Vous êtes sur la page 1sur 7

At the peak of tbe stock market boom in

"space" sbares a year or so ago, a market letter


seriously suggested a new investment yardstick
in place of tbe apparently obsolete price/earn-
ings ratio: a "price/research-budgot ratio." To-
day this reads like incredible naivete. But tbe
assumption underlying sucb a pricc/rescarcb-
budget ratio is still very much alive: viz., that

Twelve Fables "researcb" by itself is a meaningful business re-


sult or, at least, a surefire guarantee of results.
Stated so baldly (rather tban tacitly taken for
granted), the fallacy stands out. Research is a

of cost or an investment; far from delivering gttar-


anteed results, it is a highly speculative, highly
uncertain effort that requires the greatest man-
agerial eompetence to produce results. Judging
by the over-all performance of American busi-
Research ness these last ten years, all we arc really good at
so far is spending research dollars. We still have
to learn how to get results from this expenditure.

Management Pitfalls & Pratfalls


Research spending bas become tbe "growth
sector" of tbe American economy, doubling in
dollar volume every four years. More remark-
Research expenditures keep growing, able still is the shift in national attitude. A
but wbat about tbe results? region that complains of a deartb of defense
contracts is no longer just told to elect more
Democrats; it is counseled to spend more on in-
dustrial research!
By Peter F. Drucker
But do we get greater research effort from
tbe jump in research expenditures — from $4
billion to about Si 5 billion in one sbort decade?
A good deal of this amount, I am afraid, simply
represents more spending on nonresearcb and
especially on "research support" — clerks, re-
port writers, massive (and largely useless) elabo-
ration, new buildings, and complex equipment.
How many companies, for instance, are spend-
ing millions on fashionable apparatus — e.g.,
a cryogenic cbamber — witbout even asking
whether tbis beautiful machinery will ever be
used for anything as sordid as commercially
usable results? And tbere are some major space
programs where 75 cents out of every R & D
dollar goes for "engineering documentation."
To be sure, tbe number of researchers em-
ployed in industry bas doubled in the last ton
years. But tbe number of people wbo arc called
"researcbers," yet are not doing research, has
grown faster still — researcb administrators, co-
ordinators, expediters, and so on. Even witb tbe
aid of these "organizers," the research force it-
103
104 Harvard Business Review
self has to spend so many hours sitting in meet- that industrial research should be organized and
ings or producing the paper flow which sustains managed.
the research administrator and other research- Unfortunately, a great many Ameriean busi-
support people that the total number of hours nesses still go by the fables — and seem to know
actually available for real research work may little else. They are, without knowing it, or-
even be less today than it was before the "re- ganized for "nonresults" from research, and can
search explosion" of the last decade. count themselves lucky if they get as much as
And where are the research results'? Does IO cents back on their research dollar.
the rate of economic growth of the American
economy or the profit performance of Ameri-
can business show any measurable impact from How to Get Nonresults
all this research spending? For ten years or There are a round dozen research-manage-
more we have heard glowing talcs of all the ment fables which every executive should know.
new products and processes about to issue forth Each impairs the ability of the ablest researcher
in ever-increasing volume. But, on the whole, or research group to produce.
rapid advances in products and processes are
still confined to those few industries — e.g.,
plastics, Pharmaceuticals, or electronics — that
were innovation-prone well before the "research
explosion."
Some of the ahlest of our research managers
and scientists actually are afraid we will price
ourselves out of research within another decade.
Whether or not their fears are justified, I am This first fable is fairly prevalent in industry.
not sure. But I do know that the trend of re- Quite a few companies I know of have ten times
cent years cannot go on much longer. We can- as many research projects in the works as they
not afford to keep paying more unless we start have qualified researchers — say, 500 research
projects for a professional research group of 50
getting better research returns.
men. But paralyzing projectitis sets in predict-
New products and processes, entire new in-
ably unless the number of projects is very much
dustries, and, of course, very substantial profits smaller than the number of researchers — one
can be obtained from industrial research. We to five, perhaps, or even one to ten. Otherwise,
have a whole century of experience to prove it. research people are much too busy holding meet-
The research dollar should, indeed, be the most ings about what they should go to work on ever
effective and most profitable dollar spent by to do any real work, much too busy writing
business. That it is not indicates there is some- metaphysical memoranda on the "true inward-
