Business Horizons (2014) 57, 289—300

Available online at www.sciencedirect.com

ScienceDirect
www.elsevier.com/locate/bushor

TEACHING CASE

HomeGrocer.com: Anatomy of a failure

Greg Fisher a,*, Suresh Kotha b

a
Kelley School of Business, Indiana University, 1309 E. Tenth Street, Bloomington, IN 47405-1701, U.S.A.
b
Michael G. Foster School of Business, University of Washington, Box 353226, Seattle, WA 98115, U.S.A.

KEYWORDS Abstract This case study examines the initiation, financing, development, and
Business models; failed merger of an ambitious online grocery retail venture: HomeGrocer.com. It
Entrepreneurship; highlights the risks and challenges associated with developing a revolutionary ven-
Venture financing; ture. Even though the HomeGrocer.com team did many things right developmentally,
Online grocery; some key errors and unfortunate timing resulted in them expanding too quickly and
Merger; running low on cash. It also forced them to make a tough strategic decision about
Bankruptcy; whether to scale back operations and renege on their IPO commitments or merge with
Teaching case a well-funded competitor, Webvan. The case discusses lessons that can be learned
from business failure.
# 2013 Kelley School of Business, Indiana University. Published by Elsevier Inc. All
rights reserved.

1. Learning from the past? western United States. They made plans for more
locations and were in the midst of a fast rollout at
Terry Drayton, co-founder and first CEO of the now the time of HomeGrocer.com’s initial public offering
defunct HomeGrocer.com, left the meeting with his (IPO) on March 9, 2000. One month later, the
original management team feeling charged and hope- NASDAQ dropped 35% from its high of 5,132 and
ful. They had gathered to discuss the possibility of the opportunity for a secondary round of funding
reviving the online grocery business. Although the disappeared. Lacking cash to deliver on the growth
initial iteration of HomeGrocer.com had fallen victim plans laid out in its IPO prospectus, HomeGrocer.
to a failed merger with Webvan, a potential reincar- com tried to weather the financial crisis by merging
nation was now attracting interest from investors. with Webvan, a California-based online grocer that
The HomeGrocer.com saga began in 1998, when was launched in 1999.
Drayton and his Seattle-based team launched the Webvan, with its larger capital base1, took con-
service. They raised over $440 million in capital trol of the merged company. To the chagrin of
in multiple rounds of financing and opened eight HomeGrocer.com executives, Webvan’s managers
distribution centers in six locations throughout the transitioned all of the distribution centers to the

1
* Corresponding author At its 1999 peak, Webvan had a market capitalization of
E-mail addresses: [email protected] (G. Fisher), $8.8 billion compared with HomeGrocer.com’s $725.5 million
[email protected] (S. Kotha) at IPO.

0007-6813/$ — see front matter # 2013 Kelley School of Business, Indiana University. Published by Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.bushor.2013.12.002
290 G. Fisher, S. Kotha

Webvan technology and created a single Webvan It’s important to understand the drivers of supply
brand. However, despite their efforts, the Webvan and demand, and therefore how Internet-
managers burned through cash quickly while losing based distribution may affect this chain and
money. As a result, their share price plummeted whether there will be disintermediation. . . .
from $26 in December 1999 to $4 at the time of the For example, Amazon.com would be classified
merger (August 2000). By January 2001, the price of as a disintermediary, as it removes the inventory
Webvan shares was only 5 cents. In June 2001, less chain to the bookstore (e.g., brick-and-mortar
than a year after the merger was announced, Barnes & Noble stores) and potentially creates
Webvan filed for bankruptcy. better unit pricing (than the traditional stores),
Drayton had learned a great deal from his experi- better service, and so forth, while reducing the
ence at HomeGrocer.com. He strongly believed that capital tied up in inventory. It’s pretty simple–—
he and his team had done many things right and were less overhead, more efficiency and better prices
close to resolving the challenge of creating an online for consumers.
grocery business model. As he drove home from the
These comments excited investors, fueling the
meeting with his former colleagues, he wondered
funding of many online ventures via a buoyant
what he could do differently. Reflecting on his
venture capital (VC) industry. In May 1998, Drayton
experience in founding HomeGrocer.com, Drayton
raised $4 million in a ‘Series A’ funding round. Tom
thought about the start-up phase, the merger deci-
Alberg (personal communication, November 10,
sion, and the post-merger actions to try and isolate
2007) of Seattle’s Madrona Ventures, an early VC
what critical mistakes he and his team would avoid
investor in HomeGrocer.com, recalled:
the second time around–—if given the chance.
We had already invested in Amazon.com, so we
were beginning to see that online commerce
2. The start-up phase could work. This [venture] was much different
[in] everything from the delivery to the inven-
Terry Drayton and Mike MacDonald founded tory to a whole lot of issues. . . .Nonetheless,
HomeGrocer.com, an Internet-based grocer, in lots of people were frustrated with the regular
1994. Drayton had just sold Crystal Springs Water grocery stores. . .the lack of inventory some-
Company, which delivered bottled water to busi- times and having to carry [grocery] bags out,
nesses in and around Vancouver, B.C., and was looking and parking and everything else. We thought
for something new. One cold winter’s day, Drayton’s there was an opportunity and that it could
wife remarked that he ought to deliver groceries be a really big business. Unlike Amazon.com,
instead of water so that busy moms wouldn’t have though, you can’t serve the whole country from
to brave a snowstorm just to go to the store. The idea one warehouse initially. You’ve got to start by
stuck and Drayton found himself discussing it with his building out in cities. We realized it was going
friend MacDonald, president of the largest food bro- to take a lot more capital. But we were at-
ker in the Canadian province of British Columbia. tracted by Terry’s idea of doing it in a relatively
MacDonald was intrigued and agreed to put up some low-capital way.
initial capital while Drayton wrote the business plan.
Four months later, the HomeGrocer.com team
MacDonald’s friend, Ken Deering, who had technology
raised another $6 million in a ‘Series B’ funding
expertise, soon joined them as vice president round, with the high-profile Silicon Valley VC firm
of systems. The new venture would be called
of Kleiner Perkins Caufield & Byers as the lead inves-
HomeGrocer.com.
tor. This was followed by a ‘Series C’ funding round.
By this time, the market was abuzz with hopes for the
2.1. Initial funding
Internet grocery business, and HomeGrocer.com had
attracted the attention of some well-known individ-
By April 1997, Drayton had raised $350,000 in seed
uals and organizations as investors: Amazon.com; Jim
capital from friends and family. In search of more
Barksdale, an early executive with Netscape and
funding, he and the team moved their operations to
former senior executive with Federal Express;
Seattle to be in a more ‘techno-friendly’ city. This
and John Malone, the legendary cable TV operator
came at a time when respected Wall Street analysts
and Liberty Media investor. Terry Drayton (personal
and investment bankers were touting the Internet’s
communication, February 29, 2008) recalled the dy-
potential ‘disintermediate’ incumbents in a number
namics of the ‘Series C’ funding round:
of industry supply chains. Mary Meeker, the Morgan
Stanley stock analyst, noted (Meeker, Pearson, & We went out to raise $20 million, but everybody
Roach, 1997, pp. 10—14): wanted to own more of the company, so the
TEACHING CASE–—HomeGrocer.com: Anatomy of a failure 291

