, customers would be able to chat with service reps in real time -- asking questions about products or about a bill, for instance. That kind of service, which the company would contract out to dedicated staffers at Boston-based eSupportNow, would be what Hunt and Takata believed would ultimately distinguish their company not just from other online hardware stores but from their brick-and-mortar brethren as well. As Takata pointed out later, there weren't many home-improvement stores that were open 24 hours a day.

Although Hunt and Takata knew what they wanted their site to do, they didn't know much about the nuts and bolts required to make it happen. They needed professional help. And they needed it fast.

The high-speed, high-stakes scenario isn't unique to -- or even to the online home-improvement industry. Today almost any new business-to-consumer Internet company must fight for a foothold in an already crowded market. (Witness the proliferation of online pet stores, drugstores, vitamin stores, and toy stores.) Being first online remains a competitive advantage. But there's no point in being first without doing it well. As consumers on the Internet grow more sophisticated they're less willing to tolerate sites that are slow, unreliable, boring, or tough to navigate. And they absolutely won't return to sites that haven't provided stellar customer service. E-commerce sites have grown increasingly complex in reaction to the industry's ever higher standards and well-publicized successes and failures. In many cases, like's, a business simply can't hire its own team to build a site -- even if it could find the right people, it probably couldn't afford to pay them or retain them. So, like, the company opts to stake the future of its business on outside developers -- people the company doesn't know, people who must translate the entrepreneur's dreams and plans into equipment and software and code.

Xuma's approach bridges the gap between standard and optional E-commerce components.

As Takata and Hunt were setting up shop in rented space in San Francisco's financial district, Joe Cha was building his own business just a few blocks away. About a year earlier Cha had been working at his third consulting job. A friend reintroduced him to Jamie Lerner, a consultant Cha had known slightly when both had worked at Andersen Consulting several years earlier. Like Hunt and Takata, Cha and Lerner found themselves thinking along the same lines. They wanted to try something new, and they didn't want to create just another San Francisco Web-development company. Instead their thinking went like this: What if you could apply the same approach to building a Web site that Dell Computer applies to building a computer? What if you could create big, complex, flexible, reliable, customized E-commerce systems in record time simply by not reinventing the wheel for every single project?

So Cha and Lerner founded Xuma. (The name, pronounced "zoo-ma," is an ancient Chinese battle cry that the partners found perfect to describe their army of engineers charging into the E-commerce wars.) They adapted the Dell model: just as Dell combines standard and optional components to rapidly create computers, Xuma combines its standard and optional E-commerce components to quickly build Web sites.

In the venture's first year, the quiet, charming Cha (so charismatic that he was among 10 bachelors featured in a feature on "The Men of Silicon Valley") sold Xuma's services to customers ranging from health-and-beauty-products retailer to home-furnishings site By its second anniversary, in April 2000, Xuma had launched more than 70 Web-based businesses nationwide and employed 250 people in four offices. But back in mid-1999, Xuma hadn't yet tackled anything on the multimillion-dollar scale of

By the summer of 1999, Takata,'s CEO, and Hunt, its chief operating officer, had raised about $6 million in funding: close to $1 million from their own pockets and from family, friends, and angels; and the balance from the first round of venture funding. (A second round early in 2000 would yield an additional $21 million.) And the founders had begun building a staff. Their first hire: vice-president of engineering Steve Finer, who faced the daunting job of actually overseeing the Web site's construction. (See "Chronicles from the Pit," below.)

Finer, then 33, was an enthusiastic, outspoken technologist who, in a previous life, had managed nightclubs in Boston. He knew something about risk: he'd cofounded an Internet start-up that later collapsed and eventually filed for bankruptcy. And he knew something about working hard; he was always either at the office or connected to it by beeper, cell phone, or computer. (Shortly before the launch, when Finer was working 12 to 15 hours a day, he came home one night to find that his lonely dog, Cassius, had disemboweled a sofa cushion.)

After joining in August 1999, Finer faced his first and toughest task: getting his new bosses "to understand that you don't build anything -- whether it's a car or a Web site -- overnight." Especially not something as complex as And in Internet terms, what Hunt and Takata wanted was pretty close to overnight. So Finer immediately ruled out doing the job in-house. Given the tight market for top technology staffers, especially in San Francisco, he knew he couldn't build the talented team he needed to even approach that timetable.

