__ Find customers, take orders, lug deliveries, eke profits - selling groceries online is a crazy idea that'll make somebody millions. __
Johnnie Bentley is in a hurry and there's no place to park, so he saves a little time by steering his delivery truck onto the sidewalk and cutting the engine. We jump out on Beacon Hill's West Cedar Street and run to the back of the vehicle, where Bentley pops the lock, swings a bar, and pulls open the double doors. If a meter maid catches us, he's out 40 bucks, so we hustle.
We unload five full plastic grocery totes, ordered off the Web from a Boston-based service called Hannaford's HomeRuns. Bentley, a tall, muscular guy with close-cropped hair, throws on a black HomeRuns windbreaker and buzzes the customer, who lets us in after uttering five dreaded words over a crackly intercom: "It's on the top floor." We lug the stuff up four flights, Bentley running ahead in a blur of boot soles. Moments later, inside a modest, middle-income apartment, he strides past a tiny young Filipina mother and heads for the kitchen, polling customer satisfaction as he goes.
"How's the service working for you so far?" he asks. The woman smiles, nods, and softly replies that it's working fine.
In the kitchen there's a baby squirming in a high chair - right in the middle of the action. "He's a big guy," Bentley says as he piles cans and boxes on the counter and loads the fridge. A friendly golden retriever pads over for a sniff. Mom comes in and signs an invoice, and Bentley flies out the door. We're back downstairs in six minutes flat - just in time to see a parking-enforcement three-wheeler heading our way with grumpy intent. We hop in and speed off, scanning a city map to locate our next stop. Bentley's still in a hurry: He has to deliver 64 totes to 13 homes and businesses in downtown Boston today, and he's only about halfway done.
It's the busiest week ever in HomeRuns' three-year existence, and Bentley is pleased that the last customer was a typical homemaker. Still, he knows the company is far from an unqualified success. "Right now we're a luxury for rich people," he says as he navigates Beacon Hill's narrow streets. "We need to be a necessity."
Bentley is toiling in the grocery industry's consumer-direct trenches, an ecommerce realm whose big moment may finally be here, thanks to the mass-marketing power of the Web. Or so it is hoped. Though the idea sounds tantalizing - using technology to deliver groceries to homes, just like it's done with books, CDs, flowers, and pharmaceuticals - too often the reality has been a morass of tangled logistics, traffic jams, cracked eggs, and maddeningly slow adoption rates.
Several early contenders, now-defunct companies like Home Shopping Alternatives and Shopper's Express, ran out of money in the mid-'90s trying to make a go of consumer-direct. Having seen the business model rise and sputter before, some grocery-industry watchers are skeptical that enough consumers, Web or no Web, will ever allow surrogates to pick their food.
But then, this market could be bigger than Amazon.com, bigger than CDnow, bigger than Dell. Americans buy $500 billion worth of food every year, compared to the $30 billion or so they spend on books. They visit supermarkets an average of twice a week, spending 9 percent of their weekly income. After a slow start, the consumer-direct grocery industry has started to grab some of the action, hauling in an estimated $456 million in gross revenues last year.
That's only about half what online booksellers make, but industry projections say the number of consumer-direct grocery shoppers will increase dramatically in the next few years. This year, according to Forrester Research (whose market analysis of the sector tends to be conservative), 1.1 million households will buy food online, with 235,000 of those buying groceries from companies like HomeRuns and Peapod. The others will purchase specialty-food items, like fruitcake or Vermont maple syrup. By 2003, 3.3 million households are expected to receive regular deliveries of cybergroceries. A typical client household is a two-salary family with a median income of $57,230. Consumer-direct shoppers shell out roughly $100 per order, far surpassing the average order from online merchants like Amazon.com. Some companies, like Streamline, based in Westwood, Massachusetts, say their customers are ordering an average of 46 weeks out of the year. What other ecommerce players can even dream of pulling in $4,600 per customer, year in and year out?
"What we like about grocery shopping is that it has to happen, and it has to happen every week," says Jeffrey Hulton, senior VP of HighPoint Systems, a Cambridge, Massachusetts-based company that is trying to develop new information appliances - like a scanner-equipped PalmPilot - to make ordering groceries from home more convenient. Major consumer-direct grocers such as Peapod, Streamline, ShopLink, and HomeRuns are all building a base in high-density areas, with a special emphasis (at first) on capturing the high-end consumers who help underwrite the surprisingly slender profit margins of traditional brick-and-mortar grocery stores.
