Hewlett Packard The Fight Of Kittyhawk A

HARVARDIBUSINESSISCH00L
9‐ 606‐ 088
REVi OCTOBER 23.2006
CLAYTON M CHRISTENSEN

Hewlett‐ Packard:The Flight ofthe Kittyhawk(A)

In 1903′ on a wu■ dy beach in Kitty HalⅣ k′ North Carolina′ ()A/ille arLd ヽへrilbur Wright demonstrated that sustahed flightぃras POSSible.Though their Plane only fleぃ/120 fect on the first attempt′ they flew farther and higher on each sticcessive trial. Theれ、70 brOthers stoOd at the thresh01d of a new era in transPortatiOn.
In June of 1992′ Hewlett― Packard(HP)il■ troduced the smallest hard disk drive ila the world″ named the Kittyhawk.The drive′ s disks were l.3 inches h diametet the ttnit was not much larger than h″ O postage stamps(see Exhibit l fc)r a Picture of the Kittyhawk).ThC first version of the Kittyhawk supplied 20 megabytes of storago and had unique componentry cnabling the drive to withstand a three― foot droP ttrithOut any(lata loss.The possible applicatiOns of the drive in the mobile computing market seemed endless′ and the team at HP resPonSible for launching the Kittyhawk eagerly anticiPated the takeoff of their ne、vest innovation.However′ nottZ′ in lune 1994′ after two years Of effort′ Kittyhawk sales ha(l failed to meetthe team′ s targets′ and Rick Seymourノthe Kittyhawk PrOiect leader′ was struggling wiぬwhether to refocus the team′ s effort along one of three
possible paths his staff had developed or to reco■ llnend to his superiOrs that the prtDleCt be
abandoned and the tealn disbanded.
Rigid Disk Drives
Rigid disk drives(COmmOnly called″ hard dr􈎺 es′

)were magneic infom■ ation storage and
retrieval devices usedぃrith computers.The fiFSt rigid disk drive′ invented in 1956 by engineers in
IBヽ4′ s San JOSe′ california laboratories′ was the size of th/o large refrigerators placed side by side.It
could store five megabytes(MB)Of infOrntation.The technological progFeSS Shce that time was
remarkable.E〉rives wrote and read inforrnatiOn in the same sort of binary code that computers used.
Most disk d􈎸 ves comprised a read― ■、′rite head mounted at the end Of an arrn that swung over the
surface of a rotating disk il■ Inuch the same way that a phonOgraph needle and arFIヽreached over a
record,disks′ which、、アere alulninum or gitiss Platters coatcd、vith magnetic lnaterial,at least hん′o
electrical motors―a sph mOtOr that drove tlte・rotation of the disks and an actuator rnotor that rnoved
the head to the desired POSition over the disk,and a variety of electronic circuits that controlled the
drive′ s operation and its interface with the computer.
Professor Ciayton M Christensen and Research Associate Cre〔;olγ C・Ro8ers Pl・epared tile ori3hal、,ersioi、or this case′
′′
I Ie、γleu‐ Packard:T腱
Flight of the Klttyha、vk′
″HBS No 697-060 This version was pl・epared by Proressor Clayton 􈎺l Christonsen l‐ IBS cases are dweloPc・d S01ely as
the basis for class disctlssioll_Cases are not intel、ded to sewe a〔l endonttments,sources of Prinlary data′ or illustrations of efecuve Or llleffective
management.
COPyright()2006 Presidel、t and Fe1lo、νs of Harvard College l` o order copies or request perlnlsslon to repFOduCe rnaterials,ca11 1‐ 800-545‐ 7685′
write HaA7ard Busll■ ess Sc1lool Publishing′ Boston′ MA 02163′ or go to http://ヽヽvW hbsP haⅣard edLl No Part Of this Publication may be
reproduced′ stored 􈎸、a retrieval systt・m′ used il、a spreadsl■ c・et′ or transrnitted ill al、y form or by al、y means― ぞledro,、ic′ nled、allical′
PhOはOpying′ recording′ or Othe,vise― 、vitl、outthe perFussior of HaⅣ ard Busmess schO。1
Hewlett‐ Packard:The Flight ofthe Kittyhawk(A)
The read― ぃアrite head was a tiny electrOmagrtet whose polarity changed whenever the direction of
the elect􈎸 cal current running through it changed.Data、vere l、アritten onto disks by sending electrical
pulses thrOugh the head′ s electrOmagnet to create minute magl■ etic flux fields that oriented the
rnagnetic P01arity of particles(the dOman■ )on the disk surface irrtmediately beneath the heado With
dornains bcing given either positive or negative polariサin thiS way′ il■ forlrlatiOn could be stored in
binary code,Data could be retrieved from the disk′ s sttrfaCe by reversing this ProceSS・A disk drive′ s
arcl■ itecttlre was categorized by tho size of the disk′ s diameter(i.e.′ a3.5-inch disk drive).
The Disk Memory Storage Division
Fronl He、vlett― Packard′ s earliest beginl■ ings as an clectronic instruments comPany to its
domil■ation of the printer industrv′ the HP culture deeply valued technical innovation as a key to
success.HP employed a management― by― obiectiVe(MBO)prOCess to focus its businesses on financial
goals and its peoPle on the potential Patl■ S Of innovation and strategy to achieve such goals.HP
favored a decentralized organizatiOnal structure so as to allow its businesses freedom of decis10n
making and movement.
HP had fouF ma10r business OFganiZations― Test & Measurement′ Computer Systems′
Measurement Systems′ and Computer Products.Of these four′ the Computer Products Organセation
consisted of the company′ s laser pril■ ter′ inkづet printer′ personal computer′ al■d mass storage
product groups. The 卜〔ass Storage Product CrouP deve10ped and managed HP′ s storage
techn01ogies(賛e Exhibit 2 for theヽ4ass Stor(lge CrouP′S Organization chart).The Disk Memory
D􈎺ision(DMD)′ whiCh Was resPonSible for develoPing and latlnching the Kittyhawk and other disk―
drive rnodels′ resided■ vithin the Mass Storage C:roup.
Based in Boise′ Idaho′ D􈎺llD had disk― drive sales in 1992 of s519 1Ylillion′ approxllnately 80ツ3 of
which was der􈎺 ed from Original equiPment inanufacturer(OEM)orders and the rest froln HP's
internal computer businesses.l IIP′ s corPorate net revenue was at S16 billion.(See Exhibit 3 for a 10-
year comparison Of HP corporate revenues versus its disk― dr􈎺 e revenues,)At the same time′ the
hdustry leaders′ IBヽ4 and Seagate Techn01ogy′ had disk― drive sales of$4 billion and s3 billlon′
respectively.2
ノヽlthough slrlall in comparison to some of the other disk― drive manufacturers′ I)􈎺llD had a
profitable position within the market.It cOncentrated on high― perfomnance products within the 5.25‐
and 3.5-inch architectures that supplied healthy profit inargins to the division.Consequently′ HP′s
product line offered a substantially higher capacity in megabytes than the 􈎹dustry norm (see
Exhibit 4)。For many of DMD′ s FeSCaFCh and developmcnt(R&D)engineers″ the most sought´ after
pro〕ects were the ones that developed the next― gcneration drives that furnished ever― higher
capacities and faster access tinles.Concentrating on the high― end engineering workstation and
nelレ
‘ork server markets′
D􈎺ID had been among the first Ln the industry to introduce one andれャo
gigabyte dr􈎺 cs.3 TheSe disk dr􈎺 es were oxtremcly successftllin the markctplace.
11993E)isk/Trend report.
2 1bid.
3(〕
n〔:gigabyte is equivalellt to l′ 000 megabytes.
Hewlett‐ Packard:The Flight ofthe Kittyhawk(A)
The Kittyhawk
Gθ 17θ SiS O/助ご′θ′
Bruce Spenner′ the general lnanager of thc Disk 􈎺lemory E)ivision′ came to HP in 1978 from
teaching electrical cngineeril■g at Washington University in St.Louis.A few years aftcr,oining HP
Labs′ he was Part Of the companyノs toP_prioriりprOiect tO ilnplement reduced instruction set
computing(RISC)arChitecture into HP′ s nlinicomputers and workstations.IVhile other computer
companies did not fully believe the bencfits of RISC and balked in imp!ementation′ HP′ s full
comlnitnlent to the architecttlre rnade it the leader in Unix computing.Described by emP10yees as not
the tyPical″ meet the numbers″ inanager′ SPenner was、げidely vie、ved as a visionary and risk taker.
At HP Labs′ his responsibilities soon exPan,ied into overseeing a softwarc laboratory that produced
major new breakthroughs.Dick Hackborrl′ who had built HP′ s highly successftll printer business and
vvas now executive vice presidentin charge of the conlPany′ S COmputer Products C)rganization′ liked
SPenner′ s concept― driven thinking.In 1990′ Ray Smelek′ general manager of the Mass StOrage
Product Group′ PromOted Spenner to general manager ofthe Diskヽ4emory lDivision.
DMD was an anomaly within HP.It ha(l been selling drives externally to OEMs since 1984 and′
though Profitable′ was still a niche player frOFn a diSk― drive industry perspective.In contrast′ I‐IP as a
whole took pride in its ability to be a market leader.Bothered by D卜ЛD′ s PositiOn′ SPenner Often
asked′

