to develop technology strategy - forgottenbooks.com fileverses 3 4-3 6. 1 introd uc tion the...
TRANSCRIPT
to Develop Techno logy S trategy
Mart in # Read Anthony E . Gear
Ball l o l Co llege , O# ford
S SWP #35 60- 9 3
Teamworking to D evelop Techno logy S trategy
Martin # Read Anthony E . Gear
Bal l l o l College , O# ford
S SWP# 35 60- 9 3
Teamworking to Deve lop Technology Strategy
Mart in # . Read Anthony E . G earBa l l io l Co l lege , Oxford
gather a l l the food of thosegood years that come and thatfood sha l l be store to the l andaga inst the seven years o f famine
# oseph , G enes i s Ch . 4 1 ,
verses 3 4-3 6 .
1 Introduction
The strategy of an enterpr i se i s embod ied in its strategic
dec i s ions , which are bas ic to it because # they dea l with the
d irect ion of the enterpr i se , its future s i z e and pattern. o f outputs
and markets , they determine the kind of company wh ich has to
be managed and the k inds of ta lents which are ca l l ed for .#(Minkes
Techno logy Strategy i s concerned with dec i s ions to deve lop ,
exp lo it , or ma inta in the tota l ity of organ i z at iona l know-how in a
de f ined set of technologies . Re levant act ivit ies in th i s
connect ion are ident i fying and eva luat ing technologies in terms of
the ir potent ia l importance to overa l l strategy . The paper i s
concerned. with assess ing the feas ib i l ity# and s ign i f icance. of
var ious tabled techno logy deve lopment opt ions in a team sett ing .
The paper descr ibes a group dec i s ion support system ( GDSS ) known as
TEAMWORKER , des igned to fac i l itate forward p lann ing by a team , and
demonstrates its use by means of a case app l icat ion invo lving
techno logica l forecast ing for the t ime-frame 2 0 1 0 to 2 0 3 0 .
The l iterature on techno logy assessment and forecast ing i s l arge .
For revi ews o f management techn iques in these f i e lds see , for
examp le, # antsch B lake Souder [ 1 3 ] and Ford Some
techniques seek to make use o f the co l lect ive inte l l ect of an
assemb led group o f # experts # rather than that of a s ingl e # expert #,
by means of j uxtapos ing di sc ip l ines and exp er iences in order to
promote debate , creat ivity and intu it ion . The ph i losophy of thi s
paper i s based on the use o f co l l ect ive expert i se to assess future
technologica l poss ib i l it i es in a feedback ori ented approach , Gear
Computer based technology i s uti l i z ed to promote group
commun icat ion , and i s carefu l ly des igned to min imi z e de leter ious
ef fects o f con formity , dominat ion , ret icence , ind i sc ip l ine ,
deviat ion , reconditeness or vested interest .
The T EAMWORKER GDSS compr i ses hardware and software des igned for
the part icu lar group context and process , and i s descr ibed in a
var i ety of app l icat ions by G ear and Read [ 1 0 , In the
case descr ibed , the software ( somet imes known as # Groupware #
) i s
a imed at improving commun icat ion for technology assessment .
Mockl er and F in lay have recent ly reviewed var ious forms
o f groupware and GDSS for a id ing strategic management processes .
TEAMWORKER software is a imed at fac i l itat ing group di a l ogue
fo l lowing feedback of j udgements from group to individua l and from
individua l to ind ividua l . Thi s aspect of commun icat ion i s a l so
di scussed in
The software may be des igned around a lternat ive frameworks of
dec i s ion ana lys i s For example , [ 9 ] adopted a pa irwise
comparat ive framework , based on a two-leve l h ierarchy , in order to
rank a set of a lternat ive technologies to meet a de f ined need . A
cr it ica l survey of recent d eve lopments. in app l ied mu lt ip le cr iteri a
dec i s ion mak ing and the ir current status has been undertaken
Th i s work provides a usefu l input to the des ign of groupware , given
that it i s the group commun icat ion process wh ich i s of centra l
concern rather than mathemat ica l rout ines a lone .
In the case descr ibed in th i s paper , a s imp l e scor ing sca le was
made the bas i s of ind ividua l j udgements throughout the meet ing .