thing very wrong with the way industrial re- ness of research" ever to do any research itself.
search is being handled in many American busi- Research work that promises truly significant
nesses today. economic (or scientific) results will not even get
started, since such work always requires hard
Negative Knowledge thinking, concentration, careful review, and sus-
We may not yet know how to manage indus- tained effort. It requires, in other words, the
trial research to get favorable economic results. full-time attention of several good research peo-
This, I submit, is a tenable hypothesis. More- ple for a considerable period.
over, it is perhaps the only hypothesis which
explains the discrepancy between the vast liter-
ature on research management and the meager
results. After many years of work in the area, I,
for one, am a good deal more impressed by the
extent of my own ignorance than by my knowl-
edge, and can boast of many more failures than
successes. Yet those of us eoncerned with re-
seareh do know pretty well what not to do. We
do know the fables of industrial research —
the plausible, but false, beliefs about the way
Research Management 105
This fallacy is closely connected to tbe first Tbis common belief is but a sbort step from
one. No argument is more often heard in re- worsbip of tbe advanced degree as found in
searcb meetings than: "Smith isn't really good tbe tbird fable. Researcb work produces results
enough to do tbe job he took on. Let's put in its own ways. But so tloes any other work —
in Jones to bail him out. Jones, after all, is accounting, for instance, or marketing. We do
our best man." But, if used
tbis way, even tbe strongest man For more on research, turn to the Quinn and Mueller article, p. 49.
will achieve little. He will only ——— —
be able to convert weakness or failure into not besitate to hold otber functions accountable
respectable mediocrity — and mediocrity in re- for results. Wben it comes to research, bow-
search work does not pay off. If tbe job Smith ever, we bave a deep reluctance to ask the nasty
cannot do is truly a first priority, be should be question: "And what do you contribute tbat
pulled oft' and Jones should take over — minus justifies your staying on our payroll?" We seem
Smitb. And if Smith's job is not as important to suffer from tbe delusion that research ex-
as wbat Jones is doing — or should be doing — penditures are justifiable if only they enable
let's forget about Smitb and his project. Tbe healthy, intelligent, young men, with 20 years
best way to get results is to use all tbe strengtb of expensive schooling (paid for largely by tbe
tbere is to act tbem. taxpayer), to be pleasantly relaxed at $18,000
a year plus bonuses and stock options when,
otherwise, tbey migbt have to earn their living
tbe hard way like their grandfathers, hoeing
weeds in the potato patch.
Actually, researcbers — like otber people in
a business — produce the most when the de-
mands for economic results are high. You may
ask, "But does tbis also apply to obtaining new
scientific knowledge?" The answer is yes. All
The company that announces, "We are about our experience sbows that knowledge, especially
to increase our researcb force from 20 to 60 of the basic kind, does not come out of indus-
professional men within the next two years, trial research unless it is focused on economic
92% of whom will have their Pb.D.'s," can al- performance; on tbe economic needs of tbe busi-
most be guaranteed nonrcsults. All an advanced ness; on new processes, new products, and new
degree certifies is tbat a man bas sat on school markets. The experience of Bell Telepbone
benches a long time. Whether be bas learned Laboratories, Inc. sbows tbis clearly, as does
anytbing, let alone whether he can do anything that of smaller and newer companies (e.g.,
witb bis learning, has yet to be proved — and Xerox Corporation or Corning Class Works).
the burden of proof is on bim. By sheer sta- Conversely, by not "trammeling" researcb with
tistical probability, you should be able to get demands for business results, we actually impair
more able people the more people you hire, its scientific as well as its economic productivity.
but your chances of finding able men naturally Tbe origin of tbis fourtb fable is the popular
go down steeply tbe more the field is exhausted. notion of the university scientist wbo is allowed
In addition, tbe lazy and ineffectual Pb.D. is to "roam." In fact, this bolds true only for a
perhaps a greater drain on the organization mere bandful of experienced senior men with
tban is a lazy man witbout sucb intellectual records of proven achievement; the rest work
references. according to set objectives, if not to a tigbt
scbedule. And a rescarcber who eannot effec-
tively be "supervised," but must, largely, direct
himself, requires sharper objectives and greater
insistence by bis company on hard work and
significant results.
All tbis does not, however, argue for more
paper work. Yet in few areas have we become
as procedures-bappy as in research. This brings
us to tbe next fable.
106 Harvard Business Review
and when? How many cents a share is it going
to cost?" Not everything research does will fit
into this profit frame, but this is just as true of
work in other areas, e.g., of a good deal of per-
fectly intelligent, useful, and even necessary
personnel work.