How can we make HP a ma,Or player in the disk― dr􈎺 e industry?Why don′ t we have 20%
imarket share?How can DMD become the next printer business for HP?″ QuestiOns like these
SParked Spenner′ s entrepreneurial sPirit・ By 1991′ he was convincod that a new disk― drive
architecttlre、~
rith an innovative design cOuld take the computing inarket by stom and that HP、

ras
the company to create it.
DME)had cstablished itself in the 5.25-and 3.5-inch lrLarket but by 1992 had nOt inttoduced a 2.5-
inch dr􈎺 e ofthe sort used in notebook coml)uters.Speruler felt the competitors widlin that market一
PartiCularly Conner′ Quantum′ andツVestenl Digital――were too strong to attack directly and that to
succeed′ HP needed to go beyond any existing architecture.He behoved in Hackborn′ s favOrite
maxirn:″ Never take a forti■ ed hill`″
Thus′ Spenner wanted to attack an entirely new hill.He en、アisioncd tl■ e future of stOrage in the
form of large data library servcrs fed and utilized by client colnPllterS(deSktOp and nOtebook PCs)
and′ ln thc future′ handheld computers. Handheld computers and other very sinall for■ ls of
computing represented an omerging rnarket for、へ′hich Dゝ41D could lllake a suitably small disk drive.
Spenner had found his hill.He wantcd a disk drive that not only seⅣ cd the computing lnarkotplace
but transcended the traditional nlarket boundaries and could be used in any product that used a
nllcroprocessor.
To receive aPPrOval tO lnitiate the prc)ject′ Spenner decided to presont the idea directly to
Hackbom.He assigned one of his engineers′ Ceorge lDrennan′ to scope out different design concepts.
Drennan reported back、vith several differently sized rectangular boxes′ each one representing a
possible choice for HP′ s ne恥′drive.The largest represented a 2.5-inch drive′ the smallest l.3 inches.
la7。r lunch in early 1991′ SPenner placed in frollt of Hackborn the differently sized boxes′ exPlained
his vision of a neッv disk drive′ and asked′

WVell′ ンvhich one shall it be?″ Hackborn looked thern over
and′ to Spenner′ s surp􈎸 se′ picked up the l.3-inch box and said′

Do this one.″ This approval Ⅵ/as all
that Spenner needed.″Hackborn had such respect wvithin the HP organization′

stated one engillcer′
メrthat once he said ′IDo something/ evcryone seemed to fall into lhe.″ Soon after Hackborll had
approved the project′ so too did Smelek.At the tiine′ the Computer PrOducts C)rganization had iust
rePOrted record earnings for fisca1 1990.Hackborn and Smelek agreed that D􈎺 llD could afford the
Hewlett‐ Packard:The Fllghl of the Kittyhawk(A)
financial risks of the Kittyhawk.ヽ〔uch of the necessary investFnent Could be covered by prOfits frOm
the division′ s one― andれ、「o―gigabyte products.
動θ PrψεオTθα

