Professional Documents
Culture Documents
The period from 1993 to 1998 saw the UK’s upstream oil and gas industry move towards
more collaborative relationships (partnering and alliances) between the oil companies
and their contractors.
This paper is derived from observation over three years of an alliance between an oil
company and five of its contractors involved in underwater projects.
The paper uses concepts of ‘goodwill trust’ and ‘obligational contractual relationships’
which have been introduced by Sako.The ideals of obligational contractual relationships
are very similar to those of the collaborative relationships between companies in the UK
upstream oil and gas industry in the 1990s.
Attitude surveys of members of the alliance in three consecutive years attempted to
estimate the levels of goodwill trust between the people and companies in it.
It would be expected that trust levels within an alliance would rise over time as people
gain experience of working with each other. Analysis of the results from the survey did
show an overall increase in goodwill trust over the period but also showed some
company pairs where trust levels remained low.
The paper speculates that the reasons for reduced levels of trust could include: the
experiences inherited from previous relationships, project successes or failures affecting
the way people felt about each other, perceived lack of fairness in allocation of risks
and rewards, and changes in senior management.
Copyright © 2003 John Wiley & Sons, Ltd.
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
368 Richard Green
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
An oil industry alliance 369
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
370 Richard Green
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
An oil industry alliance 371
from different companies would increase as always vulnerable to shortage of good infor-
the alliance proceeds. Sydow (1998) describes mation about the conditions under which the
a slightly different aspect where if trust is sus- work was to be done. Any problems, errors or
tained, it gives the actors even more power delivery of defective equipment to the off-
and opportunities to reproduce trust. This was shore work site could result in significant
labelled ‘a self-heightening cycle of trust’. delays in completion of the projects, with the
Nevertheless, Sydow also mentions the concomitant increased costs and delays to
possibility of a ‘spiral of distrust’. start of the customer’s revenue flow. The com-
The results described in this paper arise panies were all involved at the design and
from a unique opportunity to observe the planning stage of each project but their prin-
evolution of a major alliance in the upstream cipal effort came at different stages through
oil industry over its lifetime. the project. Nevertheless, the success of a
project depended crucially on the actions of
each of the alliance companies. The aim of
The alliance on which this research
alliance companies was to work together to
is based
obtain a significant reduction in the cost of
This paper is based on research on an alliance, the projects while maintaining or improving
created in 1994, between an oil company (the the quality of their execution.
customer) and five major contractors. All the The alliance executed a series of parallel
companies had previously had long-term rela- projects and its workforce was organized in
tionships with the customer. Several of the teams. Each team contained representatives
contractors within the alliance were and still from the alliance companies, including the
are direct competitors in the open market. For customer, and was largely responsible for
reasons of commercial confidentiality, the six planning and detailed execution of its own
companies are identified in the paper as com- projects. The teams contained people working
panies A through to F, where company E is both onshore and offshore. At its peak, there
the customer. were eight teams working in the alliance. The
The alliance was terminated during restruc- overall working of the alliance was nurtured
turing of relationships following the period of by a group of very senior managers from the
low oil prices in 1998 and this paper is based customer and contracting companies, which
on data from attitude surveys of people in the met at least once per month. There were
alliance in 1995, 1996 and 1997. The main various changes in the senior managers of the
activity of the alliance was the execution of a alliance companies over the life of the alliance
series of complex engineering projects that and the senior manager of the customer
involved work on and beneath the seabed, in company changed between the second and
water depths of about 100 m. Successful com- third years of the alliance. Within the alliance
pletion of the projects was vital to the pro- there was considerable investment in team
duction of oil and gas from the customer’s building and coaching in collaboration, both
fields. The success of an individual project was at working and senior management level.
defined in terms of the quality of the work Senior managers put much effort into creating
done (long-lasting production capacity), the an alliance identity with logos, posters and
safety and environmental performance, the gifts for those involved.