round kept going up. Amazon.com insisted on retailing had experienced growth in real sales during
owning 35% and we had a valuation in mind so only 4 of the previous 10 years, and net profit margins
we just had to raise way more money than we remained thin at 1% of sales. The industry was, and
ever intended: $52.5 million in total. continues to be, highly fragmented.
According to 1998 reports, the United States
A ‘Series D’ financing round followed in April 1999
population growth was expected to decline. Survival
with HomeGrocer.com raising another $110 million.
of the grocery industry would rely on cost-contain-
Primary investors during this round of financing
ment efforts, increases in economies of scale and
included CBS Inc., the Knight-Ridder Company, Soft-
scope, heavy penetration of existing markets, and
bank Co. of Japan, and Sequoia Capital. With the
greater-than-average consumer expenditures. To
frenzy over Internet-based ventures in full swing,
increase revenues, grocers had to increase the
Drayton raised a total of $172 million from VC
breadth and depth of fresh produce, meat offerings,
investors, even though his original business plan
and organically grown items while relying on pri-
only called for $15 million in VC funding followed
vate-label products with higher margins.
by an IPO to secure another $50 million in capital.
Drayton also assembled an impressive board
2.3. The market for online groceries
of advisors including Jeff Bezos, the founder of
Amazon.com; John Doerr, the high-profile VC from
Forrester Research (www.forrester.com) forecast
Kleiner Perkins Caufield & Byers; Jim Barksdale; and
that the total United States Internet-retail com-
Martha Stewart.
merce would grow from approximately $20.3 billion
in 1999 to approximately $184.5 billion in 2004,
2.2. The grocery industry representing a compound average growth rate
(CAGR) of over 55%. It estimated that online grocery
In the late 1990s, the United States grocery industry spending would increase at a CAGR of over 100%
was mature and extremely competitive; Table 1 during the next 5 years, from $513 million in 1999 to
provides an overview. In 1998, retail supermarket $10.8 billion by 2003. Despite its size in absolute
sales totaled approximately $449 billion. Grocery terms, this spending was expected to represent less

Table 1. Grocery industry overview, 1998

Number of employees 3.5 million
Number of grocery stores 126,000
Total grocery store sales $449 billion
Total supermarket sales $346.1 billion
Number of supermarkets ($2 million+ in annual sales) 30,700
Net profit after taxes, 4/98-3/99 1.03%
Typical supermarket size 40,483 sq. ft.
Number of items in a supermarket 40,333
Labor as a % of operating expense 57.5%
Average effective income tax rate (fed, state, local) 4/96-3/97 39%
Percentage of disposable income spent on food
 food-at-home 6.6%
 food away-from-home 4.2%
Weekly sales per supermarket $331,411
Weekly sales per square foot of selling area $9.45
Sales per customer transaction $10.16
Sales per labor hour $113.21
Average # of trips per week consumers make to the supermarket, 1/98 2.2
Food basket costs, % of weekly income spent on food
 Unites States 8.8%
 Canada 10.3%
 Japan 17.6%
Sources: U.S. Department of Labor, U.S. Department of Agriculture, Progressive Grocer magazine, and U.S. Census Bureau
292 G. Fisher, S. Kotha

Table 2. The emerging online grocer industry forecasts

1999 2000 2001 2002 2003

Online Grocery Sales

Specialty $248 $659 $1,548 $3,058 $6,291
Full Service $265 $473 $911 $1,951 $4,545
Total Revenues $513 $1,132 $2,459 $5,009 $10,836
Percentage of Industry Total 0.11% 0.23% 0.49% 0.98% 2.05%

Total Number of Households Buying Groceries Electronically (000s)