DesPite Kittyhawk′ s heavンWeight endorsenlents′ DMD′ s functional management and most ofthe
R&I)section managers hesitated to suPPort the l.3‐ inch drive.They felt that the investlnent ln a ncw′
slnall architecture would conflict with the neods of the division to stay atop its cstablished lrtarkets.
In the race for inultigigabyte drives′ Seagate and PIIaxtor were neck and neck with Dヽ1]D′ and the
industヮ′
s v01ume leaders′ Quantum and Collner′ were nipping at their heels.They fclt that the
division′ s priorities should be to its next― generation higher― gigabyte product lines′ not to a tiny drive


rhose market was yet unclear.
Speruler had expected this reactiOn and moved to separate the l.3-il■ch proieCtfrom tho rest of he
division.Accordin31y′ the prOieCt team moved operatiOns out of DMD′ s main building into trailers
located at a remote corner of the division site.:3penner considered Kittyhawk the division′ s highest―
priority proiect and afforded the team the power to make timely deCiSiOns.One team member viewed
Kittyhawk as an engineer′ s dream proiect:″We were basically a start― uP business with the speed and
■exibility of entrepreneurs but with alsO the financial and technical backing of a successful high― tech
company.″ The proiect alSO rccoived execut􈎺 e support frOm the top ranks of HP.Hackborn and Lew
Platt′ HP′sCEO′ often visited the prolect trailers to see ho、、アthe development was progresslng.
Spenner wanted to n■ake sure that the Kittyha、vkぃras not governod by the division′ s traditional
development prOcesses.In order to sPeed Kittyhawk to the market′ he gave thc Kittyhawk team
autonomy to develop the drive′ find new markets′ and cultivate a customcr base.
rrhe core project team′
formed in May of 1991′ contained three functional representat􈎺 es
(manufacturing′ markeing′ and R&D)with a prOgram managerfrom R&D′ Seyinour′ as the leader.It
恥/as not hard to fill these positions.SPenner looked for risk takers、、アho、″ould be rrLOrO eXcited by the
market potential of a l.3-inch drive than by its technological capabilities.These core n■ embers were
not necessarily experienced in develoPing neν7 architectures or cultivating emerging markets but
、、アore considered to be″ can do″ people.艶yn■ o■・r had been an R&I)section rnanager for E)卜􈎽D with a
manufacturing background ln disk― drive heads and media. Although having ncver led the
development of a ne■r architecture′ he had the reputation for quick thinking and action that Spenner
believed necessary to make the prOieCt Succeed.Jeff White′ the marketing manager′ had iOined HP a
few years earlier with anヽ4BA and had a sirnilal・reputation.
As SperLner had with the core of the team′ lhe Kittyhawk′ s managers carehlly chose their staff.
Although recnliting from other I― IP divisions as、、アell′ they mOstly selected exceptional employees
from within DMD.′′
Because of the priority of our proiect′ if We wanted someol■ e from the division′ s
next― generation m′o―gigabyte prolect′

Said Seyrnour′

へ″e got hiFn.No questions asked.″
The core tean■ was wary ofteam members wvho、ぼould bring、、アith them HP′ s cultural biases.To
reinforce how differently the team needed to work′ David Woito′ the proiect′ SR&D manager′
required a■ engineers to sign a creed before they could,oin the Kittyhawk team:″ I αP7r gοれg Fο bι ιガ″′
sttα lι′′′解b′ ε力

′′isた′″

θ′
′′
Two engineers would not sign the staternent and returned to the IIP
mainstrealn.
To ensure that the team functioned wvell′ the core team extensively researched team dynamics and
group development literature,When setting tlp their いアork areas′ team members who had to
coordinate together had tlleir desks next to one another.
′′
Our organization l、アas a state― of― the‐art
teanl′

boastod a rnember.
Hewlett‐ Packard:’rhe Flight Ofthe Kittyhawk(A)
TJz′ Pro/θσ′P′ //7″ zθ た/s
SPenner drafted a proieCt Charter for Kittyhawk′ which compFiSed five goals:
1. Introduce the Kittyhaぃrk in 12 1nonths′ from start to fillish.
2.AccomPliSh a break― even time(BE■ ')of less than 36 nlontl■ s(see Exhibit 5).BET was the time
ittook to repay the ncgative cash flow incurred in develoPing and launching the product.
3. Ac・hieve a S100 1nillion revenue rato in tlvo years after launch.
4. Be the first l.3-inch drive on the rnarket― ――′′
the first on a ne、ν hill.″
5. Gro、、アfaster than the disk― drive lnarket to helP IIP beCOme a significantindustry leader.Thus′
revenue gro􈎺vth rate had to bO aroulld 35,3.
Although aggressivc/SpeFuner′ s charter did nOt aPpear tO bC Out of reach.HP′ s average cycle time
for ne、アdisk―drive platforrn develoPment` れ′as 18 rnonths.Because they could leverage off technology
that DMD had been deve10ping for its larger drives′ thc prtDieCt team believed that they could attain
the 12-rnonth introduction date.In addition′ althOugh three tinles higher than had beon Origillally
fOFCCaSt′ the S100 million rcvenue rate v、′as thoughtto be PosSible by focusing on alnd cornering high―
growth market areas.
Fjれ力776a rみ
‘Kj′
ryた′復ノたんイ″ た
`′
One week after tl■ e start of the prolect in J■ lne Of 1991′ Seymour andいhite arrived at the
Consumer Electronics Show in Chicago to took at sorntD rLarket possibilities for the Kittyhawk.They
ignored the desktop and notebook computer sections of the shoぃr′ fortified hills that theyぃ′eror and
searched for the neぃアcst mobile computing products一handheld and Pen― based computers′ othemisc
known as personal digital assistants(PDAs).Although still in the fledgling stages of development′
PDAs offered some interesting possibilities for a l.3-inch drive.
After intervie恥アing scveral leading PD/、developers in the mobilc computing section/Seymour
and Wl■ ite came upon a bOoth the size of about eight or nille other booths combined― the lヽintel■ do
exhibitiOn,Therc they found a labyrinth of interactive gamcs and turned to each other、、Zith the same
awestruck reaction:′