time taken and the final cost. Each project There was a risk–reward structure, which
involved many different engineering disci- enabled the alliance contractors to gain when
plines and success depended on close co- performance was improved, but which also
ordination of construction activities, logistics penalized them if performance fell below
and the supply chain for major equipment target. The form of the risk–reward structure
items. The activities of the alliance were was changed between the second and third
vulnerable to bad weather and success was years of the alliance life. During the life of the
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
372 Richard Green
alliance there were some spectacular suc- belonging to the other companies in the
cesses, where careful planning and close co- alliance. Thus the need to name individuals in
operation allowed projects to be completed the questionnaire was removed and the task
substantially faster and at lower cost than had of data collection became easier. However, the
been achieved previously. For these projects opportunity to analyse relationships between
the companies and in some cases the people all possible pairs of individuals was lost.
received substantial bonuses. There were also
failures, where knock-on effects of technical
Questionnaire
problems caused projects to overrun the time
and cost targets considerably. For these pro- There were four basic statements in the
jects, bonuses to companies and people were questionnaire, which related to goodwill trust,
not paid. as follows:
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
An oil industry alliance 373
alliance. Thus, the following data items are Table 1. Internal consistency of responses on goodwill
associated with a response to each statement: trust towards each company
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
374 Richard Green
Table 3. Level of goodwill trust by receiving company Table 4. Level of goodwill trust by giving company and
and year (significant changes over the years shown bold) year (significant changes over the years shown bold)
and third years the level of trust seems to have between 50% and 70% are designated as
reduced slightly. medium.
Table 4 shows the level of goodwill trust There are 36 possible pairs of trust-giving
given by the people in each company.The vari- and trust-receiving companies. For the six
ation of ‘trustingness’over the three years is not instances of internal trust, where the trust-
consistent for all the companies. Companies B giving and trust-receiving companies are the
and D seem to show rather lower levels of it same, it can be seen that the level of trust
than do the other companies. Companies A and within companies seems to be in the high or
F show large increases in positive responses medium area for all three years. It would be
between the first and second years, and surprising and potentially dysfunctional if
Company A shows an increase between the there were low levels of trust within a
second and third years. However, Company F company. For the remaining 30 pairings of
shows a decrease in the third year. Company B, trust-giving and trust-receiving companies it
on the other hand, shows a decrease in the appears that, in the first year of the alliance,
second year but an increase in the third year. 45% of the possible pairs had high or medium
These inconsistent responses might indicate levels of trust and 55% were low. In the second
that the people in the alliance companies have year the proportion of pairs with medium
different perceptions of the same situation. or high trust levels increased to 73% and this
The differing perceptions could be because proportion is maintained in the third year.
the companies do not all have the same level Thus we can claim that there has been an
and type of interaction with each other. overall increase in levels of goodwill trust over
Table 5 shows the level of goodwill trust the three years and that most of the increase
between pairs of companies. In interpreting came between the first and second years.