Th i s meant that rap id interact ions were easy , with the abi l ity to
repeat stages , formulate addit iona l items , and respond to new
requests # on the f ly #. ( see a l so [ 14 ] in the connect ion ) . For the
given purpose , techno logy assessment on long t ime-frames ,
ref inement of the techn ique i s not the ma in i ssue . As B lake [ 2 ]
has pointed out # # Over-ref inement of the system makes it poss ib le
to obscure i ssues with masses o f deta i l , in turn caus ing error in
forecast ing #. Comp lex approaches , whi le o f cons iderab le interest ,
may be tota l ly unsu itable for certa in bus ines s uses .
The under lying management technique behind the on-l ine system ,
TEAMWORKER , in app l icat ion to techno logy assessment i s GROUP
DE# P# I . Some of the advantages of GROUP DE# P# I over and above
standard DE# P# I are d i scussed by
A further important aspect to take into account in the des ign phase
of prepar ing process support software i s the cu ltura l sett ing and
h i story of the s ituat ion , together with the sk i l l s and backgrounds
of the part ic ipants . For examp le , [ 1 6 ] has descr ibed the # Emerg ing
Technol ogy Roadmap # framework of th ink ing , wh ich i s updated by a
sma l l committee of ind ividua l s with in Motoro la .
Each o f the above i ssues needs to be born in mind in order to
des ign a framework for the group process wh ich can form the bas i s
of group support software . I n Sect ion 3 , a case study i s descr ibed
in order to i l lustrate des ign cons iderat ions . The group dec i s ion
support system ( G DSS ) used in th i s app l icat ion i s descr ibed in
Sect ion 2 , and some conc lus ions are ident i f ied in Sect ion 4 .
The system compr i ses a set of hand-he ld devices , one for each group
member . D ig ita l s igna l s from each un it are transmitted to a remote
rece iver , and fed into a tmic ro-proc essoru The s igna l s are ana lyz ed
in accordance with the software wh ich i s be ing used . Processed
informat ion i s d i sp l ayed on a large screen vi s ib le to the group .
The screen i s used to
request informat ion from each group member dur ing the
meet ing .
di sp lay processed feedback informat ion at 21
number of stages .
provide expert advice and menu Opt ions dur ing
the meet ing .
The arrangement of the system of feedback i s shown in F igure 1 .
Certa in j udgmenta l processes are espec ia l ly usefu l , depend ing on
the i ssues invo lved . Some important ones are itemi z ed be low #
Vot ing
Uses inc lude se lect ing an item or opt ion from a set ,
and mu lt ip le choice quest ions .
( i i ) Scoring
Uses inc lude eva luat ing each of a set of items or
opt ions on a pre-def ined sca le , scoring as a means of
produc ing l i sts in rank order , scor ing as a means of
qu ick ly f i lter ing long l i sts to short-l i sts for c loser
attent ion .
( i i i )
Uses inc lude we ight ing the re lat ive importance of a set
of cr iter ia by means o f pa irwi se compari son , eva luat ing
subj ect ive factors and parameters , assess ing subj ect ive
probabi l it ies and r i sks .
Uses inc lude the est imat ion o f parameters ,
probab i l it ies , r i sks and other factors by d irect and
subj ect ive ly based input va lues .
The system des ign fac i l itates communicat ion at a meet ing in support
of group work on four re lated act ivit i es
poo l ing o f informat ion , expert i se , op in ions and
be l ie f s .
( i i ) c o m p a r i s o n a n d d e b a t e o f a r e a s o f
ag reement / d isag r eement .
( i i i ) ident i f icat ion of key e lements of the dec i s ion task for
further group attent ion .
( iv ) revi s i on and rede f in it ion of the task , opt ions ,
criter ia,etc . as the group progresses .
3 . Ca se S tudy # T echno logy Assessment
B ackground # Th i s case study concerns the Research and Deve lopment
department of a large mult inat iona l o i l company . It i s known that
the future market for petro leum products i s uncerta in , owing to the
r i s ing costs of product ion and di scovery , and the increas ing
environmenta l concerns assoc i ated with the use o f petro leum
products .
These and other i ssues led the R&D department to cons ider the
sc ience and techno logies it shou ld be invo lved with by the year
2 0 1 0 . Thi s i s part of a cont inu ing process led by a # Core Team #
invo lving the study of two key aspects #
( a )# Bus iness Dr ivers # , i . e . Trends cu lminat ing in
opportunit ies or threats to the organ i z at ion
( b )# Bus iness Opt ions # , i . e . A set of proj ected
capab i l it ies wh ich would enable impl ementation
o f a spec i f ic techno logy for a de f ined
bus iness purpose at some future t ime per iod .