Look at all the stOrage Possibilities here!L7ve COuld fit 50 ofthese games onto one
Kittyhawk.″ They pulled the lゞintendo rnarketing managcr aside and asked hiFn if Nintendo might
be interested in a new small stOrage device for its game cartridges.The lゞintendo lnanager replied′

Absolutely′ the soft■ vare writers′ dream is to have more cheap storage.We′ re al■ vays looking to
create more comPleX games.″ IIe then ernPilasized that the irnPerativeぃ′ord hereぃアas″ cheaP/′ more
accurately about S50.As they left/Seymour asked′

Hol~I many ofthese game cartridges do yotl shiP
per year?″

ヽVell′ to give you an idea′

th4〕Inanager resPonded′

in the(=hris廿1las soason、ve ship
aboutl.5 rnillion per day.″
After returnmg to Boise′ いhite continued market researtth for the proleCt・He read research
rePOrtS On ne、アmarkets in electronics and contacted companies to explore their ilturc product plans.
White also talked to many people lvithir■ HP itself in his search for insight about wherc the
electronics industFy waS going to exP10(le,White compiled a list of five possibilitiesl mobile
information technologies′ cornmunications technologies′ constliner electronics′ alltomotive
electronics′ and some new developmentsin standard computer technology.
After considerable deliberatioll over tltese target markets′ the Kittyhawk team narroぃイed the
strategy for the KitサhaWk dOWn to two pOssibilities/either a disk drive speciAcally focused at the
Hewlett‐ Packard:The Flight of the Klttyhawk(A)
mobile computing lnarket or a drive so inexpensive that it could be used in aPPliCations hアhere disk
drives previously had not been econonlically fcasible.The team struggled bet、、アeen these h、「o chOices.
Seymour andンVhite had not forgotten their experience in the Nintendo exhibition′ certall■ly not the

1.5 milliOn″ statistic.But at the same tirno′ th(,y also believed that a S50 disk drive′ by itscli lYlight
nOt SPawn a large rrtarket fast enOugh to achieve Spenner′ s desired break― even tirne.The 10ぃァest unit
cost that had been achieved thus far in the industry for a fully featttred disk drive of any capacity was
about$130.For the industl,′ aS a whole′ this S130 seemed to have acted as a cost floor ofsorts― ―it cost
that inuch to purchase and assemble the basic components.Designers、vere able to reduce the cost
per Fnegabyte by persistently increasing the】Υtegabytes per drive.But the cost′ θr′ γjυ′Seemed to
stubbornly remain above S130. Henceノ deSigrling a S50 drive w、7ould require a significant design
breakthrough.
􈎺16bile computing markets wOuld require brcakthrough techn01ogies of a very differcnt sort―in
PartiCular′ the ability to PaCk lnOre rnegabytes ofinfOrrnation per square inch Of disk surface than had
even been done in a small drive.ButifthOy cOuld do it′ the lnarket seemed attractive.Every company
that、、ras developing a PDA showed intense interestl・1/heneveF VVhite or someone fronl his lnarketing
staff asked if they cOuld■lse a smaller disk drive.ツVhite noted′
′′
〕νlobile computing v、Zas still in its
lnfancy when the Kittyhawk、vas being developed.Everyone lvho knew something about technology
thought that PDAsぃ′ould be the next biggest thil■ g to hit the markct.″ Companies such as Apple′
IBヽ4′ ヽ4otorola′ AT&T′ and even HP 、verc investing hundrcds of millions of dollars in tlle
deve10pment oftheir own PDAs.
Because the rnobile computing rnarket volul■ e seell■ ed nearer at hand′ the team decided that the
best strategy would be to start by designing a drive suited for the nlobile computing nlarket and then
eventually′ through high― voltlrne production′ reach the S50 price point through volunle
manufacturing and next― generation ProduCt redesign.Fronl their beachhead in mobile computing′
the tearn imagined a disk drive that would be utilized in all sorts of consumer electronics such as
Nintendo game cartridges and cellular phones.They salヽ′it uscd as a″ super′

floppy diskノ予Vhere
computer users could carry the Kittyhalt7k arOund in their pocketsぃアith prograrns already loaded
onto it and be able to sit down at Public― access computers′ plug in the l,3-inch driver and begin using
the applications that they broughtル、アith them.The team reveled in the possibilities.
Seymour subsequently presentcd a three― pこige strategy documcnt to Spenner.On each Page′ in
bold lettering′ was Printed one obleCt􈎺 e:
● Lead industry in l.3-inch forin factor.
o Ride the rnobile computing explosion to get to loぃr cost.
。 ″
I′ u sell you a drive for s49.95.″
Seymour explained the sirnPliCity behind these objectives:″ Thisぃras not a 20-Page Strategy that
you would ordinarily see with a proieCt Of this rnagnitude.Instead′ we lust wanted peoPle to start to
irnagine the ilnrnense possibilities of the Kittyhat、ヽrk.′ ′
To reassure Spenner that FnObile computing represented the right high― grolへ′th market for the
Kittyhawk′ the team contracted with a highly reputable market resoarch firm that specialized in high―
tech rrlarkets to independently gauge thc magrtitude of Kittyhawk′ s opporttu■ity.The firll■ りTiCally
would talk to existing customers and industry experts to analyze、、′here the market wvas headed.
However′ in this rnarkct′ when itw、´as not yet clear、、アho the larger customers、、アould ultimately be′ the
firn■ found that its normal rnethOdologies led it no、