this table, a value of less than 50% trust has Examining the changes between the first
been taken as evidence of a low level of good- and second years and excluding the values for
will trust from one company to the other and internal trust (where giving and receiving
a value of 70% or over has been taken as evi- companies are the same) shows that of the
dence for a high level of goodwill trust. Values 20 pairings for which there are data:
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
An oil industry alliance 375
Table 5. Levels of goodwill trust between pairs of companies by year (significant changes over the years
shown bold)
A B C D E F
Giving Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust
Company 1st 55 1st 30 1st 1st 38 1st 17 1st 22
A 2nd 81 2nd 59 2nd 63 2nd 66 2nd 59 2nd 57
3rd 89 3rd 79 3rd 89 3rd 82 3rd 55 3rd 75
Giving Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust
Company 1st 33 1st 77 1st 1st 55 1st 50 1st 23
B 2nd 22 2nd 56 2nd 50 2nd 57 2nd 47 2nd 25
3rd 43 3rd 83 3rd 53 3rd 61 3rd 75 3rd 48
Giving Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust
Company 1st 1st 1st 1st 1st 1st
C 2nd 64 2nd 84 2nd 77 2nd 79 2nd 80 2nd 88
3rd 74 3rd 82 3rd 82 3rd 85 3rd 85 3rd 64
Giving Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust
Company 1st 88 1st 92 1st 1st 82 1st 22 1st 5
D 2nd 58 2nd 79 2nd 45 2nd 74 2nd 42 2nd 43
3rd 35 3rd 37 3rd 47 3rd 74 3rd 38 3rd 29
Giving Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust
Company 1st 50 1st 60 1st 1st 81 1st 64 1st 57
E 2nd 51 2nd 64 2nd 53 2nd 71 2nd 64 2nd 44
3rd 60 3rd 64 3rd 36 3rd 66 3rd 79 3rd 50
Giving Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust Year % Trust
Company 1st 100 1st 58 1st 1st 46 1st 36 1st 74
F 2nd 66 2nd 77 2nd 47 2nd 76 2nd 79 2nd 95
3rd 58 3rd 68 3rd 85 3rd 66 3rd 65 3rd 87
Four pairings stayed in the low trust level. Consistent low levels of trust between com-
Seven pairings stayed in the medium or high panies are of interest and could be a cause
trust level. for concern. In Table 5:
Seven pairings moved from low trust to
medium or high. Company B has a continuing low level of
Two pairings dropped from medium to low trust for Company A. These two companies
trust level. are competitors in the open market. Inter-
estingly, Company A’s trust for Company B
The changes between the second and third was low in the first year but rose to a high
years and excluding the values for internal level in the third year. Company B also has
trust (giving and receiving companies are consistent low levels of trust for Company
the same) show that of the 30 possible F.
pairings: Company D displays puzzling behaviour. Its
levels of trust for Companies A and B are
Five pairings stayed in the low trust levels. very high in the first year but drop to very
Nineteen pairings stayed in the medium or low in the third year. It also displays con-
high trust levels. sistent low levels of trust for Companies E
Three pairings moved from low trust to and F.
medium or high.
Three pairings dropped from the high or It is interesting that trust levels are not neces-
medium to low trust levels. sarily reciprocated between the companies.
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
376 Richard Green
For example, Company D gives low levels of The people I deal with in ·company nameÒ
trust for most of the other companies in the will do whatever is necessary to ensure
third year but it consistently receives high or the success of the alliance even if it in-
medium levels of trust. Company F receives volves tasks to which they had not agreed
low levels of trust from most of the other previously.
companies but its people have high trust for If I have a problem that can influence the
the other companies. success of the alliance, the people I deal
with in ·company nameÒ will work hard
to help me solve the problem.
The people I deal with in ·company nameÒ
Trust levels are not
will not withhold any information that I
necessarily reciprocated need to help me perform well in the
between the companies project.
The people I deal with in ·company nameÒ
will not exploit to their advantage any tem-
porary weakness of myself or my company.
Discussion of results
How representative is the sample? At the peak, The first and second statements were designed
when eight parallel project teams were oper- to gather views on the commitment of others
ating, there were in excess of 500 people and their willingness to take initiatives over
involved in the alliance.Their jobs ranged from and above what was explicitly promised, and
senior managers through middle managers, the third and fourth statements aimed to
technical specialists and offshore supervisors capture views on the absence of opportunis-
to offshore craftsmen and labourers. Much tic behaviour. Expectation of commitment of
effort was invested in promoting the identity others and absence of opportunistic behav-
of the alliance and its ideals to all the people iour are key components of goodwill trust as
involved. However, it is likely that the middle described by Sako (1992), and it is considered
management, technical specialists and off- that responses to the four statements would
shore supervisors were most aware of the reflect these components. It is also considered
alliance. When carrying out the survey the reasonable to assume that the proportion of
intention was to attempt a census of every people agreeing or strongly agreeing with the
person who might have an involvement with statements would be an indication of the level
the alliance and its work and about 500 of goodwill trust between groups of people.
questionnaires were distributed in the second The reasonably high values of the Cronbach’s
and third years. The response rates in years alpha coefficient give encouragement that
1 through 3 were 72%, 28% and 37% the four statements do relate to the same
respectively. construct.