PC-Based 920 1,839 3,678 6,252 10,628
Specialty Other Net Devices 35 106 318 1,114 4,454
Total Specialty 955 1,945 3,996 7,366 15,082
PC-Based 187 327 621 1,242 2,546
Full-Service Other Net Devices 6 12 35 124 495
Phone/Fax 42 42 38 34 31
Total Full Service 235 381 694 1,400 3,072
Total Households doing both 59 115 243 630 1,536

Total participating households 1,131 2,211 4,447 8,136 16,618

Source: Forrester Research, Inc. (www.forrester.com)

than 2% of the total United States market for grocery Tom Alberg (personal communication, November 10,
products in 2003. 2007) noted:
Forrester categorized the online grocery industry
I think there have always been people who have
into two segments: full-service retailers and spe-
said, ‘‘[Online markets] will never work for
cialty stores. Full-service retailers are those selling
groceries. Margins are too low. How can you
a complete range of grocery products, including
purchase at the same competitive prices?’’ A lot
perishables. Specialty stores, in contrast, offer a
of which is true, but there are also online
limited selection of gifts, hard-to-find items, or bulk
advantages. You don’t have to have the expen-
replenishment products via the Internet. Table 2
sive space, parking lots, and as many employ-
provides a detailed breakdown of predicted elec-
ees. These are huge advantages.
tronic grocery spending in the U.S. from 1999—2003.
At that time, Forrester Research estimated that by At the time, online grocers were not responding to
2003, 16 million households would buy groceries pent-up demand. Demand had to be created. Be-
online. cause this kind of service was so novel, potential
Traditional brick-and-mortar retailers were slow customers had to be educated. Many questions
to embrace the Internet as a medium for selling about the transaction process had to be answered.
groceries. Albertson’s and Hannaford Brothers were Despite these concerns, a number of online grocers
among the few traditional grocers that experi- began to sell over the Internet and deliver to cus-
mented with an online model. A major uncertainty tomers’ homes. Initially there was little head-to-
for those launching online grocery ventures was if, head competition among new entrants, since each
or when, other major chains might enter the online focused on a specific geographic region.
space. Given their buying power, traditional re-
tailers had the ability to source products at a lower 2.4. Launching in Seattle
cost compared to new entrants. While some detrac-
tors of online grocery retailing cited many down- The HomeGrocer.com business plan dictated that
sides to the online model–—including supply chain the company would spend its first 12—15 months
complexities, consumer distrust, and technology establishing a distribution center in Seattle and then
and payment challenges–—others were much more roll out new facilities every 3 months in other U.S.
positive about the prospects of selling groceries cities. The idea centered on learning as much as
online due to potential cost savings of the model. possible from the launch of the first distribution
TEACHING CASE–—HomeGrocer.com: Anatomy of a failure 293

Table 3. Financial projections for HomeGrocer.com from the business plan ($ 000s)
CORPORATE FINANCIAL PERFORMANCE ($ Thousands)
Year 1998 1999 2000 2001 2002 2003

Number of Centers 1 3 7 15 31 31
Revenue 11,199 49,960 125,718 291,816 634,333 1,125,123
Gross Margin (%) 19.4% 16.7% 15.3% 15.0% 14.7% 12.0%
Net Income Before Tax (1,349) (2,901) (2,821) (1,065) 2,685 12,467
Net Income (%) (12.0%) (6.2%) (2.2%) 0.4% 0.4% 1.1%
Members 7,700 27,600 69,700 157,000 336,400 547,200
Working Capital 9,622 4,606 44,377 34,289 18,980 25,415
Total Assets 12,149 10,557 57,506 62,185 76,672 92,810
Employees 95 323 800 1,771 3,753 5,684

PROTOTYPE CENTER FINANCIAL PERFORMANCE ($ Thousands)

Year 1 2 3 4 5

Revenue 12,960 30,948 47,108 60,309 72,152

Gross Margin (%) 18.5% 13.4% 13.0% 12.8% 12.7%
Net Income Before Tax (629) 618 1,962 2,931 3,903
Net Income (%) (4.9%) 2.0% 4.2% 4.9% 5.4%
Members 8,400 15,600 21,600 26,400 31,200
Employees 91 313 784 1,753 3,733
Working Capital 532 918 2,666 5,616 9,656
Total Assets 2,122 3,215 5,572 8,819 13,038
Source: Extracted directly from the HomeGrocer.com business plan dated December 31, 1997

center to perfect the model before replicating it in centers were opened at a rate much faster than
other regions. The HomeGrocer.com website went originally called for in the business plan. Drayton
‘live’ in May 1998, and commercial operations were (personal communication, February 29, 2008)
officially underway in Seattle with a staff of 60. pointed out:
During its first month, HomeGrocer.com attracted
The biggest consequence of raising a lot of ven-
300 customers; on average, each placed three $75
ture capital was [that] with all this money, ev-
orders per month. According to Drayton’s business
eryone expects you to grow that much faster.
projections (Tables 3 and 4), the Seattle facility
That’s very difficult to do in a logistics-intensive
would break even at 1,100 orders per day.
business like HomeGrocer, especially when your
business is not yet fully baked. If you’re the
2.5. Expanding beyond Seattle
entrepreneur, what you’d love to do is stuff
the money in the bank and take your time,
A year later, in May 1999, HomeGrocer.com entered
knowing that you don’t actually have to go spend
Portland, Oregon, with a distribution site. However,
more time raising money. Then you can focus on
the company had not yet broken even. By September
perfecting the venture. Venture capitalists don’t
1999, it built another distribution center–—this one in
like that. They want you to go full speed all the
Orange County, California, where HomeGrocer.com
time with the gas pedal pegged to the floor.
employed a more automated system. Items were
picked in zones and batches, and conveyors were
used instead of having people walk around to make 2.6. The HomeGrocer.com approach
selections. Rather than its usual 50,000-square-foot
facility, the Orange County center was double that HomeGrocer.com was established as a full-service
size. The number of employees also increased rapid- online grocer offering perishable and non-
ly, doubling every 5 or 6 months. By November 1999, perishable items. The company enabled customers
the payroll reached over 1,000 employees. living within a designated region to buy groceries via
Over the next few months, HomeGrocer.com its website and then have the purchases delivered to
opened five new distribution centers: second facili- their homes. The company invested heavily in cre-
ties in Orange County and Seattle, two in Los An- ating a website that replicated many aspects of the
geles, and one in San Diego. The new distribution shopping experience for customers used to buying
294 G. Fisher, S. Kotha