Instead, at his recommendation, looked outside, holding what Cha describes as a "bake-off" for potential developers late in the summer of 1999. Xuma wasn't the oldest or the biggest or the best-known contestant. But Hunt, Takata, and Finer liked what Xuma had cooked up. The decisive factor: speed. Cha, Xuma's CEO, and Lerner, its chairman and chief technology officer, then both 29, promised to do the job faster than anybody else -- within six months. They also promised to build systems and databases that would "scale," or grow quickly without having to be replaced. That's what needed -- and that's why Xuma walked away with a contract worth between $750,000 and $1 million. (Takata says the balance of's launch budget went for interface design, software licensing, equipment, product photography, and related costs.)

By Xuma's standards today -- less than a year later -- the contract is a relatively small one. But at the time it was a huge coup, providing, if all went well, a link in the chain leading to bigger jobs. So Cha took the kind of risk that he would later say no developer should ever take: he went ahead without any built-in contingency plan -- no plan B -- in case of crisis.

True to its own business model, Xuma would build the site using many preexisting components -- a standard credit-card-processing system, for instance. Still, the Xuma team, headed by senior project manager Phil Lew, then 26, knew that building such a complex E-commerce site wouldn't be easy. Xuma anticipated it would spend five to six months, beginning in October 1999, building, testing, debugging, and launching the site. The schedule, though ambitious, seemed entirely possible.

That is, as long as nothing went wrong.

Although and Xuma were both new, fast-growing San Francisco­based start-ups, their cultures were entirely different. Sure, they both hired the best they could find:'s hiring coups included a Home Depot senior vice-president, a top producer from CNet, and several home-improvement authors and writers, while Xuma lured dozens of "rock-star engineers" away from other Web developers. But theirs were very different workplaces. At, a middle-aged artist or writer in a flannel shirt and jeans might sit in meetings with a college-age kid with a nose ring. It was rare for anybody to spend the whole night at work (with the possible exception of Finer, who worked around the clock in the countdown to the launch). In general, it was quiet, especially since most employees worked one or two days at home. (There wasn't enough office space for everyone to be there at the same time.) In contrast to that relative calm, at Xuma nobody had a private office. Engineers racing to meet project deadlines spent days in the big war room known as "the pit," living on trucked-in pizza or Thai food, working elbow to elbow at food-littered tables lined with computers. It was a noisy, messy, overwhelmingly youthful atmosphere. For Lew, it was exhausting, but it was also fun. His team bonded in a way that can come only from eating three meals a day together, working side by side until after midnight, then car-pooling home through unusually silent streets. And that bonding meant that together they felt they could do anything, Lew says.

They would need to. In late fall, when Lew's team was already spending most of its time in the pit simply trying to hit the original March launch date, something did go wrong.

In mid-November, a competitor,, launched earlier than anybody had expected. About the same time, launched its home-improvement store, and Ace began putting online. And funding was beginning to dry up for consumer dot-coms in favor of business-to-business ventures. Takata and Hunt decided they had no choice: they had to move the stealth launch from late March to January 15, and follow that with the public launch a few weeks later. They delivered the bad news to Finer, their liaison with Xuma.

"They told me, 'If we wait till March, we're out of business,' " Finer recalls. "At that point I'm holding my stomach."

Finer reluctantly asked Xuma to shave close to six weeks off the initial launch date. Xuma agreed to try moving it up to January 15. "The trouble with being the vendor is that the customer is always right," Cha says with a sigh. In this case, being right required heroics from Lew's project team. "We had guys here that didn't see their families, that were living here 24/7," Cha says. (See Lew's diary, below.) "We killed ourselves. But we got on-the-job training there. We learned."

What followed was a series of compromises made by both sides. Five days before the new launch date, the Xuma project team begged for an extension, saying they needed the extra testing time to make sure the site worked well. They asked for two more weeks. Takata and Hunt agreed to wait 10 more days. They knew they wouldn't be doing themselves any favors by launching sooner if the site frustrated the very people they wanted to attract.