′here.Vヽ「hite recalled′

It was liko trying to learn

Hewictt‐ Packard:The Flight of the Kittyhawk(A)
Swahili without the help of anyone else、、ィho knoぃ7s the language.The research firllll ended up talking
to us rnore than anyone else.Naturally′ they came to bclieve whatぃアe believed.″
As White worked、4th budding PDA nlakers′ solne of the clearest early inPut came from DaytOn
Electronics Corporation′ a leading comput,ar maker that was developing a pen― based computer to be
used as an electronic cliPboard by delivery personnel in ovemight Package delivery companies.
Dayton′ s lead customer had a specification that the computer had to be able to、″ithstand a thFee― f00t
drop onto concrete.At the tilne′ the avcrage hard disk drive couldぃ4thstand a droP′ ぃアhile operating′
of about three inchesマvithout data lo6s.
C)ther of the Kittyhawk′ s design specifications l~ 7ere taken iけom HP/s Con/allis Division.ンVhercas
most notebook and subnotebook comput(trs employed 2.5… inch drives′ tlle CoⅣ allis Divisiol■ was
designing a″ super″ subnotebook that wvould not have sPace fOr a 2.5-inch drive.As a result′ the
Kittyhawk team worked c10sely with tlle CoⅣallis group to meet their operating requirements一
PartiCularly′ lowぃreight and lovv POw、ィer c01lsumption.
Through their wOrk with PDA developers′ tl■ e Kittyhawk toam developed a view abotlt which Of
the PDA developers would succeed.They felt that those handheld computers that addressed an
applicatiOn― specific niche il■ the market―一such as Dayton′ s PDA for package dolivory一would be
more likely to succeed becauso of the focused functionality they required.Portable check― in devices
for car rental collnpanies had similar characieristics,
TItθ Cθ

′θF′ιθγ5
Seymour and White felt thatれ、70 teChnologies lnight compete with Kittyha、アk.The first was flash
memory′ a nonvolatile integrated memory circuit that retained inforrltation stored on it even if the
pow、アer乃げas turned off.Because it had no m()ving Parts/11ash■ 、ras exceptionally rugged.Although the
flash chiP cOminanded about s50 pertnegal)yte′ 10 t􈎺 nes rnore cxpensive than the average disk drive′
it becalne competitiveぃrith disk rnemory at the small end,Six lnegabytes of iash lnemory cost about
S300.Because of the apparent unit― cost floor for disk drives discussed abovc′ this lneant that for very
low― capacity drives′ nash rrLemOry would be morc closely cost competitive than for higher― caPaCity
drives,The Kittyhawk engineers consider(:d the dynalnics of this competition v、「hen decidhg that
their l.3-inch drivc sho■ lld have atleast 20 n■ egabytes of storage.
The second POtential competitor、、アas the l.8-inch disk drive,Although not yet available vvhen the
Kittyhawk proiect Started′ industry sour(les believed that several other companies planned to
introduce l.8-inch rnodels in 1992.I‐ Iistorically′ in the progression frorr1 8-inch to 5,25-inch′ 3.5-inch to
2.5-inch disks′ the surface area per disk in each newv architect■ lre was half of the preceding
generatiOn′ s surface area(Exhibit 6 shoⅥ ′s ths Pattern).HenCe′ a l.8-inch drive would be the
industry′ s lnost predictable next step after the 2.5-inch drive,the Kittyhawk essentially leapfrogged
ahead one generation beyond the l.8-inch forn■ factor.Seymour tcnded to discount the l.8-inch
threat′ however. Though it would have greater capacity′ it ivould be larger and consune morc
po、げer.
Tttθ Pγθグタε
`
Seymour deliberately workcd to instill fear amongst his colleagues:″ I、νanted the team to imaginc
that everyone in the industrv was 30ing to beat us to the punch.″ Integral Peripherals′ introduction of
the first l.8-inch disk dr􈎺 e fOr subnotebot,k computers in September 1991 sPurred the KitサhaWk
group to rnOve even faster.′
′We tradcd most eveFything to rnoetthe schedule:PorfOnnance′ features′
Hewlett‐ Packard:・rhe Flight of the Kittyhawk(A)
cost―一everything but reliability′
″said Scynlour.Though Seymour andツVhite felt they prObably had a
12-rnonth lead on tlle competition′ they told everyone SiX rnonths tO elevatclthe sense of urgency.
Designing the Kittyhawk to lneet the key l)erforrLanCe mandates White′ s market research had
identified一一the three― foot droP and 20 MB of capacity′ in particular― ―required that three unique
techno10gies be deve10ped.The firstぃ′as a neh7 stlbstrato material for the disks.E)isks in larger drives
generally collsisted of polished Platters of altttninum that lvere coated M′ ith thin films of rnagnetic
metal.To meet the height requirement of the Kittyhawk′ these disks needed to be reduced tO the
thickness of foil一making aluminum an unsuitably weak substratt・.The prolcCt team developed′ with
a supplior′ a glass substrato that was thin but strong eno■ lgh.lt could be PoliShed so nat and smOoth
that heads could fly closer to the disk′ a1lo、vilrlg data to be PackOd moro densely.Teanl llnembers
believed that this disk technology′ combined ぃrith other custom conlponentsノ `

″Ould a1loぃア the
Kittyhawk to reach up to 200ヽ4B of capc■ city by 1995。
The secOnd technology、

ras a newv level of integration for the Kittyhawk′ s electronics.Fortunatelyノ
since 1989 a group of D􈎺(D engincers had becn、″orkmg on the problern of rnanaging the drive′ s
operations and computer interface、vithin a rntich smaller number of custOm―designed integrated
circuits.1年hile a typical l.8-inch disk drive had 20 1o 30 chips′ the Kittyha、「k team mtegrated even
bctter flmctionalitb/On only 5 chiPs・ThiS meant that the l.3-inch modtlle would use less power′ be
lightcr′ and be rnanufacttarable at lo、ャer cost.
The key to nleeting the three― fOot drop requiremcnt centered on a proprietaり′ SiX… axis
piezoelectric accelerometer.This、vas a shock―seilsing inechanism that could detectin■ pending inlPaCt
on both lillear and rotational clxes and caused the drive to revert to a mode that protected against
data loss……aCting much like an airbag collisio】l sensor on an automobile. Seymour described his
lnitial reaction to the innovation:″ The technologyン、ras amaz蛇・An elegant design all around.The
only problem was thatthis cOmponent alone cOst over S10 to rnake.Btlt′ man′ was it cool.″
The proiect team decided not to manufacture their drive in―house.They looked instead for an
extemal supplier with proven expertise in milliaturized manufacturing and found a perfect match in
Japan′ S Citizen Watch Corporation.Citizen designed and built an autolnated production line fOr the
KitサhaWk・Prepared for future growth′ the lhe had a capaciサof 150′000 units Per month・
The Kittyhawk was introduced right on sclledulo in June Of 1992ノeXaCtly 12 months from the
begirlning of develoPnlent.Although retired and not having been Present at a product launch in 10
ycars′ ヽ