The organ i z at ion dec ided that part of th i s cont inuing process
shou ld be a three-day workshop invo lving sen ior execut ives from the
R&D funct ion and from a cross-sect ion of other bus iness funct ions .
A tota l of 4 2 sen ior execut ives took part in the workshop, 3 0 from
the R&D funct ion and 12 from other bus iness areas .
The purpose of the workshop #was to exchange views and. poo l
j udgements re l at ing to the bus iness opt ions be ing cons idered by the
organ i z at ion , and to provide gu idance to the Core Team on the h igh
grade opt ions , the options to drop , and the opt ions requir ing
further work . It was stressed that th i s workshop was not so much
a dec i s ion-mak ing forum , bur rather an opportun ity to exchange
views , eva luate and grade the opt ions .
Workshop Framework
The workshop mi ss ion was to de f ine a statement that the Core Team
cou ld use to he lp frame a Strateg ic P l an . In th i s respect , the
Core Team was seen as the # customer # of the workshop .
I n it ia l ly , a report on each of the key bus iness dr ivers was
presented and d i scussed . Th i s provided a framework for the
subsequent assessment of the bus iness opt ions . A fter th i s , the
fu l l set of bus iness opt ions generated by the Core Team over the
preced ing months was presented by the Core Team . .At thi s po int the
T EAMWORKER system was used as part o f a structured process ,
invo lving a combinat ion of d i scuss ion and po l l ing , to determine a
sub-set of up to 1 3 opt ions to be studied more intens ive ly at the
workshop,
and then to become the subj ect o f post-workshop
act ivit ies . Th i s was ca l led the # Screening Process #.
The next stage invo lved a number of sma l l teams each work ing on one
of the sub-set o f opt ions . Each team was organ i z ed to have at
least one de legate from a bus iness funct ion . Presentat ions were
then made by each team in a plenary sess ion , deta i l ing the
advantages and d i sadvantages of the bus iness opt ion studied by that
team .
Fo l lowing th i s , the T EAMWORKER system was used to score the sub-set
the opt ion would be feas ib l e in the t imesca le under cons iderat ion ,
and the re lat ive commerc ia l imp act that the opt ion wou ld have on
the organ i z at ion in th i s t imesca le , Th i s was known as the # Mapp ing
Process # . The output from th i s process cons i sted of a two
dimens iona l gr id showing the workshop # s eva luat ion of each of the
sub-set o f opt ions on the two cr iteria .
The outputs from these two processes were used to formul ate the
f ina l gu id ing statement to the Core Team . These processes are now
descr ibed in more deta i l be low .
The Screen ing Process
Each o f the set of 3 5 Bus iness Opt ions was presented by se l ected
part ic ipants . The deta i l s of these opt ions cannot be l i sted in
th i s paper , but inc luded items such as # Chemica l Veh ic les
Deve l opment #.
Fo l lowing the presentat ions , and some further c lar i fying
d i scuss ions , each workshop de legate used a 1 to 5 sca le to score
each of the bus iness opt ions . The bus iness opt ion cons idered of
most importance ( in the opinion of a part icu lar person ) was g iven
a 5 , and the opt ion cons idered o f least importance g iven a 1 . A l l
other opt ions were then given a unitary va lue of between 1 and 5
re l at ive to the individua l ly def ined most and l east important
opt ions . Thi s approach served to anchor the end po ints of the
subj ect ive scor ing sca l e of each part ic ipant independent ly .
TEAMWORKER was used to co l lect the scores and feedback summari z ed
informat ion to the workshop for d i scuss ion and debate . The f irst
feedback screen provided to the de legates , fo l lowing a part icu lar
opt ion vote i s shown in Fig . 2 . Th i s i s a h i stogram showing the
numbers of scores obta ined for each va lue on the 1-5 sca l e , the
overa l l average and standard error , and the leve l of conf idence
assoc i ated with statement that the sub-sets o f scores from those
R&D Funct ion members were s ign i f icant ly d i f ferent from those of the
Bus iness Function members . The average used was not a s imple
1 0
average of a l l 4 3 de legates , but the mean of the averages of the
two sub-groups . Th i s form of average was subsequent ly used to a id
determinat ion o f the opt ions to be cons idered at the next stage of
the workshop , and the Core Team did not want the larger number of
R&D de legates to havez a more than equa l ( quant itat ive ) inf luence in
the process . # i stograms of the type shown in F igure 2 c learly
showed how much di sagreement there was in the workshop for each
opt ion . ZU1 add it ion , the ca lcu lated conf idence leve l ind icated
whether there was substant ia l d i sagreement between the two groups
of members . A feedback screen was used in order to compare the
sets of R&D and Bus iness scores ( F ig . Each of these feedback
screens frequent ly led to further di scuss ion and debate ,
part icu l arly when s ign i f icant d i f ferences were revea led,
often
resu lt ing in rescoring of an option by the group .