The online grocers aren't making money yet - typical of ecommerce, all are operating at a loss, in anticipation of future IPOs and profits - but they believe that the success of Amazon.com has redefined what's possible. After all, in 1996 and 1997, consumers were just getting used to buying books on the Web. Purchasing bread and milk online was still outside their comfort zone.
As that comfort zone expands, the numbers could get very big very quickly. Andersen Consulting predicts that by 2007, at least 8 percent of US grocery sales will occur online. And if that happens, it creates a new $60 billion ecommerce stream.
That $60 billion could be some of the hardest-won money the Internet economy has ever seen. Any ecommerce niche entails difficult logistics; with consumer-direct, it's trickier. You're not just tossing a book in a box and shipping it. You're procuring, say, a side of beef from a reputable supplier, cutting it up using equipment and procedures that adhere to USDA standards, packaging the cuts, storing them properly, selecting a package to fill an order, and keeping it cold as it travels from shelf to tote to truck to home.
It's a headache, which is why the rapid advance of consumer-direct grocery delivery is also leaving behind a few carcasses. But the effort is worth it - there's a fortune waiting for whoever makes it fly.
The most established player to date - Peapod, with an annual gross of $70 million - covers only 5 percent of the country, with headquarters in Skokie, Illinois, and a presence in New York, Boston, San Francisco, and five other markets. Streamline, HomeRuns, and ShopLink, three newer rivals with national aspirations, all orbit Boston, though Streamline is opening a warehouse to serve the Washington, DC, area in advance of a planned IPO in late spring.
Of all American cities, Boston is the most active proving ground for consumer-direct - its population density, wealthy suburbs, and high Internet penetration make it the perfect setting. And in Boston, no one is closer to the front lines than Tom Furber, a plaid-clad, soft-spoken guy - his epithet of choice is "jeepers" - who looks like he should be running the youth ministry at a suburban church. Furber is the president of HomeRuns, a risky, multimillion-dollar experiment launched in 1995 by Hannaford Bros., a $3 billion supermarket concern based in Portland, Maine. HomeRuns is the only established outfit owned by an existing grocery chain. (The others are stand-alone start-ups, supplied by wholesalers or partnered with retailers.)
Some of Hannaford's old guard have been skeptical of the ecommerce fray, and since HomeRuns launched, Wall Street has fretted about Hannaford's investment in a money-losing division. At the moment, Hannaford is evaluating whether consumer-direct is worth all the trouble and is looking for "strategic partners" to share the risk. Insiders say this move may mean parting with a majority stake in the start-up.
For his part, Furber has always been passionate about consumer-direct. Back in 1995, Shopper's Express approached Hannaford to discuss a possible consumer-direct partnership. Furber argued that Hannaford could pull off home delivery itself. "I didn't like the economics of doing it with a partner," he says, "and I didn't like the lack of control."
Ironically, Furber's chief objection to existing consumer-direct companies was that they took an old-fashioned approach: filling orders off the shelves of a traditional supermarket. Stores aren't designed for order-filling efficiency; they're designed to sell food to people in the aisles. Furber proposed running the operation straight out of a warehouse.
Hannaford's old-timers grumbled, arguing that the company would lose money chasing a niche market, but in the summer of 1995 Furber got the go-ahead for a trial project, Pilot Foods. "We wanted to play with the idea on a skunk-works basis," says Paul Fritzson, a Hannaford VP who supports the project.
For the trial run, Pilot Foods served Hannaford employees in the Portland area, who could place orders by phone, fax, set-top box, PC, or Newton notepad. Participation was voluntary; about 50 employees wound up ordering regularly.
In the fall, Hannaford's board of directors convened to decide the project's fate. "The results from Pilot Foods said that there could be a considerable profit," recalls Fritzson. "We said, 'Why wouldn't consumers want to shop this way?' A groundswell of enthusiasm overtook the skepticism." Furber got the green light to start his business in Boston, where it wouldn't cannibalize Hannaford's core business in Maine. By late November he was up and running with an office in Woburn and a spirited new name: HomeRuns.