rilliam Hewlett′ HP′ s cofounder′ PlヽeSided at the press conference announcil■ g the
Kittyhawk′ s iatlnch.It measured O.4 il■ ches by 2.O inches by l,44 mches and wcighcd about l o■ ll■ce.
The Kittyhawk was almost halfthe total sLe anct one― third the、veight of the l.8-inch disk drives that
had been introduced to market iust rnOnths bebre.The Kittyhawk announcement garnered more
press coverage than any ne、~r―product announcement in the history of Heヽ~
rlctt―Packard.The design
won several Prestigious technology and new― product a、~
アards for 1992.CEC)Platt developed the habit
Of carrying a Kittyhawk in his pocket as a convorsation Piece for customers and analysts.
艶ymour、vas not sure ho􈎺 v hc felt about all the attention:″ The great news was that、

′e、~
rere in a
project with a lot of 􈎼sibility.That was also the bad news.We had llnbelievable supPort.If We made
the Kittyhawk lly′ it、vould fly high′ but ifit crashed′ there、vas gOil■ g to be ono he1l of an exPlosiOn.″
HP shiPped its first Kittyhawk on June 23′ with high― volume(DEM Pricing at about S250.Based on
their read of the market′ the Kittyhawk′ s marketing staff was proiecting the next two― year demand
fronl the PE)A market to be over 500′000 units. At this pricing′ Kittyhalヽ/k looked as it if would
achieve Spenner′ s obieCtiVes both for revonue rate and break― even time.
Hewlett‐ Packard:The Flight ofthe Kittyhawk(A)
T77θ C夕s′θ7″θ

By July Of 1992′ the Kittyhawk team had design wins il■ new PDAs beillg developed at six
computer cOmpanies.It appeared to be besting botll flash inemory and the l.8-inch drive fOr the rnOst
attractive applications.Tl■ e teanl also got designマvins in Other applications.Ol■ e examPle`Vas with a
COmPany that utilized the drive as backuP StOFage in the portable check― in devicos it inade for car
rental companies.The delmand for the Kiltyhawk looked to be on target with the prolect′ S gOals.
ツVhat was rnore′ many of these custOmers at the tirne、へ7ere COnsidered to be the bltlest of blue― chip
companies― ―Apple′ IBヽt and even HP itselJ.
In its third inonth of production′ howcver′ the prolcCt hit a road bunlP,IIP/s CoA/allis Division
decided that a l.3-inch drive vvould not be able to rncet the Future storage retluirements of its super―
subnotebook computers.Even though tho Kittyhawk team had Plans tO introduce a sccond―
gcneration drive l、アith 40 Fnegabytes′ that would still not be cnough.Conアallis instead decided to use
lntegral Peripherals′ 1.8-inch drive.Kittyhawk had lost its first llaalor pOtential custonler.
An even bigger road bump、、′as thatthe PDA market never emerged as exPected.FOr nlost of the
PDAsノKitサhaWk′ S perfornlancc was more than stlfficient.But othor new technologics uPOn WhiCh
the PE)As depended′ such as handマvriting―recognition softvvare and nelv integrated circuitry/Proved
to be inadequate.Literally every PDA mallufacturor found its PDA sales to be disappointing′ and
mostぃrithdre、アfroln the inarket.The l■ andl:eld computers that、vere conlrnercially successA■ l tended
to be technologically mOdest′ lol~′er― priced dcvicesぃアhose needs for storago could bc cconorflically
satisfied by flash-lnemoゥr chips.
But there was never an unanlbiguous′ definitive signal to the tealn that the PDA market、、ァould
not materialize.It seemed that for every cl■ stolner that stulnbled′ stlch as APPle、`′ith its Iヽeぃ/ton′
another rePutable′ technologically competellt customer stoPPed uP、Vith a different type of PDノヽthat
it was confident l、7ould hit the right price and Perforlllance polnts u■ the lnarket.IIence′ though
Kittyhawk′ s volume ramp was delayed′ ProsperiりalWays seemed right arOurld the corner.For
example′ just as HP annO‐■ced the Kittyhawk H′ witl■ 43 MB oF stOrage and the abiliサtO Witl■ stand
10%greater shock and consume 25%less Power thall thtS KitサhaWk I′ a rnalor COmputer company′
Chicago Controls′ designed Kittyhal、′k into an 􈎹dustrial pen― based device for recording and
analyzhg data for statistical process control in nlanufacttlring plallts.KitけhaWk′ S ruggo(hoss was
attractive because stored dataぃ′ould remail■ intact desPite any ro■ lgh handling that might occur On
the factory lloor.Microsoft、、「as creating a vcrsion ofits PC oPerating systemゃvith graphical interface
for this PartiCular PDA.The requirements lvere that the operating systern had to be able to fit On the
soon― to― be―introduced 40-megabytc Kittyhawk and still leave enough capacity for Other programs.
VVhite′ so、ヘアn research、、アith end uscrs supPc,rted Chicago Contr01s′ enthusiasnl for this product,if it
took oft it alone、へ7ould catapult Kittvhalvk l)ack on Plan.
To bolster HP′ s comn■itlnent to Kittyhawk′ Seymour brokered a meeting beれveen Spenner′
Smelek′ and Chicago Contr01s′ CEOr and it worked.Tlley left reassured with the KitwhaWkノs
ipotentialo Shortly after the rneeting′ hoぃアever′ ヽ4icrosoft annoしlnced that its operating system for this
apPlication would need rnore than the 43 1ヽ43 of storage the Kittyhawk offered.This derailed the
whOle cOncept ofthis PDA for factoゥr COntrol.
Even while prosperity kept loo■ ling aro■lnd the next corner in nlobile coIIllputing′ several
linexpected customers started to show lnterest in IIP′ sl.3-inch drive.The first was a Japanese
company whose portable word processor pr􈎹 ted Japanese K/7″ ノi(ChheSC Characters)when wOrds
、vere sPelled On its keyboard.The device was used in both the home and officc.It requircd a stOrage
device that not only had cnOugh capacity to store the necessa恥′transiation PrOgrams but also could
l″ithstand the shocks tO which POrtable devices、vere exPosed.The Kittyhaゃ、rk′s shock resistance also
Hewlett~ Packard:The Flight of tlle Kittyhawk(A)
attracted manufacturers of cash registers.Nlost Of the newest cash registers 􈎺vere run through an
operating systom that recOrded daily transactions vvithil■ a ccntral conlpllter systein, The
manufacturers、vere looking for a stOrage dl・iv(〕that could act as a backup if the centl・al computer
failed while also being able tO、、ァithstalld the shOck from the slamnling shut of the cash register
dral● 7er.The 20-inegabyte Kittyhal、′k seⅣ ed the necds Of both these aPplicatiOns Perl‐ t・Ctly.
Another customer sa、ヽ′the Kittvha􈎺vk as a″ filn■ cartridge″ that could be removed from the digital
camera it was develoPing and inserted into v"ン、ring and printing devices.The Kittyha、vk′ s shock
resistanceン、アas PartiCularly attractive for this aPPlicatiOn. The camera′ s success dcpended on hν o
enabling techn01ogies.The first、、アas rugged stor;ige.´ 、lthough flash ll■ emory could solve this Piecc Of
the puzzle′ the KitサhaWk Served the need better due to its loぃたer cost per rllegabyto.The second
technology was a charge― coupled device(CCD)tl■ at transformed images into digitalおrmat.The
CCD technology tlltiinately proved techllically feasible′ but tinfortunately its cost pushed the
camera′ s retail price to Sl′ 500ノいアhere tulit volunlesャ、rerc disappointing.
As a result′ the list Of KitサhaWk CustOmeFS(3hoWn in Exhibit 7)was very diffel’ent缶Om wilatthe
teanl had origil■ ally planned.
A氏眠(夕Bcg:η 77ブルlg2
E)esPite this string of disaPPOintlments PepperCd by a fcゃ、′successes from unexPeCted quarters′
lnterest in the KitサhaWk kept rolling in toヽ