Table 4. Break-even analysis for HomeGrocer.com from the business plan: Analysis per center

These projections assume an average order size of $100 with members shopping twice per month. Only 15% of orders
are assumed to be less than $75 requiring a $9.95 delivery fee. A membership of $35 is collected from 90% of members.
Hence an average member purchases $2,467 per year in products from HomeGrocer.com as shown below:
Amount % of average
grocery sales
Average Sale $2,400.00 100.0%
Average Delivery Fee 35.82 1.5%
Average Membership Fee 31.50 1.3%
2,467.32 102.8%
Average Cost of Sales
Groceries/shrinkage 1,735.20 72.3%
Delivery 223.20 9.3%
Picking 208.80 8.7%
Transaction 19.20 0.8%
Subtotal 2,186.40 91.0%
Annual Gross Margin per member $280.92 11.8%
The above analysis uses costs at the end of 1 year and excludes all revenues forecast from manufacturer programs.
Revenues from these activities are forecast at $1,740,000 in the first year and $600,000 thereafter.

Fixed costs to operate a HomeGrocer.com site are approximately $3 million annually, which are determined as follows:

Annual Fixed Costs to Operate a HomeGrocer.com Center

Cost Amount
Warehouse Rental $243,000
Triple Net and Utilities 102,000
Warehouse Equipment Maintenance 48,000
Insurance 24,000
Office Equipment Maintenance 12,000
Supplies 60,000
Legal and Accounting 60,000
Communications 90,000
Administration 784,700
Member Services 176,229
Sales & Marketing 1,236,000
Depreciation 137,550
Miscellaneous 60,000
Total $3,033,875
As operating efficiencies improve over the year and some expenses are front-end loaded, actual monthly breakeven is forecast to
occur in month 14. So each site is forecast to achieve a break-even level of operation in just over 1 year.
Source: Extracted directly from the HomeGrocer.com business plan dated December 31, 1997

goods in a traditional grocery store. The website was 2.7. The ‘online’ value proposition
user friendly, time efficient, and featured 9,000—
12,000 items2 via an intuitively organized list. The The key advantages to the consumer were conve-
customer could either browse the shelves or use the nience, quality, and service. HomeGrocer.com
search function to locate items. Customers would go customers could shop 24/7 from the comfort of
online, order groceries, and select a 90-minute their couch, and goods were delivered directly to
window of time between 1 p.m. and 9 p.m. the their homes; they no longer needed to carry heavy
following day to have the items delivered. Orders grocery bags to and from their cars. Because the
over $75 had no delivery charge, while orders under perishable items sold online hadn’t been put out
$75 accrued a $10 delivery charge. for display or handled by other shoppers, they
were of higher quality and lasted longer than
those purchased in a traditional grocery store. John
2
A traditional store carried over 100,000 items or SKUs. Landers (personal communication, March 6, 2008),
TEACHING CASE–—HomeGrocer.com: Anatomy of a failure 295

vice president of marketing at HomeGrocer.com, HomeGrocer.com’s advertising and promotion