Meanwhile, Lew was learning that both Hunt and Takata were demanding, detail-oriented, hands-on managers. "I used to say 'Retail is detail,' " Takata says. "Now I say 'E-tail is detail.' " Both Hunt and Takata closely tracked the site's development, sometimes requesting changes that would take days of engineering time to complete. "Or they'd say, 'We need 700 pages [of Web-site content before launch],' " Lew says. "We'd say, 'We can do 200.' "

Then there was the titanic tinkering on the day before the rescheduled launch. In the afternoon of January 24, Hunt decided the site needed another level of search hierarchy, or ways for customers to view products and information. While other team members frantically tested the site, one engineer spent six hours building in the new function, letting it go live around 8 p.m. After viewing the site that evening, Hunt changed his mind. Lew describes it this way: "Peter sees it. He doesn't like it. I say, 'It's exactly what you guys asked for.' He says, 'I want it back like it was this morning.' " Lew asked Finer to intercede; Finer returned with this message from Hunt: "Sorry, but it has to happen. And you have to tell me when it's done." The same staffer spent the next three hours reversing his earlier work, finishing at about 1:30 a.m., just hours before the quiet launch.

And yet Cha, ever the diplomat, doesn't regard as particularly exacting. "All of our customers are very demanding," he says. In fact, he adds, was a relatively easy client because, unlike many enthusiastic dot-com start-ups with ill-defined business plans, from the very beginning Hunt and Takata had clear ideas about what they wanted to accomplish.

On January 25 at 3:30 a.m., Lew drove his entire team home from work, and, at 4:45 a.m., finally slept. had launched -- without fanfare and without any major problems. It also launched without some of the things its founders had wanted. These were the trade-offs: had been photographing about 800 products a day, but even at that rate the company couldn't shoot 37,000 products before the launch. Instead it posted a representative sampling from each category. (Says Hunt, "If you have 72 hammers on the site, do you really need pictures of all 72 from day one?") It also launched with no way of issuing returns to customers' credit-card bills. (Initially, refunds would be made by check.) And it launched with fewer products and less content than Hunt and Takata had wanted. But nothing crashed, and the products advertised were available, poised on the shelves in the Kansas City warehouse.

Surprisingly, Takata rates the launch at about 95% of what he'd hoped for. "One of the lessons I have learned about the Internet space in general is that you can't be a perfectionist," he says. "The Internet is a game of weeks. If you can get your site up four weeks earlier and have a complete customer experience [even without some desired features], I'd say do it."

A month later the public launch went off without any major hitches. Since then,'s traffic has grown steadily, with Xuma continuing to run the site. Hunt and Takata won't release figures except to say that they had more traffic in April than in the entire first quarter. As for the conversion rate -- the percentage of people who actually buy something -- "some days it's 19% or 20%," Hunt says. "And we have days where it's 1%." Meanwhile, the purchasing of big-ticket items has increased: in addition to batteries and lightbulbs, customers are buying bathroom vanities and power tools.

These days the founders are still keeping an eye on the competition, especially that big orange company from Atlanta. "I'd be lying if I told you we don't worry about Home Depot," Hunt says. "But we don't lose any sleep over it," because, he says, he doesn't believe the giant retailer will duplicate the business model or its real-time online customer service.

As for that tree house, Hunt did build it much later. But he ended up creating it from a kit that he got from a brick-and-mortar retailer. Ironically, he got so busy starting up that he didn't have time to build one from scratch.

And in returning to that project, Hunt revisited one big lesson that he and Takata had discovered throughout the building of their business: it's all about making compromises.

Anne Stuart is a senior writer at Inc. Technology.

Chronicles from the Pit

Steve Finer, vice-president of engineering for, and Phil Lew, Xuma's project team leader, each kept diaries for several weeks between the Web site's "stealth launch," in January, and its first major marketing pushes, in early March. Here are some excerpts:

Steve Finer

Thursday, February 17
At 3:45 a.m., we added another Sun Enterprise 4500 server with 16 CPUs. Serious horsepower! As I got off the elevator this morning my coworkers gave me a high five because the new server made the site so much faster.