イhite′ s nlarketing tcam一―inteFCSt from highly crediblc
companies、

rith solid ne、ν product ideas.'「hese scjcmed to fall into t、νo groups.The first、ヽ/cre those
described above for which the Kittyhavvk′ s ruggcdness based upon its accclcrorrteter techllologyン、アas
its lnOst PriZed attribute.The second grollp of potential ctlstomers、、アas sounding a very different
then■ e:they nceded a cheaP′ sirnPIC drive priced at around$50.
′′
It took a couPle Of years of the Kittyhawk being in the marketplace before peoplc figLlred Out
、νhat they needed′ ″
VVllite recalled.″ BefOre the Kittyhaッvk′ Inost of our customers never even
thought abOut disk storage as a、、′ay to imprOve their prodtlcts.The Kittyha■ vk got theFn thinkil■ g′
and thcn they started conting to us、、アith νvhat they reallyゃ、アanted.〃lゞintendO/for examPlc′ ShOや、red
the Kittyhawk teanl an entertainmont systenl they had already designed′ tlnbekrio、、7nSttO HP′ vvitll a
slot br a Kittyhawk drive tO bo plugged in― → mOdule that could cOntalll many more gaΠ les′ with
much n■ ore sophisticated graphics/than lゞinten(lo′ sぐonvcntional cartridge aPProach・
′′
The systein is
all set/′ the lヾintendo representative persuaded.″ All you have to do is sell tls your drivc f()r s50.″
When White protested that the Kittyhawk′ s accelerometer alone(required for shock resistance)
rendered the s50 price impossible′ 1ヾintendo re〔;Ponded that they did nOt need the acceleromcter― ―
they iust nOedcd 20 megabytes at S50′ cheap and simPle・
A fax― ll■ achine manufacturor exPlored with the tcam the possibiliりof offering an cmbedded
Kittyhawk as an option in its high― end nlodel t()handle graphics― il■tensive transrnission

I learned
PrettV early on that you don′ t、、アantto be dcsigned in as an optionノ
′ノ
white recalled.″ You、vantto be
paFt Of their standard ProduCt′ Or you have no、、ray of forecastulg、、「hat kind of volume yotl′ re
looking at.″ White polnted at olllc Of the custon■ cr′ s loッver―PriCed fax machines and asked ho、、アrnany
of thosc they sold.To the ans、ver of 5 inillion units per year″ Seymour rcsPonded′

That/s interesting.
Could you use the Kittyhavvk for those?″ The(lustOmer revic、ved his IItaterials list and ansゃ、アered′
′′
Sure′ as long as you can deliver the drive to tts for S48.〃″How about ifン、「e hit S100?″ Seymour
qucried. The custolner shook his head and saidァ
プ′
If you hit S52シ、アe stili couldll′ t use it.″ ンVhite
recalled 10 other comPanies 􈎺vith sinlilar PrOPOSitiOns at the S50 prico poillt.
10
Ilewlett‐ Packard:The Flight ofthe Kittyhawk(A)
錠s翼濯TTlel露導ご翼¶ £ 鼻I鵠
覆ぬe職りhWk∞ rts tt_3m〕ねttetthrmheぬdr .es oinerged frort the disctlssiOll.The First、
vas to
署育:∬視撚h甘ぶ器t殿:′鷲ム糧
::貯T:llillll::′脚i胞1よ∫尋聡i「il!檸19:lil還
〕d intercst from HP Inanagenlent.
M∬緊1暇懺翼PI肌鍔1蠣n飢」脚il響