Customers often rave about the convenience of consumer-direct - no more checkout lines! - but they can be unforgiving when the experience doesn't match their expectations. Nancy Zerbey, a freelance editor in Wayland, Massachusetts, won three months of free deliveries from ShopLink at a school charity auction (the flat delivery fee is $25 per month, another drawback for some) along with $25 in free groceries. But one day the delivery truck knocked down her mailbox. Another time her entire order was sent to a neighbor across the street. "They're trying hard, and it's a good idea," she says. "They just don't have it down yet."
Moreover, while in surveys consumers claim to hate the time they spend cruising supermarket linoleum, it's a deeply ingrained ritual, which some supermarket-industry stalwarts find reassuring.
"Stand in a grocery store in the suburbs or a rural area," says a marketing executive at a major supermarket chain who asked not to be identified. "Some are just hearing about the Internet now. If you look at the average consumer, you'll see how unsophisticated they are."
Tim DeMello, Streamline's founder and CEO, likes that kind of talk. "We want skepticism among traditional supermarkets," he says. "All the skeptics are going to stay out of the business. It's a big opportunity for us."
The consumer-direct strategy involves plundering the high end - aiming for already-wired customers with discretionary cash and upscale tastes - and widening the base over time.
"We'll go from the best customers to the next echelon down," says John Icke, CEO and president of ShopLink. "That leaves the supermarkets with the lowest-margin customers."
Meanwhile, the newer operations are already starting to chisel away at consumer reservations.
"I used to have to walk to the Star Market, take a cab home, and then carry the bags up to my unit," says Penny Cherubino, a HomeRuns customer who lives in Boston's Back Bay, a densely populated neighborhood of four- and five-story brownstones. "It was a major undertaking, and it would eat up half an afternoon on Saturday. The quality of the food is excellent, and I've recruited other people in my building to join the service."
Cherubino uses her Apple PowerBook to assemble her order every Monday, which takes about 45 minutes. She says the time required isn't an issue, though HomeRuns and others are working on ways to speed up the process.
The major services produce either Web sites or CD-ROMs that serve as product catalogs - usually offering about half as many items as you'd find in a typical grocery store. Cherubino has a master list of items she buys regularly, and she browses the site or uses a printed catalog to add impulse buys. "I do 90 to 95 percent of my shopping this way," she says, "and the rest are things like fresh fish or fruit."
Consumer-direct moguls believe that once more shoppers like Cherubino are hooked, they can start thinking in the anything-is-possible terms of Amazon.com founder Jeff Bezos. It's not just about food, they'll tell you, it's about everything.
DeMello, echoing other ecommercialists, says consumer-direct is a chance to "seize" the last hundred feet to the home, to control a pipeline of atoms (not bits) that leads directly to the consumer's pantry or fridge.
To that end, Streamline already offers perks that Amazon.com and Dell can only dream about, since those companies rely on FedEx and UPS as their distribution channels. Streamline will, for instance, bring you a Blockbuster video and pick it up a week later, or grab a dry-cleaning order from your garage - the door will be opened via a special keypad. And the consumer-direct companies don't intend to stop there.
"Once you've established that you can replenish food products - and you have that kind of trusted relationship with your customers - you can start to latch on to a much broader range of products and services," says John Icke of ShopLink. "Down the road, I can see things like flowers, paint and home do-it-yourself materials, garden supplies, beer and wine, and even CDs and books. Our drivers replace lightbulbs or unpack groceries for elderly people.
"You can build on that type of relationship," he says. "It's not like your FedEx guy handing you a box at the front door."
In early 1996, HomeRuns started small by mailing a brochure to 50 households within a two-block area - its infrastructure couldn't support more than a handful of users, and Furber felt that he still didn't understand everything that could possibly go wrong.
One by one, people signed up. "We had a whiteboard, and our customer-order reps would write down the orders we'd received the day before," recalls Ray Giandrea, HomeRuns' VP of information technology. "Some mornings there would be one order, some mornings none, and sometimes we'd celebrate - because we'd gotten two orders."
Every night at midnight, the cutoff point for next-day delivery, a HomeRuns employee would sit down with the handful of addresses and manually plan the trucks' routes. They'd email the orders to a Hannaford supermarket in Lowell, Massachusetts, where "pick sheets" would be printed out and orders packed.