noted: initially focused on the convenience of buying on-
line. After Jon Landers joined as vice president of
Clearly, our produce was fresher than that bought
marketing in November 1998, the company began
at a traditional grocery store. We operated on
promoting the freshness and quality of its products.
a just-in-time inventory basis and did not require
Landers decided to stress this advantage by sending
surplus merchandise for display. Theoretically
free sample produce bags with each order.
[in a traditional grocery store], customers could
In addition to distributing free fresh produce bags
select produce to their liking from the available
in targeted neighborhoods, the HomeGrocer.com
produce by touching and scrutinizing. However,
team gave away coupons to encourage customers
customers damage a good deal of produce by this
to buy, and eliminated delivery charges for first-time
inspection and are limited to selecting from a
buyers. This was a costly exercise. The company
picked-over assortment. HomeGrocer.com pick-
spent approximately $300 per customer in acquisition
ers were able to select fruit to the customer’s
costs but struggled to retain these clients; many
specifications without damaging the rest of the
would stop buying after just three or four purchases.
produce.
Picking up on this, Landers implemented a psycho-
HomeGrocer.com focused on quality customer ser- graphic profiling process, using GIS (Geographic In-
vice. This entailed responding to customer requests, formation Systems) mapping to ensure that the team
following up after completion of a delivery, and was targeting areas with customers who had the
training drivers to be sensitive to customer needs. appropriate profile. The move helped bring down
The driver was seen as a company representative acquisition costs and improved customer retention
since he was the only employee the customer ever rates, but did not completely solve the problem.
saw. Delivery personnel were instructed to carry Because all commerce was completed via the
groceries into the kitchen or pantry after asking Internet, boosting web traffic was critical to gener-
the customer whether they could enter the house. ating sales. This was at a time before email or search
Wearing surgical booties over their shoes was man- engine marketing. To drive traffic to the company
datory in order to leave no impact on the customer’s website, HomeGrocer.com used traditional media
home. Such attentiveness made a big impression on including direct mail pamphlets and newspaper and
HomeGrocer.com’s customers. Ken Deering (personal television ads. It also used unconventional marketing
communication, November 14, 2007) observed: in the form of Peach Parties, a type of Tupperware
party for early Internet users; and extensive sponsor-
We’d get comments from people calling and
ships of local community organizations, such as do-
saying, ‘‘The best thing about your business is
nating apples and bottled water to school PTA
I never have to pick up another 20-pound bag of
members who handed them out with direct mail
dog food in the store, and the driver puts it in the
pamphlets that promoted a gift of $25 for the orga-
garage. He knows that’s where we keep it.’’ Or,
nization if people tried HomeGrocer.com. The com-
‘‘I can’t believe he took the time to put the soap
pany also entered into partnerships with AOL and
on top of my washing machine on the way out the
Amazon.com. Of these partnerships, Drayton (per-
door.’’
sonal communication, April 18, 2008) remarked:
In addition, each first-time customer received a
These were basically the big players in the Inter-
phone call from a customer service representative
net, especially for customer acquisition. We had
to make sure he/she was completely satisfied with
Amazon.com as a shareholder [and so] we had to
the order.
deal with them. We signed a big AOL deal be-
cause at that stage AOL was so dominant, they
2.8. Marketing and customer acquisition
were about 40% of the overall market, but for us
they were all our target customers, and so it was
The company targeted busy middle- to upper-class
probably a disproportionate share. Most people
families ($75,000+ per year) with school-age chil-
who used AOL thought that it was the Internet
dren and both parents in professional occupations.
at the time.
Such families were typically short on time, spent
enough on groceries to make buying online attrac- Even though HomeGrocer.com was able to create
tive, and had access to the Internet at home. customer interest and get people to explore the
According to the company’s market research, company’s website, it encountered barriers to first-
women made the buying decisions in these homes. time and repeat orders. For many individuals, buy-
As a result, the website was designed to appeal ing groceries at a store was such a natural way of
to them. doing things that they struggled to move away from
296 G. Fisher, S. Kotha

the practice. Deering (personal communication, area. In this respect, the online grocery business
November 14, 2007) called this the ‘gallon of milk’ departed from other Internet retailers like Ama-
problem: zon.com, which focused on a wide geographical
area. HomeGrocer.com boasted an impressive 98%
Your spouse is on his way home and you call up
accurate delivery record. It also used a sophisticat-
and say, ‘‘We’re out of milk. Could you please
ed third-party program, similar to that employed by
stop and pick up some milk on your way back?’’
UPS, to route deliveries efficiently. According to
He stops to buy a gallon of milk. While buying
Drayton, the company was a pioneer in its technol-
milk, he does other grocery shopping. Because
ogy. At the same time, software development led to
of this, this family ends up not shopping online
significant cost overruns; website development
that week.
alone reportedly cost $750,000.

2.9. Operations and delivery 2.10. Competitors

HomeGrocer.com opted to build its own distribu- A number of other companies in the United States
tion centers and purchase refrigerated delivery were also attempting the online grocery retail mod-
trucks. The distribution centers were built close el, including Peapod, which delivered products from
to the neighborhoods the company served. Because existing retail stores; Netgrocer.com, which deliv-
of the high upfront investment in infrastructure ered nonperishable packaged items using courier
and systems, each facility needed 1,100 orders per services like FedEx and UPS; and Webvan, the grand
day, at an average order value of $100, in order project of Louis Borders, co-founder of Borders
to break even and become profitable. The first Books and Music. Webvan attracted large amounts
55,000-square-foot facility was rented just outside of capital from high-profile VCs with its promise to
of Seattle in Bellevue, Washington. revolutionize retail by creating a web-based store-
Since small operations do not enjoy the strong front that sold ‘‘everything from groceries to Palm
wholesaler relationships that large brick-and- Pilots.’’ With its massive 330,000-square-foot Bay
mortar stores typically do, Homegrocer.com gener- Area warehouse, fleet of trucks, and highly sophis-
ally paid more for its inventory. In spite of the ticated automated picking system, Webvan was
increased cost of goods sold relative to larger re- considered HomeGrocer.com’s fiercest rival. Both
tailers, HomeGrocer.com’s original business plan companies were striving hard to be first to market
projected net profit margins of 7% over the long in establishing a web-based retailing presence in the
term (compared with 1%—2% for traditional grocery biggest cities in the United States.
chains). These higher projections were based on its
lower location costs and not needing to build retail 2.11. Going public
displays or hire checkout clerks.
Logistically, the business concept presented the In 1999, the investment climate for Internet IPOs
company with significant challenges. HomeGrocer. seemed strong, and HomeGrocer.com prepared to
com’s top management agreed that the most com- go public. The company’s financials revealed that it
plicated part of the business was the supply chain. had lost $7.9 million on $1.1 million in revenue in
When an order came in, a picker would start at one 1998 and $84 million on $21.6 million in revenue in
end of the facility and work his/her way to the other, 1999. Prior to its public offering, the board decided
selecting all of the ordered items. If something to bring in an experienced CEO to replace Terry
wasn’t in stock, an employee would have to rush Drayton, along with a new CFO. While Drayton
out and buy it from a local grocery store. While its was not overly pleased, he remained supportive
competitors would simply ship an incomplete order of the plan. Board members considered candidates
if they did not have an item in stock, HomeGrocer. with the relevant logistics, financing, and manage-
com’s zeal for customer service prompted it to ment experience. They hired Mary Alice Taylor–—
ensure that customers received exactly what they former CIO at Citigroup and, prior to that, a senior
ordered. This policy had significant cost and time executive of operations at Federal Express–—as CEO.
implications in that pickers could spend up to 50% of As CFO they hired Dan Lee, former CEO of Mirage
their time filling a few items on an order. Resorts for the preceding 10 years.
The company’s trucks featured compartments HomeGrocer.com went public on March 10, 2000.
with three different temperature settings: one for The stock opened at $12.88, reached a high
frozen food, a cooler for produce, and room tem- of $16.25 on the first day, and closed at $14.12.
perature for dry goods. Each truck could deliver HomeGrocer.com’s top management team had
30—40 orders per day within a narrow geographical hoped to bring in between $500 and $600 million
TEACHING CASE–—HomeGrocer.com: Anatomy of a failure 297

