Anda di halaman 1dari 4

Importance-Performance Analysis

Author(s): John A. Martilla and John C. James

Source: Journal of Marketing, Vol. 41, No. 1 (Jan., 1977), pp. 77-79
Published by: American Marketing Association
Stable URL: .
Accessed: 28/09/2013 13:10

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact

American Marketing Association is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Marketing.

This content downloaded from on Sat, 28 Sep 2013 13:10:27 PM

All use subject to JSTOR Terms and Conditions
John A. Martilla and John C. James

An easily-appliedtechniquefor measuringattributeimportance
and performance
canfurtherthe development of effectivemarketingprograms.

IRMSconductingattributeresearchto measure An Example

consumer acceptance of particular features of
their marketing programs frequently encounter An automobile dealer's service records indicated
that only 37% of its new car buyers remained loyal
problems in translating the results into action.
Several factors may contribute to this situation, but service customers after the 6,000 mile service. The
two are particularlytroublesome: firm hoped to increase that figure to 50%as a means
of improving service department profits as well as
1. Management may find it difficult to under- stimulating repeat sales of new vehicles. A literature
stand the practical significance of research search and conversations with service and sales
findings expressed in terms of "coefficients of department personnel and factory representatives
determination"and "levels of stress." identified 14 attributes which were felt to affect
2. The researchmay have examinedonly one side service department patronage.2Respondents were
of the consumer acceptancequestion - either then asked two questions about each attribute:
attribute importanceor attribute performance How important is this feature?
- rather than both. How well did the dealer perform?
Yet empiricalresearch has demonstrated that
consumer satisfaction is a function of both expecta- Questionnaires were mailedto 634 individuals
tions related to certain important attributes and who had purchased a new car from the dealer be-
judgments of attribute performance.' In light of tween one and two years earlier, and 284 usable
these considerations, importance-performance returns (45%) were received after one follow-up
analysis has been found to be a useful technique for mailing.
evaluating the elements of a marketing program. The mean importance and performance
The technique draws on conceptualcontribu- ratings for the 14 automotive service attributes are
tions to be found in many places in the literature. presented in Exhibit1. Itwas confirmedthat virtual-
It is hoped that the specific example offered ly identical results would have been obtained using
here will highlight the approach- to present a case median rather than mean values.
where the technique was clearly useful. An attractive feature of importance-perfor-
mance analysis is that the results may be graphi-
cally displayedon an easily-interpreted,two-dimen-
About the Authors sional grid. The 28 attribute ratings from Exhibit1
JOHN A. MARTILLAis Associate Professor of Market- are plotted as 14 points on the importance-
ing in the School of Business Administration, Pacific performancegrid in Exhibit2. The numbersrefer to
Lutheran University, Tacoma, WA. the attributes listed in Exhibit 1. The labels of the
JOHN C. JAMES is Assistant Vice President of Frank quadrants A, B, C, & D refer to marketing effort.
Russell Co., Inc., Tacoma, WA.
For example, "Concentrate Here" denotes an
area (A) where attributes are important and also

This content downloaded from on Sat, 28 Sep 2013 13:10:27 PM

All use subject to JSTOR Terms and Conditions
78 / Journalof Marketing,January1977

where performance can be improved. Concen-

trating constructive action in this area would EXHIBIT2
produce maximum results. Importance-Performance Grid with
Attribute Ratings for Automobile Dealer's
Service Department
Importance and Performance Ratings for Extremely
Automobile Dealer's Service Department Important
A. Concentrate Here B. Keep Up The Good Work
Mean Mean
Attribute ImportancePerformance 1
Number AttributeDescription Ratinga Ratingb 2
1 Job done right the first time 3.83 2.63 4
2 Fast action on complaints 3.63 2.73 5
3 Promptwarrantywork 3.60 3.15
4 Able to do any job needed 3.56
5 Service availablewhen needed 3.41 3.05 7
6 Courteous and friendly service 3.41 3.29 8
7 Car ready when promised 3.38 3.03 9
8 Performonly necessary work 3.37 3.11 10
Fair Excellent
9 Low prices on service 3.29 2.00
10 Clean up after service work 3.27 3.02 Performance Performance
11 Convenient to home 2.52 2.25
12 Convenient to work 2.43 2.49
13 Courtesy buses and rental cars 2.37 2.35 11
14 Send out maintenance notices 2.05 3.33
a Ratings obtained from a four-pointscale of "extremelyimportant,"
"important,""slightlyimportant,"and "not important."
b Ratings obtained from a four-point scale of "excellent," "good," 14
"fair,"and "poor."A "nobasis forjudgment"category was also provided.