By the fall of 1996, HomeRuns was delivering several hundred orders a week, and it had become nearly impossible to pick them all from the supermarket in Lowell. On Labor Day, HomeRuns moved to a new facility in Newton, another Boston suburb, and started building its army of order takers, pickers, baggers, and drivers.
Order volume slowly grew. Soon, HomeRuns was receiving enough orders for meat that it could purchase and cut its own sides. The company moved away from proprietary ordering software and launched a Web-based system, but the vast majority of clients were still using phone and fax.
On April 1, 1997, HomeRuns got proof that it was making inroads - along with an insurmountable challenge - when the April Fools' Day blizzard dumped 27 inches of snow on the Boston area. The previous night, HomeRuns had seen its business spike upward as customers ordered supplies instead of dealing with mobbed supermarkets. Here was an opportunity for HomeRuns to win the hearts of customers in a crisis and truly prove its value.
That morning, Furber drove to the office on snow-buried roads. "I was basically plowing the road with my Explorer," he says. When he arrived at 5:30 a.m., he discovered that several employees had spent the night there, concerned that they wouldn't be able to get to work in the morning. Others had walked four or five miles through the snow.
In the kind of bonding experience that's essential to every start-up, the skeleton crew began packing groceries. But they soon realized that even if they could fill the landslide of orders, there was no way they could deliver them - snow was still falling as the sun came up. It was the first time that the company would be unable to deliver.
When Furber and his crew began calling households to apologize, they found customers to be pleasantly surprised. Most Boston residents, stuck at home, were amazed that anyone had shown up for work to even try delivering the groceries. HomeRuns' promise to deliver the next day proved a miscalculation. The service had received so many orders - the influx continued on the day of the blizzard, too - that it took days to catch up.
"We worked 16-hour shifts trying, though," says VP of marketing and sales Alison Berglund.
I'm holding a Palm device in my left hand and aiming it at a box of Raisin Bran. I click a button on the device and a red laser shoots out the front, scanning the UPC code on the cereal box. What looks like a shopping list appears onscreen, with "20 oz Raisin Bran" appended to the list. A pulldown menu lets me adjust the quantity - one box, two, or three.
A litany of problems can crop up when ordering groceries via PC, and Jonathan Green, CEO of HighPoint, reels them off as I zap other UPC-emblazoned items in his Cambridge office. He argues that what's really holding the industry back is that it hasn't given shoppers the right device for placing orders. George Colony, founder of Forrester Research, has long envisioned a garbage can that scans UPC codes as you throw away empty packages, automatically recording how to restock your cupboard, and in February, a UK firm called Electrolux announced a Net-connected refrigerator with a scanner - and a 13-inch screen - built into the door.
At present, four of the major online grocery companies provide customers with a Web interface or proprietary software that helps them browse through available items and assemble their weekly shopping lists. Given the state of the Web, using the proprietary software is faster for consumers than clicking on items at a Web site - but even then, some complain that it's not much quicker than going to the store.
"People order less frequently because using a PC is cumbersome," Green says. "Also, with the PC, you have to be connected to the Net every time you want to add a single item to your order. Boot up, connect, search for 'milk,' find it, and add it. Using a PC is slow. Our research shows that it takes 47 minutes to assemble an order on a computer - compared to 45 minutes that a consumer spends in a traditional store. We think we can get the process down to 9 minutes."
He plans to shave off those 38 minutes by writing software for the scanner-equipped Palm, which was developed by Palm Computing and Symbol Technologies, and by porting HighPoint's application to a scanning pen that Symbol developed with Cross. The Palm I used was developed for warehouse and inventory-control applications, not especially for grocery ordering. But HighPoint has embarked on a project to develop its own device - similar to the Palm, but with better display capabilities - in conjunction with the industrial-design firm Ideo.
The notion is not only to make online grocery ordering easier and faster but also to enable consumers without PCs to order their groceries electronically. Green says that a Palm-like device, equipped with a modem for keeping track of what you've run out of, uploading orders, and downloading pricing and delivery information, would cost about $550 today. But he expects the price to drop to about $300 by year-end.