When a l l the options had been cons idered , a # # igh-# ow # chart
showing the average score for each option , together with the
standard error , was presented to the group ( Fig . I nc luded with
thi s chart was a # grey area # wh ich dep icted the z one where the cut
of f for the sub-set of opt ions to be stud ied further was located .
Th i s z one was establ i shed in a debate . Further di scuss ion
concentrated on th i s grey area , with some rescoring of opt ions by
the group .
The outcome of th i s process was a de f ined sub-set of 12 opt ions for
further attent ion at the workshop ( the# Mapp ing Severa l
1 1
other opt ions were cons idered worthy of further work a fter the
workshop ( somet imes because there was not yet enough informat ion to
make a j udgement ) , and other opt ions were recommended to be
d i scarded .
The Mapp ing Process
The 4 2 de legates were d ivided into 12 work teams , each to deve lop
one o f the se lected opt ions establ i shed in the Screen ing Process ,
by studying the feas ib i l ity and marketabi l ity of the ir Opt ion .
Each team had a min imum o f 3 and a maximum o f 4 members , inc lud ing
one Bus iness Funct ion Member , and spent a day deve lop ing the ir
opt ion , and prepar ing l iterature wh ich cou ld be presented to , and
di scussed and eva luated by a l l , the other workshop members . A.
ser i es of presentat ions was g iven by each team , when the resu lts of
team de l iberat ions were d i sseminated to the rest of the workshop .
Fo l lowing the presentat ions , a l l de legates used a 1 to 5 sca le to
score each of the subset of options on each of two cr iter ia # the
t imesca l e and the re lat ive impact the opt ion would have on the
organ i z at ion i f it was feas ible . As with the Screen ing Process ,
each de legate f irst ( private ly ) gave the top opt ion on each
cr iterion a score o f 5 , and the bottom opt ion a score of 1 . A l l
other opt ions were then given a score of 1 to 5 , re lat ive to the
12
The T EAMWORKER system was undoubted ly use fu l in terms of provid ing
a degree o f structure to a comp l ex task carr ied out by a large
group of experts . It a l so enabled rap id ident i f icat ion of areas of
strong di sagreement , mak ing it easy to prompt re levant debate . A
number of aspects or i ssues which ar i se from th i s case app l i cat ion
are ident i f ied as fo l lows #
The Core Team wanted to use a scor ing approach wh ich
was easy to understand and imp lement , so a 1-5 sca le
was adopted in order to score each opt ion rather than,
for examp le , a pa irwise or rat io sca le approach .
Each part ic ipant was asked to ident i fy independent ly
the ir persona l lowe st. and . highest. rated options , and to
a l locate scores of 1 and. 5 to these respect ive ly# Th i s
ensured that the ful l sca le was used , and serves as a
bas i s for produc ing a ranked l i st of opt ions from each
de legate .
( i i i ) The scores themse lves were used in order to ca lcu late
average and standard deviat ion va lues . The r igour of
th i s form of aggregat ion i s quest ionable . An
a lternat ive approach cou ld be aggregate rank ings rather
than scores . In <discuss ion , the Core Team pre ferred to
1 4
work with the scores in terms of c lar ity of the
feedback , bear ing in mind that th i s was intended to
generate debate rather than create excess ive dependence
on a mathemat ica l rout ine and assoc iated output .
( iv ) The group process achieved a h igh l eve l of susta ined
attent ion and app l icat ion , invo lving a l l the de legates
over severa l days .
( v ) The feedback screens frequent ly resulted in verba l
express ions of surpr i se , typ ica l ly concerned with
d i f ferences of j udgement , and eas i ly served to generate
focussed debate .
( vi ) Re-scor ing a fter feedback and d iscuss ion d id not
invar iably reduce di f ferences of opin ion , but
nonethe l ess opin ions frequent ly changed at these
stages .
( vi i ) There was an occas iona l review of the de f in it ions of
opt ions , espec ia l ly dur ing debate fo l lowing feedback .