Wednesday, February 23
Two new applications went into alpha testing. One rotates products featured on the home page. The other is a media-tracking application [which tracks the effectiveness of promotional campaigns].
Unique visitors to the site have doubled since last week. Today the site was accessed from all over the world, including from Taiwan, Slovenia, Thailand, Malaysia, Israel, and Japan.
Actually shaved for the first time in weeks. Wanted to be presentable for a taping we did today with Xuma and the Mark Bunting video crew [for a business video to be shown on United Airlines and TWA].

Thursday, February 24
Biggest challenge: preparing for the March marketing campaign [a newspaper and online ad campaign with coupons]. We need to be able to support the traffic.

Friday, February 25
Added 6,000 new images. Finished quality-assurance process for media tracker and product-feature applications. Also [a New York Times] article gave us a nice boost in traffic and tripled the number of people who went to the customer-support line yesterday.

Monday, February 28
We're all really busy. The engineering team has a lot of projects that need to be completed, including the media tracker and the product-feature device.

Tuesday, February 29
Public launch. It has been a difficult day. All departments are asking for additions to the site. It's a challenge to satisfy all requests and prioritize them properly. Our lack of space [is a problem]. My job would be so much easier if we could hire people to supplement Xuma's activities.

Friday, March 3
We're really focusing on driving traffic. We added a new disk drive to the development server and a new storage device to help manage all our images. We also added more than 40 new how-to articles.
Don Johnson and Cheech Marin filmed their TV show, Nash Bridges, outside our office. We passed out hats to the crew, which they all wore during the filming.

Saturday, March 4
Had one last meeting with Xuma and my staff to make sure we have the right staffing in place to support [Sunday's marketing campaign].

Monday, March 6
The marketing campaign went off without a hitch. We were able to support all the additional traffic. It blew my mind! Daily traffic today was more than double normal, and we had 10 times as many people buying products as we have on a typical day. Next week the campaign will branch out to a larger part of the country. Since everything today went so well, I'm not too worried. But it is still keeping me up at night.

Phil Lew

Thursday, February 24
Development seems to be going better than planned. [ CIO Ken Hite] called me in the morning wanting to track an order number for an order where the money wasn't captured. The problem was with the file from the third-party fulfillment house. Lesson learned: We need to build in more robust error checking. We cannot assume that the third-party fulfillment house will always give us the correct formatted file.
The need to develop a robust process to keep the content and data fresh on the live site is giving me grief. [The process was so slow that when updated many items, it could take days for the updates to take effect.] We are working on another solution to load data but don't know when that will be in place.
The media-tracker application needs to be done by tomorrow (that is, in the hands of QA). Things are going great, but I've been down this road before. I need to keep the pressure on development to make sure that they follow through on our delivery dates.

Friday, February 25
Had our first meeting with [new executive producer] Alice Hill. She had some fantastic ideas about the site direction. I look forward to working with her; she's going to be able to streamline the decision-making process since we will not have to wait on [COO Peter Hunt and CEO Rich Takata] in the future for decisions about where the site is going to be heading.

Sunday, February 27
What I'll remember most about today: talking to Bill [Meehan, Xuma's lead engineer on the CornerHardware project] at midnight on a Sunday night about -- again. We make this site go. Both of us take a lot of pride in that. We are both emotionally attached to this project and want to be the best it can be. That makes it easier to stay up late on Sunday nights to do things for

Monday, February 28
Everyone is very excited about the big day tomorrow [the official launch]. I think we have all done due diligence to get ready for this big day, but until the day comes you never know.

Tuesday, February 29
Crazy, crazy day. In the afternoon, informed us that [there were] 40 additional content pages to be attached to a new front page. At 5 p.m., 6 p.m., 7 p.m., 8 p.m., this content had still not passed the QA check. The new front page will not be able to go up until tomorrow.
Traffic was higher than usual. Two articles [about] came across my desk: one was from CNet and the other from ZDNet.
Didn't get much sleep, since we were at work until 5 a.m.

Sunday, March 5
First thing I did when I woke up today was log on to the Internet via my DSL to check on the site [following that morning's newspaper coupon campaign]. I can see that the orders are already rolling in from Washington and Utah.
By 10:30 a.m., we are already at 15 orders for the day. I called Steve Finer at home (I think I woke him up) and let him know the good news: people are hitting the site and buying things.

Please e-mail your comments to