in capital from the IPO, yet ended the day raising a April 2000, the bankers said: ‘‘It may be a blip;
total of just $268 million. maybe by the summer it will reopen.’’ And then
by about June 2001, it was: ‘‘It’s all over. Who
2.12. The stock market collapse knows when it will reopen.’’
One option was to scale back expansion plans and
In April 2000, the NASDAQ index declined sharply as focus on reaching profitability by improving existing
the Internet bubble burst. Overnight, Wall Street’s operations. At one point, company executives enter-
Internet mantra changed from ‘get big, fast’ to ‘get tained the idea of restructuring HomeGrocer.com
to profitability.’ HomeGrocer.com had spent 18 under Chapter 11. However, since the company
months gearing up for a rapid rollout and slowing had just gone public, declaring Chapter 11 would
it down would be tough. The company lacked the almost guarantee shareholder lawsuits. The com-
capital needed to achieve positive operating cash pany’s board was sure to oppose that route.
flows on its eight distribution centers. In order to Another option was merging with a wealthier
remain a viable operation, it needed between $300 competitor like Webvan. Louis Borders founded
and $500 million in additional capital, yet it was Webvan with the bold vision of ‘‘selling and
unclear when–—or even if–—the financing window delivering anything and everything.’’ For him,
would reopen. Given such uncertainty, growth plans delivering groceries was just the beginning, a
were put on hold. At the same time, rival companies way of getting into consumers’ homes. His plans
began to compete head-to-head in some markets. for Webvan were audacious, as he ‘‘sought to
Without additional capital, it was unlikely that outsmart the biggest players around, from Ama-
HomeGrocer.com could support its large workforce zon.com and Wal-Mart to UPS and the U.S. Postal
or fund the expansion plans promised in its IPO Service’’ (Helft, 2001). Within months of launching
prospectus. Both management and the board were Webvan, he was able to raise over a billion dollars
concerned about lawsuits from investors if expan- through a successful IPO. One HomeGrocer.com
sion plans were not carried out. executive (personal communication, January 15,
These were difficult times for HomeGrocer.com. 2008) recalled:
Although the company raised $268 million at the
IPO, financial losses were mounting. By this time, [Webvan] had done a secondary offering. They
HomeGrocer.com had over 2,300 employees and the had tons of cash. We had eight operating cen-
company’s stock was trading at $5 a month after the ters and they had one. If we merged [we would
IPO, representing a 64% drop in value in 1 month be] able to get out from most of the commit-
since listing. Terry Drayton had recently left the ments we had made. We could say: ‘‘Now that
company due to differences with Mary Alice Taylor, we’ve merged the companies, everything’s
the new CEO. Under these trying conditions, she changed and we’ve got to cut this back and
began exploring ways to move forward. we’ll only open their facilities.’’ And so we
thought, ‘‘Here’s our way out.’’
If HomeGrocer.com merged with Webvan, the
3. Key decisions combined company would have $500 million in cash
(Hobson, 2000)–—enough to weather the storm and
The stock market meltdown that occurred days move the combined entity toward profitability. But
after HomeGrocer.com’s March IPO squelched the the two companies had very different cost struc-
company’s ability to raise more capital. The top tures (Drayton, personal communication, February
management team, under Taylor, faced limited op- 29, 2008):
tions. Drayton (personal communication, April 18,
At that stage, one of the key metrics was
2008) recalled:
average order size. You need at least a $100
We’d just gone public, and when you go public order to generate $30 in gross margins to pay for
you issue a prospectus in which you make all the picking, to pay for the shipping, and all
sorts of commitments. We committed to open- that. [Webvan’s] picking costs were high be-
ing new markets and distribution facilities. So cause they had these really automated facili-
one of the biggest questions and hardest dis- ties. They needed staggering numbers of orders
cussions we ever had with the investment bank- to cover costs. It cost them something like
ers was if and when the market would ‘reopen.’ $100,000 a day whether they picked anything
We needed more capital to meet our commit- or not to keep the facility open, so divide that
ments. We asked: ‘‘Is the market going to re- by 500 orders and all of a sudden it’s really
open in time? Or is it ever going to reopen?’’ In expensive to pick your orders. Whereas if we
298 G. Fisher, S. Kotha