( vi i i ) The workshop resu lted in the group reach ing a
co l lect ive dec i s ion # a de f ined short-l i st of
opt ions , to each o f which a sma l l group was ass igned
for further indepth review and assessment a fter the
1 5
Workshop . There appeared to be commitment to th i s
on-going work , perhaps as a resu lt of the h igh l eve l
o f involvement of the part ic ipants .
( ix ) The workshop was a co l lect ive h igh l ight or ep i sode in
an on-going process of technology assessment and
strategic p lanning , rather than a start or f in i sh in
its own r ight .
We conc lude th i s paper with a quotat ion made by the organ i z ing
of f icer short ly a fter the event # # We might have had some other ,
more convent iona l , meet ing but I don # t think it would have been
near ly as product ive #
1 6
Minkes , A1 ( 19 8 7 )
B l ake , Stewart P ( 19 7 8 )
F in l ay , Pau l and Marples ,
Chr i s ( 19 9 2 )
Ford , Dav id ( 19 88 )
Gear , T and Read , M #
( 19 88 )
Gear , Anthony E and Read ,
Mart in # ( 19 9 3 )
# ant sch , E ( 19 7 2 )
Mockler , Robert # andDo lo g ite , DC ( 19 9 1 )
Prasad , AVS andand S omasekhara N . ( 19 9 0 )
Read , M and Gear , T ( 19 8 9 )
Read , Mart in Gear , T
( 19 9 3 )
Stewart , T # ( 19 9 2 )
The Entrepreneuria l Manager #
,
Fengp in , p . 138 .
# Managing for Respons iveResearch Deve lopment #
. W . N . FreemanC o . , San Franc i sco .
Strateg ic Group Dec i s ion Support SystemsA Gu ide for the Unwary #
. # ong Range
Deve lop your Techno logy Strategy #
# ong Range P l ann ing Vo l 2 1 , 5 , pp . 8 5 -9 5 .
# Oh -# ine Group Dec i s ion SupportProceed ings o f VIIIth Internat iona lCon ference on Mu lt ip l e Criter ia Dec i s ionMaking , Manchester . Pub l i shed by Spr ingerVer l ag .
Oh -# ine Group Proces s Support #. Ac cepted
for Pub l icat ion in OMEGA .
Techno l og ica l P l ann ing and Soc ia l Futu res .
W i ley , N ew # ork .
Us ing Computer So ftware to Improve GroupDec i s ion-Making #
. # ong-Range P l ann ing . Vol .
4 , pp . 4 4-5 7 .
# The Ana lyt ic # ierarchy Proce s s for Cho iceo f Techno log ies
#.
Techno logica l Forecast ing and Soc ia l Change ,
3 8 , pp . 15 1-15 8 .
Interact ive Group Dec i s ion Support #.
Proceed ing s o f MCDM Internat iona lWorkshop on Mu lt ip l e Cr iter ia Dec i s ionSupport , # e l s inki . Pub l i shed by SpringerVer l ag .
# Dec i s ion Support for Management Meet ingsAccepted for Pub l icat ion in O R Ins ight .
A Cr it ica l Survey on the Statu s o fMu lt iple Cr iteri a Dec i s ion Making Theory andPract ice . Ome a , Vo l . 20 , N o . pp5 6 9 -5 8 6 .
24
Souder , W i l l i am E A Rev iew o f Creat iv ity and Probl em So lv ingZ ieg ler , Robert W ( 19 7 7 ) Techn iques
# Research Management , # u ly pp . 3 4-4 2 .
To z ar , Edwin E ( 19 8 6 ) Deve loping Strateg ie s for Management In format ionSystems #
. # ong Range P l ann ing . Vo l 19 , 4
pp . 3 1-40 .
Weble r , Thomas # # ev ine A N ove l Approach to Reduc ing Uncerta intyDebra # Rake l , # orst # and Techno log ica l Forecast ing and Soc i a l Change .
Renn Ortwin ( 19 9 1 ) Vol 3 9 , pp . 3-2 6 3 .
W i l lyard , Char les # ( 19 8 7 ) Motoro l a # s Techno logy Roadmap Proces s #.
Research Management , Sept-Oct . , pp . 13 -19 .
Vetsc he ra , Rudolph ( 19 9 1 )#
Integrat ing Databases and Pre ference Eva luat ions
lO NITOR
Figure l. TEAMW ORKER FeedbackSystem
Figure 2. # istogram ofScores forComplete Group
Figure 6 . # istogram ofScores forRelative # ikelihood and Relative Impact
2 3 4
R e a ve # k e h o o d
Figure 7 . Grid Show ing Average Scores and Standard Errors forEachOption