The first goal of effective research manage-


ment should be to convert the maximum of re-
search dollars into actual research time. For
only people do research — not books, equip-
ment, forms, or procedures — and the only re-
sources of people are knowledge and time —
both of which are scarce and perishable.
The research department in which nonre-
search work takes more than a fraction of avail-
ahle time (with i o % an upper limit, perhaps) Companies have to decide what to do research
wastes researeh dollars and, worse, misdirects on. In the absence of genuine business-focused
researchers' energies and vision. I ran across a research objectives, following the leader is the
good example of what not to do in one well- easy way. But does it pay?
known company. Thus: One needs to know, of course, what tbe lead-
A "truly seientifie" personnel-appraisal form was ers — a Bell Lab, a General Electric, a Dow
distributed whieh asked each researeh man to fill Chemical — are doing. But one needs to know
out a 2 0-page questionnaire on himself, on every it so as not to follow them. To climb on tbe
one of his subordinates, and on his associates. It bandwagon is, in research, the sure way not
contained multiple-ehoice questions, such as: "Is to have economic results. At best, one comes
he in your opinion (mark one) firm, determined, out with some results — but maybe three years
rigid, dogmatic?" and "Is he In your opinion (mark behind the pacemaker, with a product that is
one) cooperative, friendly, ingratiating, spineless?" "almost as good."
Anyone taking tbis seriously would be foreed to
work on nonreseareh praetieally full-time! Of course, by climbing on the bandwagon
one avoids those disagreeable arguments which
might be caused by proposing something new,
original, and therefore questionable. For in-
stance, nobody is likely to quarrel with such a
plausible statement as: "To be a chemical com-
pany today one has to be in plastics on a broad
front." Yet the deceptive appeal of this argu-
ment largely explains the poor profit perform-
ance of several American chemical eompanies
This is, perhaps, the most common miscon- in the last few years.
ception. Industrial research uses seientifie and
technical knowledge (Who doesn't today in any
area of a business?). But its goal is not scientific
or technical knowledge; it is economic results.
It therefore needs business objectives. To be
sure, research objectives often will lead to a
change in business objectives — if only because
research opens up new profit potentials or new
market opportunities. But the starting point for
a research program is properly the needs and
goals of the company; the crucial questions
when selecting a research project are: "How
many cents a share is it likely to contribute —
Research Management 107
This is just as common as tbe previous fahle
— and even more dangerous. Compromise proj-
ects make certain researchers happy, for there
is a little real research in them, some real un-
certainty, and some challenge. They make the
"practical" people — the sales manager and con-
troller — very happy, for they promise reason-
ably probable results in a reasonable period of
time. At the same time, they have neither the
small but sure results of the "quickies," the com-
Everybody knows — or sbould know — that
mercial improvements which need only a little
work and imagination (hence, are too easy), nor one cannot "market research" a truly new prod-
the massive potential of the really big, long term, uet. One can only research a product, service,
uncertain projects which might change an entire or market that ean be defined and described.
industry and create a new technology, if success- A market study of tbe electronic computer in
ful against all odds (too challenging). What 1938 would have been as senseless as would
these compromise projects always have, how- one of space travel today. By limiting work to
ever, is the built-in arithmetic of compound in- product ideas that can be market-researched,
terest which normally offsets any return they one therefore makes sure that no truly new prod-
might yield. uct or process ever comes out. For novelty re-
quires judgment and courage in addition to sta-
tistics. It requires insight into economic, so-
cial, and technological developments, potentials,
and changes, not just figures on the present and
the past.