Table 5. HomeGrocer.com and Webvan: A summary of revenues and profits ($ 000s)

HomeGrocer Webvan
1998 1999 2000 * 1998 1999 2000 **
Net Sales 1,094 21,648 21,215 - 13,305 16,269
Cost of merchandise sold 1,018 19,515 17,515 - 11,289 12,138
Gross Profit 76 2,133 3,700 - 2,016 4,131
Sales, general, and marketing expenses 7,455 59,208 13,899 8,825 104,152 47,352
Technology expenses 6,466 3,010 15,237 5,523
Stock-based compensation expense 412 28,158 8,143 1,060 36,520 17,720
Customer fulfillment center expenses - - 17,644 - - -
Pre-opening expenses - - 2,015 - - -
Loss from operations -7,791 -85,233 -44,467 -12,895 -153,893 -66,464

Interest Expense -172 -384 -663 -32 -2,156 -150

Interest Income 54 2,232 1,688 923 11,480 8,799
Other income/(expense) -608 -23 - - -
Net Loss -7,909 -83,993 -43,465 -12,004 -144,569 -57,815
Unrealized Gain (Loss) on Marketable Securities 4 -729 990
Comprehensive Loss -7,909 -83,993 -43,465 -12,000 -145,298 -56,825
*
13 weeks ended April 1, 2000
**
13 weeks ended March 31, 2000
Source: Hoovers.com

had about 500 orders a day, we were doing fine; of the merger, Webvan had about $500 million in
at 1,000 we’d break even. I think they needed cash, and HomeGrocer.com about $150 million.
3,500 orders a day to break even. And no facility
ever, in any of the markets we were operating,
got close to 3,500. We did about 2,100 in Seat-
tle once, which was a record for one day.
4. The death spiral
For 2 years, Webvan and HomeGrocer.com had been Once Webvan took control, its executives exerted
fierce competitors. Merging with the enemy seemed influence on the merged company and things started
like a bitter pill to swallow for certain members of to quickly unravel. Webvan switched all HomeGrocer.
the HomeGrocer.com management team. Table 5 com facilities to its technology platform. Immediate-
provides a summary of HomeGrocer and Webvan’s ly following the change, customer orders in San Diego
financial performance at the time of the merger dropped from 1,000 to 500 a day, and within a week to
talks. On June 26, 2000, HomeGrocer.com an- 300 per day. The decline soon repeated itself seven
nounced a merger with Webvan in an all-stock additional times at other facilities. As Terry Drayton
deal involving 138 million Webvan shares valued at (personal communication, April 18, 2008) recalled:
$1.2 billion. Webvan was to exchange 1.076 shares for
each share of HomeGrocer. For Webvan, the deal Webvan’s technology did not work with AOL.
eliminated a tough competitor and gave it an imme- The ‘Buy’ buttons and the checkout on the
diate presence in six new markets. The combined website did not work, so clients could not make
company projected it would save an annual $20—$30 purchases even if they wanted to. Plus, it was
million in redundant overhead costs. brutally slow since they only tested it on T1 and
The acquisition was completed in September T3 lines used by their developers. They did not
2000. By then, Webvan’s stock price stood at test it under the conditions that AOL soccer
$3.88. Its purchase of HomeGrocer.com was now moms would experience it, using a dial-up mo-
valued at $535 million, substantially less than its dem. We [HomeGrocer] had an in-house lab
initial offer of $1.2 billion. George Shaheen, CEO of that always tested new releases with represen-
Webvan, remained the president and CEO of the tative users. I could forgive them for inade-
merged entity. At year’s end, the combined compa- quate testing on the first conversion, but not for
ny operated in 13 markets and was set to open two repeating the mistake seven more times over
distribution centers in 2001. Its new goal was to the following months. Unfortunately, Webvan
operate 25 facilities by the end of 2002. At the time management repeated this mistake with every
TEACHING CASE–—HomeGrocer.com: Anatomy of a failure 299