C. Low Priority D. Possible Overkill


Management Applications
Interpreting the Results Developing Marketing Strategies
Interpretationof the importance-performancegrid Importance-performanceanalysisprovidesmanage-
may be illustratedwith examples taken from each of ment with a useful focus for developing marketing
the four quadrants.
strategies. In the case of service prices, for example,
A. Concentrate here Customers feel that low at least three strategies, or combinations of
service prices (Attribute 9) are very important strategies, might be considered:
but indicate low satisfaction with the dealer's First The dealercouldmeet the priceissue head-on
performance. through informative advertising showing that the
B. Keep up with the good work Customers dealer's rates are lower than service station prices
value courteous and friendlyservice (Attribute or by running coupon specials for service work
6) and are pleased with the dealer's per- performed dack periods of the week.
formance. Second Given the high importance attached to
C. Low priority The dealer is rated low in terms doing the job right the first time (Attribute 1) and
of providing courtesy buses and rental cars fast action on complaints (Attribute 2), the dealer
(Attribute 13), but customers do not perceive might attempt to make customers feel that his
this feature to be very important. service rates are worth paying by improving his
D. Possible overkill The dealer is judged to bee performed during slack periods of the week.
doing a good job of sending out maintenance Third The dealer might lessen the impact of his
notices (Attribute 14), but customers attach perceivedhigh service pricesby selling customers an
only slight importance to them. (However, extended warranty which would reduce out-of-
there may be other good reasons for contin- pocket repaircharges and encouragetheir returning
uing this practice.) to the dealer for regular maintenance.

This content downloaded from on Sat, 28 Sep 2013 13:10:27 PM

All use subject to JSTOR Terms and Conditions
Importance-Performance Analysis / 79

Tips on Using Importance- Analyzing the importance-performance

is systematically accomplishedby considering each
PerformanceAnalysis attribute in order of its relative importance,moving
Determining what attributes to measure from the top to the bottom of the grid. Particular
is critical, for if evaluative factors important to the attention should be given to the extreme observa-
customer are overlooked, the usefulness of impor- tions since they indicate the greatest disparity
tance-performanceanalysis will be severly limited. between importance and performance and may be
Development of the attribute list should begin with key indicators of customer dissatisfaction.
identifying key features of the marketing mix.
Previous research in the same or related areas, Differences between loyal and disloyal cus-
various qualitative research techniques, such as tomer ratings may reveal important strategy impli-
focus groups and unstructuredpersonal interviews, cations as well as provide validity checks. In the
and managerialjudgment, all are useful in identify- above example, both groups rated low service prices
ing potentially important factors which might as being high in importance and low in performance,
otherwise be missed. These sources can also pro- so this attribute by itself would not appear to explain
vide guidance for screening the attribute list down differences in dealer patronage. Prompt warranty
to a manageablesize in order to avoid low response work, on the other hand, was also rated high in
rates and unnecessary data manipulation. importance by both groups, but loyal service cus-
tomers rated the dealer'sperformanceconsiderably
Separating the importance measures and the higher than the disloyalgroup. Ingeneral, when dif-
performance measures helps to minimize com- ferences are present between the ratings of loyal
pounding and order effects. If the respondent were and disloyalcustomer groups, there is greater confi-
asked in one question about the importanceof price dence that the research provides valid attribute
and in the next question about his satisfaction with measures influencing the buying decision.
current price levels, his answer to the first may
influence his response to the second. By grouping all Summary
of the importancemeasures in one section and all of
the performance measures in a later section, the Importance-performanceanalysis offers a number
of advantages for evaluating consumer acceptance
respondent moves in a natural progression from of a marketing program. It is a low-cost, easily-
general to more specific questions with a distinct understood technique that can yield important
separation between his ratings for each attribute. insights into which aspect of the marketing mix a
Positioning the vertical and horizontal axes firm should devote more attention as well as identify
on the grid is a matter of judgment.The value of this areas that may be consuming too many resources.
approach lies in identifying relative, rather than Presentation of the results on the importance-
absolute, levels of importance and performance. performancegrid facilitates management interpre-
Frequently a five- or seven-point scale will yield a tation of the data and increases their usefulness in
good spread of ratings, and the middle position will making strategic marketing decisions.
constitute a useful division for the grid. Occasion-
ally, as in the above example, the absence of low
importanceand performanceratings may argue for ENDNOTES
1. James H. Myers and Mark I. Alpers, "Determining
moving the axes over one position on the scale. Attributes: Meaning and Measurement," Journal of
Median values as a measure of central Marketing, Vol. 32 No. 4 (October 1968), pp. 13-20, and
tendency are theoretically preferable to means John G. Swan and Linda Jones Coombs, "ProductPer-
because a true interval scale may not exist. How- formance and Consumer Satisfaction:A New Concept,"
ever, the investigator may wish to compute both Journalof Marketing,Vol. 40 No. 2 (April1976), pp.25-33.
2. WalterJ. Salmon and Stanton G. Cort, Implicationsof
values and, if the two consistently appear reason- Car Owner After-Market Preferences for Car Dealer's
ably close, use the means to avoid discarding the Future Role in Automotive Distribution Boston, MA:
additional information they contain. Since tests of (HarvardU. GraduateSchool of Business Admin., 1971).
significance are not being used, distortions intro- 3. John A. Martilla and Davis W. Carvey, "Four Subtle
duced by minor violations of the interval-scale Sins in MarketingResearch,"Journalof Marketing, Vol.
39 No. 1 (January1975), pg. 10.
assumption are unlikely to be serious.3

This content downloaded from on Sat, 28 Sep 2013 13:10:27 PM

All use subject to JSTOR Terms and Conditions