The obvious question is, Who will pay for a dedicated grocery-ordering device? Green has a few theories. Consumer-direct companies could rent them to customers for a few dollars a month or provide them free to their best clients. Some customers might want to buy the devices outright to make shopping easier. Maybe consumer-goods manufacturers like General Foods or Procter & Gamble would cover part of the price in exchange for onscreen advertisements. Imagine scanning an empty 2-liter bottle of Pepsi and getting an instant message from Coca-Cola offering you 25 cents off its product if you switch.
While Tesco, a UK-based supermarket chain, has scheduled a trial of the Palm device, ShopLink is the only consumer-direct company in the US with plans to test it. HomeRuns' Furber says the appliances are simply too expensive: "We're waiting for the cost to come down, since we're reluctant to charge the consumer for using them." Streamline's DeMello suggests customers would rather order in one session than spread it over the week by scanning containers as they're discarded.
But Green is steadfast. He believes his firm will provide the tool these companies need to be profitable. "Consumer-direct won't hit its numbers, because they're limited by the awkwardness of placing an order by PC and the expense of taking an order by phone or fax," he says, brandishing the Cross CyberPen. "They need something simple that can be adopted by even the technophobes."
By late April 1997, HomeRuns had recovered from the blizzard and assembled a crisis-management team to better cope with future storms. On April 23, the company dropped 40,000 brochures in the mail, confident that it could handle a new wave of customers.
The next day at 11:20 a.m., Alison Berglund heard a rumbling sound while she was in a staff meeting in the Newton offices. "It felt like an airplane was flying overhead - about 50 feet above us," she recalls.
A few seconds later, the building began to shimmy, and blocks of concrete came crashing through the wall closest to HomeRuns' data center. No one knew what was happening. A few people hit the deck, covering their heads. Others dashed out of the building into the parking lot.
When the dust settled, the strange reality became clear: A demolition crew working on an adjoining building had inadvertently bashed HomeRuns' HQ, causing a partial collapse. Miraculously, no one was hurt, and somehow a few of HomeRuns' systems survived. The digital telephone switches had been knocked out, though, which instantly brought down HomeRuns' Web site, and the system that supported order fulfillment was disabled. Even worse, the city of Newton condemned the building. Employees were allowed just a few minutes to collect their personal belongings. IT head Ray Giandrea - thinking he might have to use rented computers to get the business running again - went back in, grabbed his jacket, put it on, and smuggled out HomeRuns' backup tapes.
Furber put on a brave face for his employees and the media. He told The Boston Globe that Hannaford wouldn't use the accident as an excuse to shutter the money-losing service. In reality, that scenario wasn't far from Hannaford executives' minds.
"We thought about it in the depression of having the thing crash - no doubt about it," admits Fritzson, the Hannaford VP to whom Furber reports. "We took a step back and said, 'We're losing a pile of money.'"
Furber knew his business hung in the balance. "I remember being pretty discouraged," he says. "I was wildly thankful that no one was hurt, and angry that it occurred. I also had huge concerns about the customer impact." But he had no intention of giving in. He intensely lobbied the city of Newton to let his employees into the condemned building to retrieve paperwork and equipment. Three days after the collapse, he won: A few employees would be allowed entry for half an hour.
"We hired a computer-moving firm," recalls Giandrea. "They sent us six burly guys, and in 30 minutes we managed to bring most of our equipment out. We took some of it to the corporate office in Maine and some of it to Manchester, New Hampshire, where one of our vendors lent us some space." Before rebooting the machines, he adds, "we bought a Shop-Vac, opened up the computer cases, and sucked out all the cement dust."
Furber kept his pickers and drivers on the payroll throughout May. Since there were no orders to assemble or routes to drive, the rank and file spent their days working at Hannaford stores, cleaning up city parks, and volunteering at food banks. The management team rented a trailer, worked with laptops, and kept the call center staffed. "It was important to answer the phones," explains Berglund. "We needed to build confidence that we were coming back."
It would take six weeks for the HomeRuns trucks to get back on the streets. In the meantime, customers had to reacquaint themselves with their neighborhood supermarket. "All of a sudden we took away the new behavior that people were experimenting with," says Berglund, "and they had no alternative but to go back to the old way."
Johnnie Bentley and I are covering Route 32, one of 37 delivery runs carrying about 500 orders among them. It's been a great week for HomeRuns - more important, the company is growing again.