Ъ
蟹」箇iPtti驚有ξ盤繁猟堪lI ~習
:鷺i:RttFL∬胤鳥壁
ih穏i:翼1脚農llltth∫:1』l:l::∫ lЪttT配著詭1
lClrket earnillg enougll tf a F
􈏛轄喜≒構鱗職:駐量e鮮選メ臨􈉍苗轟􈉍″織i鳳搬
o it.Designing a 40 MB l.3-inch drivc that cOuld
drop three feet ontO a concrete f100r vvas nO si:I:liili[[lfサ
her・ The s50 drivc wOuict bc a very
different challenge′ butit probablyルvouldn′ l be n10
賊慮穏翼e璧亀鳥i謂織:∬憲廿rtr機迭Iw穐淵為X漁管綸1∫ 渤驚な?還i
having that decisiOn lnade for hinl by Spenner or his boss.
Hewlett‐ Packard:The Flight Ofthe Kittyhawk(A)
Exhibit l The Hewiett― Packard l.3-inch Kittyhaぃアk Disk Drivc
So■lrce: Hewlett― Packttd.
12
606‐ 088 -13‐
Exhibit 2 HP′ s 􈎺lass StOragc CrOup Or3anization Chart(1992)
President ind CE0
・ι
‘″
IPra″ ‐

rcst&、4easllrelllcnt C)rg.
COm,Iter PⅢ ldlltS OrganiZa,Om
‐ I D′6″ 劇隔
`驚
bθrrr l
Conlputcr Systcms Org. Mcasurelllent Systcms(Dr8.
Printing Systcr1ls Personal I11ねril〕ation Products Ink― Jet Products Salcs&3 Suppo員
M■ss Sloragl GrOup
.R″ S税むル■ |
Boisc Site Opcrations
Computer Pcl‐ iphcrals BI・istol Div.
Co10rado Mclllory S、:stems
Grecley Storage Divisioll
DisヒMemprylpiViSi。.
Sottrce: (lasc、γritcr
Exhibit 3 HP CorPoratC卜」et Revontle versus Disk‐ Drive Revenue′ 198〔卜1992(fiscal year ends October 31,$millions)
1986 1987 1988 1989 1990 1991
606-088 ‐14‐
1983 1984 1985 1992
Corporate Net Revenues
Disk‐Drive Revenues
Disk¨ Drive Revenues as a
Percentage of Corporate’s
4,710.0 6,044.0 6,505,0
269.1 315.5 259.1
5。7% 5.2%
7,102.0 8,090.0
251.3 328.1
9,831.0 11,899.0 13,200.0
420.6 533.4 402.2
3.0%
14,494.0 16,410.0
280.7 519.4
4.098 3.598 4.10/. 4.3・/. 4.5% 2.0% 3.2%
18000.0
16000.0
14000.0
12000.0
10000.0
8000.0
6000.0
4000.0
2000.0
0.0
1983 1985
窃c載〓E輸)3 ●co>o配
Corporate
1987
Disk Drivc
1989 1991
Sourccl Disk/Trend clnd Hc、vlett Pa(lkar(i corporatc reports.
Hewlett‐ Packard:The Flighl of the Kittyhawk(A)
Exhibit 4 A Comparison Of HP′ s Product Position′ Relative to C)thcr Leading Disk― Drive l■ lakers
(VertiCal lines indicate the caPaciサOf tlle medial、modelsold by Dヽ4D and PrinciPal
competitors′ compared to the industry.)
90
80
70
60
50
40
30
Conncr*and Quantum*
20
10
0

Foc■ is On deSktoP and notebOok Perl,01lal col■ ptlteri
t*Focus on engineerulg、vorkstations′ filtt sel■ rers′ and otller]arge coFnputers.
Source: 1992 Disk/・rrend reP()rt
0一OC〓“>イ、∽【●「0】と」O 破
〇一一‥一〇〇
〇〇寸‥【〇吟
〇〇﹇‥【CN
〇〇∞‥一一ト
〇〇卜‐一〇つ
〇〇Ci【C崎
→T COひマ一一”【
Φ一卜一‥一〇C 一
〇〇C 一‐一〇一一
CC寸】‥一〇〔【
Cal)ac
15
Hewlett‐ Packard:The Flight ofthe Kittyhawk(A)
Exhibit 5 Hewlett― Packard′ s 􈎺fethOd for Evaluating Proiect Success:The Break― Even Tirrle CalculatiOn
PrOJeCt MOnths
bご高品応
‘・
とお悦1」

L;:… 品苺よ􈎹品品`キ島もごdⅢ:だ島;i
This schcdllle was a stalldal・d HP Fo「mat for measurillg pFtteCt Success.
Sotlrcei Case、γriter
16
Hewlett‐ Packard:The Flight ofthe Kittyhawk(A)
Exhibit 6 The Pattern in Rcduction of Disk Surface in lヽe,v/Sma11_Architecttires
Surface Area= R2
Sourcei Casewriter.
5,25"Disk Drive
SoA.=22 Sq.In.
3.5"Disk Drive
S.A.=10 Sq.In。
2.5’'Disk Drive
S.A.=5 Sq.In`
1.8"Disk Driv
S.A.=2.5 Sq.In。
1.3"Disk Drivご
StA.=1.3 Sq.I11.
17
Exhibit 7 Planned versus Actual Production L,vels and Product APPlicatiOns
Plannedヽ7ersus lttctual Production Leveis
700
600
500
Kl■yha、vk
Uni協(in 400
thousands)
300
200
100
Planned Versus Actual】嘔ajor Product
Hewiett‐ Packard:rhe rlight of the Kittyhawk(A)
1994
″ofUnits Sold Over the
Pian: Actual: l`ife ofthe Project(000S):
・Pcrsonal Digital Assistants
OSub― notcbook
・Hard Copy Dc􈎼 ces
・Prillters
・Copicrs
・Fax 􈎺lachines
・Japancscムλlord Proccssors ‐………―――――‐―‐‐――‐‐ 100
0Pcrson〔tl Digital Assistants‐ ―……………………35
=􈉍
l憲鮮l_三llilIIIilillllili1 2:

´
rclcc。lllm.Sヽ、itching Sゝ,stenl s ‐―‐―‐――――‐……… 2
Sourrci Caseゝ:riter compiled■ on■ col■ pally ttoctlillcnts.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License