facility, with the same result. . . .The end re- released a dismal annual report in April. It included
sult was a 50% immediate drop in sales, and a warning from its auditor that expressed ‘‘substan-
down to 33% of sales within a week. tial doubt’’ the company could survive the year. The
report stated that Webvan would probably need to
In a related blunder, Webvan retained most of
raise additional capital by the fourth quarter. By this
HomeGrocer.com’s employees at a 50% premium
time investors had soured on Internet-based ventures
over their original salaries, but then gave them
and the company could not raise more capital. As The
nothing to do (Helft, 2001):
Seattle Times pointed out (Liedtke, 2001):
[Webvan] kept employees with identical jobs on
The gloomy outlook is reflected in Webvan’s
both sides, paying scores of workers retention
stock, which closed this week at 12 cents, a
bonuses and time-and-a-half on their already
far cry from its all-time high of $34. The com-
high salaries–—even as many of them had little
pany has warned that its shares are in danger of
to do. ‘‘The last month I sat in my office an- being delisted from the NASDAQ if the stock
swering maybe one question a day,’’ says Mike doesn’t start trading above $1.
Smith, Director of Distribution for HomeGrocer.
com. Shortly after the merger, Smith had gone On April 13, 2001, CEO George Shaheen resigned
to Foster City, Calif., to meet with his counter- and the company laid off 1,150 employees. New
parts. At the meeting, five Webvan employees CEO Robert Swan was unable to stop the bleeding or
introduced themselves and described their re- turn things around. In May, Terry Drayton made a
sponsibilities. When his turn came, Smith said: presentation to Swan suggesting ways to rescue the
‘‘I am director of distribution, and I do all of your company, including re-branding operations under
jobs.’’ It was all part of a culture where money the HomeGrocer.com name and rehiring former
was no object. HomeGrocer.com executives. He even offered to
buy back the company’s facilities in Seattle and the
By December, the media began reporting that cuts
HomeGrocer.com name. However, nothing came of
were expected at HomeGrocer.com as Webvan
the meeting.
worked to establish a single brand and integrate
On July 9, 2001, Webvan filed for bankruptcy and
the two technology platforms. Even though Home-
ceased all operations. Over 2,500 employees lost
Grocer.com had six locations and Webvan had just
their jobs, while 400,000 customers in Seattle, Port-
one, and although the HomeGrocer.com brand had
land, San Francisco, Chicago, Dallas, Atlanta, and
started to gain customer recognition for the peach Southern California lost their home grocery service.
logo on its website and delivery trucks, a strange
A former senior technical manager at Webvan was
switch occurred (Helft, 2001):
quoted as saying: ‘‘We bought them out, we killed
Blinded by their grand vision, Webvan’s execs their stock, we killed their company, and then we
made a string of bad decisions. Even as it was killed ourselves’’ (Helft, 2001). Alberg (personal
running out of money, the company spent mil- communication, November 10, 2007), the VC inves-
lions on a rebranding campaign to avoid being tor from Madrona, reflected on the merger and
pegged as a mere grocer: The marketing push everything that followed:
promoted Webvan’s sterile new blue-and-green
We [HomeGrocer] overextended ourselves by
‘‘W’’ logo. Gone were Webvan’s earthy grocery
trying to open too many markets too quickly.
bag and HomeGrocer’s fuzzy peach.
Then we kind of got in this panic situation like,
This move left many customers in Seattle, Portland, ‘‘Oh, we’re going to run out of capital.’’ We’d
and California unsure about what the change might gone public by then. ‘‘We’re going to run out of
mean. As cash continued to run out, management capital, we’ve got to do this merger.’’ Manage-
initiated cutbacks that only made matters worse. ment just became quite fixated that we had to
For example, during the last quarter of 2000 they do the merger as opposed to really trying to find
decided to slash marketing expenses by 28%. As a other solutions like slashing expenses, cutting
result, total orders fell 8% and orders from new back our expansion, and just letting Webvan
customers slid 48%. By January 2001, in order to fail. If we hadn’t done the merger, Webvan
conserve cash, Webvan postponed planned openings would have failed anyway. . . .The CEO and
in Washington D.C., Baltimore, and New Jersey. the CFO were absolutely convinced this was
There was some good news: In March, the com- essential or the company would fail. Well, we
pany’s Orange County facility #2 turned a profit, the did the merger and it failed. I think once the
first and only standalone Internet grocer to do so. merger happened, there was no chance of suc-
Overshadowing this achievement, the company cess. I mean, it was just preordained. . . .
300 G. Fisher, S. Kotha

We invest in a lot of companies that succeed rule out the possibility of making another go
and a lot of them fail, and sometimes it’s just of it.
totally the wrong idea. I don’t think web-
Noted David Billstrom, a managing partner at FBR
grocery sales is the wrong idea; it’s just hard
CoMotion, a VC firm in Seattle (Cook, 2001):
to execute. How do you figure out how to do
it, and how do you do it without spending too [I would] definitely take the meeting if Terry
much capital to get there? knocked on my door. But he would have to look
methodically at the failure points of Webvan/
HomeGrocer.com and then he would have to
show how he would overcome those failure
5. The post-mortem
points in the new incarnation. And then, most
importantly, he would have to get a track re-
Soon after Webvan filed for bankruptcy, there was
cord showing that the new and improved ver-
talk about HomeGrocer employees reviving the orig-
sion did not have the problems of the old.
inal company. Many showed support for the original
founder, Terry Drayton, and his team. Former cus-
tomers sent emails to the Seattle Post-Intelligencer References
with subject lines like ‘‘Bring back the Peach’’ and
‘‘HomeGrocer.com resurrection.’’ John Cook Cook, J. (2001). Venture capital: Rebuilding online grocer is tall
(2001), a reporter and columnist at the newspaper, order. Retrieved May 9, 2009, from http://www.seattlepi.
wrote: com/news/article/Venture-Capital-Rebuilding-online-grocer-
is-tall-1059693.php
To call the two-dozen e-mails I received pas- Helft, M. (2001). The end of the road. Retrieved June 10, 2009,
from http://kerchmax.tripod.com/webvan.html
sionate would be an understatement. These Hobson, K. (2000). Out to lunch? Webvan/HomeGrocer deal
online grocery shoppers clearly want to see doesn’t deliver all the answers. Retrieved October 12, 2008,
the well-polished, peach-colored HomeGro- from http://www.thestreet.com/tech/internet/977452.html
cer.com delivery trucks rumbling down their Liedtke, M. (2001). Webvan’s CEO resigns. Retrieved October 15,
streets again. . . .[Also] some of the venture 2008, from http://archives.seattletimes.nwsource.com/
cgi-bin/texis.cgi/web/vortex/display?slug=webvan14&date=
capitalists and angel investors I spoke to this 20010414
week–—including a few who invested in Home- Meeker, M., Pearson, S., & Roach, S. (1997). Internet retail: The
Grocer.com in the early venture rounds–—didn’t Internet Retailing Report. New York: Morgan Stanley.