The job types of the people who did
respond covered the full range from senior
Anticipated versus actual results
managers to offshore labourers. The majority
of people who responded were from job types A priori, it could be expected that the level of
that should be aware of the alliance and its goodwill trust between the people from the
aims. Thus it can be argued that the sample is alliance companies would increase as they
sufficiently representative that some credence gained experience of working together. We
can be placed on the results of the survey. would expect increases in trust to be espe-
How effective is the questionnaire in cially marked if the alliance was successful in
measuring goodwill trust? The results in this its projects. Overall the results do show that
paper rely on responses to the four statements levels of goodwill trust have increased over
shown below. the three years — 73% of possible pairings of
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
An oil industry alliance 377
companies are in the medium to high trust deliver improved performance. They will
category in year 3 compared to 45% in year 1. continue to be motivated if they believe that
Most of the gains appeared to come between their efforts are recognized and if they per-
years 1 and 2. However, for some pairings the ceive that their company is receiving an
levels of trust remained low for all three years appropriate share of the benefits arising
and for others, the levels dropped back into from the improved performance. In this
the low category in the third year. Examples alliance the senior management struggled
of this pattern are shown in the trust from to find a risk–reward structure with which
Company B for Companies A and F, and in the all companies felt comfortable and the struc-
trust from Company D for Companies A, B, C, ture was changed between the second and
E and F. Although the survey did not set out to third years. There were occasions when
gather data on possible reasons why the level Company D felt that it was not being fairly
of goodwill trust remained low or fell during rewarded for its efforts. It is possible that
the period of the alliance, it is possible to spec- this may have influenced the level of good-
ulate what those reasons might be. If the same will trust from Company D.
team of people is engaged in a series of similar Continuity of senior management. The
projects over time and their level of success importance of commitment and example of
increases as each project is completed, then senior management to the success of col-
we would expect the level of trust between laborative relationships has been noted in
the people to increase as they gain experience the literature (e.g. Green and Keogh, 2000).
of working with each other. However, other In the alliance under study the senior man-
influences which may affect a person’s feel- agement teams remained fairly constant for
ing of trust for others include: the first and second years but in the third
year there were several changes including
Experiences from previous relationships. the departure of the senior manager, who
All the contractor companies in the alliance had set up the alliance from the customer
had previously had a long-term working company. The new manager, although being
relationship with the customer company. very supportive of the alliance, changed the
Company B was a competitor of Company way it was managed and this too may have
A outside the alliance. Company D had also influenced the results from the survey in the
worked with the other contractors on many third year.
occasions in previous adversarial relation-
ships and it appears that its built-up sus-
Conclusions
picion and distrust were slow to dissipate
during the life of the alliance. This paper has described the use of statements
Degree of success or failure. It is often said in an attitude survey to estimate the levels of
that a true test of a collaborative relation- goodwill trust between people in an oil indus-
ship occurs when difficulties are being try alliance of six large companies over three
experienced. It may well be that an individ- years. Goodwill trust is described by Sako
ual’s view of the alliance and his or her level as being very important to the success of
of goodwill trust for others may be affected obligational contractual relationships, which
by how well the current projects are pro- are very similar to the collaborative relation-
gressing. The alliance teams experienced ships embraced by the UK upstream oil indus-
both successes and non-successes in their try in the mid to late 1990s.
projects and this could well have affected It was suggested that levels of goodwill trust
their responses to the questionnaire. should increase with time as people gain expe-
Perceived fairness of risk–reward struc- rience of working together. Analysis of the
ture. Within an alliance that is working well, results shows that levels of trust did increase
people are motivated to work hard to over the three years, and that most of the
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
378 Richard Green
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003
An oil industry alliance 379
Copyright © 2003 John Wiley & Sons, Ltd. Strategic Change, November 2003