This fahle is my special pet peeve. Concrete-


ly, it means that a research group is not encour-
aged to do "preventive" research, to find new
and better products to take the place of the pres-
ent successful product line. The prevailing logic Although just about everyone says that he
seems to be that if anyone is to make the com- knows tbis to be wrong, it seems to be practiced
pany's product obsolete, let it be the competitor. by a great many companies. I often associate it
If a product is selling well in the market today, with another fable, which we might identify as:
it would be much too upsetting to the sales peo-
ple — and to the financial people, as well — if
research went to work to find the premium prod-
uct which would take its place seven or ten years
hence. The philosophy is: let the research
people stick to "defensive" research, which pours
a lot of money into efforts to catch up with the
competitor or to slow down the decline of yes-
terday's products, or let them go in for "blue
sky" research (provided, of course, that they
don't go any further than a "feasihility study").
But don't encourage preventive research which
might produce economic results.
If a company does not want research to "rock There is no law of nature that says a good
the boat," wouldn't it be better off having no researcher cannot be a good research director;
research? At least then it would not have to it is merely unlikely that he will be. There is
pay for researeh. no law of nature that says a poor researcher
108 Harvard Business Review
cannot be a good research direetor; it is merely money, market, and customers. This appraisal
unlikely. Nor is it a law of nature that an ahle of development work as "implementation" leads
husiness manager with some technological im- to its organizational subordination as a junior
agination will work out as researeh director. component of a research department that has
However, this is the least unlikely of the gen- technical rather than business orientation and
eral propositions. What is needed is a man, not knowledge. This setup, more than anything
a type; and wherever one needs a man, one has else, explains the widening time-gap between
to look for a man and look at a man. research results and product introduction in
Research management is work. It is impor- American industry (but not in Europe or Japan).
tant work, considering how much is at stake. It It is tbat lapse whieh so often deprives genuine
is difficult work, considering how rarely it is scientific and technical achievement of its eco-
done well. What it is comprised of, we may nomic results.
not yet really know. But it is not "research";
it is "management."
Conclusion
Ridding ourselves of the "fables" will not
guarantee effective research management and
profitable research results. Research will remain
high-cost effort and risky investment. But, hy
not hamstringing ourselves with delusions and
old wives' tales, we can, at least, begin working
on the difficult, demanding, and important job
of making research yield its high-return poten-
Admittedly, this fable is not consistent with tial. At its best, research management (and per-
all the others, yet it has spread through much haps all business management) has yet to catch
of industry. Development work is regarded as up with the proverbial farmer who already knew
dowdy. Unlike "real researeh," which is called how to farm twice as well as he actually did.
"challenging," "imaginative," and "creative," it But we surely need not — as we do today in
merely has the task of integrating into one so many American businesses — spend time,
harmonious, money-making course of aetion money, and effort to manage research as poorly
scientific knowledge, manufacturing processes. as we know how.

GE brings experienee and sound judgment. "Experience is the best teacher";


A "Old men are wise and young men, impetuous"; and so on through all the
eliehes, aphorisms, and old wives' tales, which the elders have used for generations
to restrain their juniors. True, judgment is important and years of experienee
should hring sound judgment. It is also true, unfortunately, that experienee tends
to be specific and to limit a man's vision to a narrow area. It is difficult to keep the
mind open to innovations and to sift out of past experienee those eonsiderations
whieh are pertinent to a new problem. The problem is partieularly acute in men
who have themselves earried through important development projects earlier in
their careers. Development work is a tough way for an engineer or scientist to
make a living. It is a young man's type of work, requiring discontent with the past
and unbounded optimism for improvement. Uncertainty is their constant com-
panion.
M. P. O'Brien, "Teehnologieai Planning and Misplanning," in Technological
Planning on the Corporate Level, edited by James R. Bright
Boston, Division of Research, Harvard Business School, 1962, p. 90.

Vous aimerez peut-être aussi