Product Strategy is perhaps the most important function of a company. It must take in account the capabilities in terms of engineering, of production, of distribution (sales) existing in the company or of time to acquire them (by hiring or by mergers). It must evaluate the customers expectations at the time of delivery. It must guestimate the competition (including new entrants) probable moves to enter the same market.
Product strategy by Bull appeared sometimes erratic and not coordinated, specially during the periods where product lines run independently. However, it has been dominated by very old trends rooted in the Sales Network during the 1950s defining Bull's market around the business applications, and fighting against the sole IBM as competitor.
So, the company adopted its version of IBM's business
model, following IBM with a variable delay, in the domain of products, price and
market following. Sometimes new opportunities appeared and some innovative products were
developed, (e.g. time-sharing in GE time, smart card applications) but they faded as
marginalized by the Sales Network. In fact, the Sales Network was not conscious of the
pressure it exerted on Planning and Engineering. Often, it focalized on IBM's short term
moves, ignoring the reasons for those moves (sometimes due to legal constraints, sometimes
by internal fighting inside IBM, other times because other competitors moves).
While IBM's influence on Bull was extremely important, the reverse existed sometimes (1). Dispute between IBM World Trade and IBM US domestic may have been fueled by some worry of IBM European salesmen about some Bull's (and GE's or Honeywell's) products.
The capability of Bull to match IBM's offer on the market never existed. Before the GE's merger, Bull did not address the US market directly and by consequence excluded itself from the market segments needing the quantities only addressed by a worldwide market (such as large scientific computers). Another market that was ignored (knowledgeably) early was the small scientific market; its margins did not matched the corporate model.
Bull never did a comparable investment to IBM's in the technology area. Each time it (or its American associates) tries a significant move, the success did not reward it. The reasons of the failure were multiple: overestimation of the return on investment, lack of a long term perspective (that existed in architecture and software), size of market. Some more specific problems were due to the lack of experience in fundamental physics, themselves related to the isolation of the engineers.
For historical reasons related to the acquisition of
a park of customers and for "political" reasons, Bull did not succeed to shut
down a product line before the 1990s. Its resource limitations did not allow to embark in
the simultaneous developments of more than one or a couple of compatible processors at the
same time. Product Planning had to prepare several product line plans and to invent models
within each product line to match the competition prices and performances. Models
were developed from a single engineering design with the same manufacturing cost by
slowing down the processor clock or adding dummy cycles and/or by reducing the
"connectivity" of the system.
When the performances exceeded IBM's target, the system was not sold at full speed to avoid the risk of undercutting IBM future announcements' price and keeping some reserve power to react against a competition "mid-life kicker".
New higher models were also created by unleashing the design constraints after one year. New lower models were created by slowing down a bit already shipped processors.
This strategy worked well as far as the manufacturer controlled completely the customer configuration by leasing the systems. The first evolution of the model was the advent of clones manufacturers. They obviously attacked IBM's market but GE, Honeywell and Bull strategists ordered to take all measures, sometimes detrimental to product and service costs, to escape cloners. The architecture or the assembler of the machines remained confidential, source and object code of programs was secrete, network architecture was not available even to peripheral suppliers, peripheral interfaces were modified and the differences kept in vaults... Bull argued to the persons objecting the strategy (suppliers, other manufacturers, customers ) that it would respect the "de jure" standards (such as ISO's or ANSI's) but that it did not have to follow the "de facto" standards (such as IBM's). That changed in the 1980s when "Open Systems" became Bull's religion.
Another IBM decision impacted the business model, it was unbundling. While the IBM pricing was more or less related to development and manufacturing costs, adopting the same price for Bull's items where software, for instance, was reproduced in far smaller number of copies, lead to a disconnect between decisions to produce and customers acceptation. Specially in the late 1970s and the 1980s, Bull embarked in many developments with a very low production rate, but they were asked to match the IBM's catalog. Later, in the late 1980s, the competition with open systems, lead to some re-bundling of the offer (the word was "packaging") where for instance associate a purchased data base system with a memory bank and even an additional processor.
See: Strategy Methodology, Competition
(1) Bull was not yet related to Honeywell at the time where the Honeywell H-200 was seen as a strong menace on IBM1401.