We drop off a couple of totes at a senior citizens' complex in Roxbury, where a one-legged man in a wheelchair turns down our offer to stock his refrigerator and cupboards. "I have fun putting them away," he says, opening the door to a perfectly regimented freezer that would make Martha Stewart proud. A father on River Street in Beacon Hill asks us to be quiet as we unload a dozen bottles of Poland Spring - his baby is sleeping. He gives us 12 empties (he'll be credited for the deposit on the bottles) and we slip silently out the door. On Brimmer Street, we make a large delivery to a Catholic school, where we wind up unpacking part of the order on the ground floor, some on the third, and the rest on the fourth - which Bentley says is against the company's rules.
He doesn't mention that, of course, until we're back in the truck. During the delivery, we smile and happily cart our totes up and down on the elevator, even waiting patiently as a school administrator searches at great length for our check. We don't mention that the truck is blocking Brimmer Street's single lane of traffic or that we're running late. You can't afford to alienate a single customer.
Later in the day, when we've replenished a kitchen in the Back Bay, the customer tries to slip Bentley a five. He refuses, saying that HomeRuns' delivery professionals can't accept tips.
"Just tell your friends about us," Bentley says nobly, sounding a bit like Batman. A moment later, we're back in the truck, racing the clock, trying to prove that you can make money doing this.
The year 1997 continued to be a tumultuous one for HomeRuns and its parent company. After reopening in Newton on June 4, the company had to pick up and move to a new location five weeks later - the landlord had canceled the lease. The company moved to a 120,000-square-foot brick warehouse in Somerville, just across the Charles River from Boston. By mid-November, order volume was back at the level it had reached before the building collapse. Even better, 70 percent of orders had migrated to the Web after HomeRuns began charging a $5 fee for phone and fax orders, which helped cut costs. In addition, delivery hours had been expanded an hour, running from 11 a.m. to 8 p.m.
The workers in HomeRuns' new warehouse were now using wrist-mounted computers from Symbol Technologies, equipped with laser scanners fastened with Velcro to a fingertip, to improve picking speed and order accuracy.
But making progress didn't seem to be enough. When he released HomeRuns' third-quarter earnings report last year, Hannaford's CEO announced that he was seeking outside investors for the customer-direct service. The project had been too much of a drain on the company, accounting for about 12 cents' worth of losses per share in 1998. And while weekly orders were approaching 2,000, the service needed to hit 8,000 to break even.
Competitors hinted that Hannaford was desperate to unload as much as 90 percent of HomeRuns, but Hannaford execs said it was unlikely they'd sell that much.
Furber, true to character, didn't seem downcast about Hannaford's plans to find partners. Instead, he looked at it as an intriguing paradox: How can an old-fashioned supermarket company stay relevant if the stock market penalizes it for experimentation?
"Wall Street tends to look at a supermarket company's earnings flow in a pretty basic way," he says, "and this is viewed as an earnings drain. Supermarkets aren't supposed to be involved in R&D."
Meanwhile, everyone else in the consumer-direct sector was sucking in huge amounts of capital. Nordstrom Inc., the department store chain, invested $22.8 million in Streamline. HighPoint raised $6.4 million from a group of investors that included United Parcel Service, and Andersen Consulting became a minority shareholder in ShopLink.
"We're trying to create something new for the consumer, to meet an unarticulated need," Furber says a few days later, sitting in a conference room next to his office. "We want to give them some time back." That, he hopes, will be the hook for consumers: Unlike Amazon.com, which is about selection and convenience, HomeRuns and the other grocery services are about reallocating time. (A Streamline brochure muses, "Suppose you could get back the time you spend on errands.")
Furber anticipates the moment of combustion - when the possibilities suddenly begin to resonate, consumer-direct catches on, and people start signing up at a furious pace. When it becomes impossible to dismiss his as just a little niche market. After a long slog, it finally looks like he and his three competitors might actually live to see widespread acceptance of their vision.
"Everything's a niche market when it starts," he says. "Electricity was a niche market." He pauses for a second. Furber seems constitutionally unable to rush his thoughts, or this business he's building. Patience is one of the virtues - rarely evident in a start-up - that he brings to this endeavor.
"It's too logical," he says. "Somebody's going to keep pounding away at this until it works."
