Kevin Lane Keller, Susan E. Heckler, & Michael J. Houston ...assets.csom.umn.edu/assets/95269.pdf · The Effects of Brand Name Suggestiveness on Advertising Recall ... The choice

Kevin Lane Keller, Susan E. Heckler, & Michael J. Houston

The Effects of Brand NameSuggestiveness onAdvertising Recall

The authors report the results of a laboratory experiment examining the effects of the meaningfulness of brandnames on recall of advertising. The findings indicate that a brand name explicitly conveying a product benefit {e.g.,PicturePerfect televisions) leads to higher recall of an advertised benefit claim consistent in meaning with the brandname compared with a nonsuggestive brand name (e.g., Emporium televisions), Conversely, a suggestive brandname leads to lower recall of a subsequently advertised benefit claim unrelated In product meaning (e.g.. superiorsound) compared with a nonsuggestive brand name. The authors discuss implications of these findings for mar-keters with respect to advertising strategies and the optimal use of meaningful brand names in building and man-aging brand equity.

Brand names come in many different forms—they canbe based on real peopie, places, animals, birds,things, and objects or just be made up. The choice of

a brand name has been suggested as one important means lobuild brand equity for a new product (Aaker 1991. 1996;Keller 1993, 1998). Choosing the proper brand name—oftenthe centerpiece of introductory marketing programs—canenhance brand awareness and/or help create a favorablebrand image for a newly introduced product. Recognizingthe important and complex role of brand names as part ofmarketing strategy, several different possible criteria havebeen proposed for choosing brand names to build brandequity (Robertson 1987).

One oflen-noted branding objective is to choose "inher-ently meaningful" brand names, so that the name itself con-veys relevant product information. Brand names can bemade meaningful in a variety of different ways. For exam-ple, brand names can be ehosen to reinforce semantically thecorresponding product category (e.g.. Lean Cuisine low-calorie frozen foods. JustJuice juices. New.sweek weeklynews magazine) or a particular attribute or benefit makingup the main selling point of a brand (e.g.. DieHard auto bat-teries. Mop'n Glow floor cleaner. Beautyrest mattres.ses).The first branding strategy should enhance brand nameawareness and identification with the product category. Ourinterest is in the second branding strategy^which we referto as choosing a "suggestive" brand name—^and its costs andbenefits.

Kevin Lane Kelier is Visiting Professor of Marketing, Fuqua School of Busi-ness, Duke Universily, Susan Heckler is Associate Professor of fvlarketing,Karl Eller Graduate School of Management, University of Arizona. MichaelJ, Houston is Ecolab/Grieve Chair in International Marketing, Curtis L.Carlson School of Management, University of Minnesota, Comments fromMerrie Brucks; Dipankar Chakravarti; Brian Gibbs; participants of semi-nars al Ihe Universily of California at Irvine, Duke University, and the Stan-ford Marketing Camp; and the JM reviewers are greatly appreciated.

48 / Journal of Marketing, January 1998

To illustrate some of the key issues involved with sug-gestive brand names, assume that a new brand of luggage isto be positioned initially as "durable." In sucb a case, is iteasier to develop a strong brand image initially by giving ita brand name suggestive of that positioning, such as Life-Long, compared with giving it a nonsuggestive brand name,such as Ocean? On the basis of an associative strength the-ory of memory, we argue that judiciously choosing sugges-tive brand names can facilitate initial brand positioning.Under the same scenario, if the brand later were to be adver-tised as having a "fashionable appearance." which brandname would be more beneficial? If branding the luggage asLifeLong initially does create stronger brand associations inmemory, do these more developed knowledge structuresfacilitate the linkage of new associations?

On the basis of interference theories of memory, weargue that choosing suggestive brand names actually canhamper subsequent marketing communicaiion elTorts toreposition the brand in new, unrelated directions. That is,when strong links have been formed in memory between asuggestive brand name and its original product positioning,consumers might fail to create new brand associations whenexposed to marketing communications designed to reposi-tion the brand. Moreover, even if new brand associalions areformed, consumers might overlook them in favor of brandassociations related to the original product positioning whenlater tliinkiiig about the brand. Consequenlly. it may be eas-ier to add new brand associations if the luggage initially isnamed Ocean than if named LiteLong.

Thus, our general researcb interest is in how the sugges-tiveness of a brand name influences markeiing communica-tion effectiveness. Our specific researcb questions arewhether (1) a suggestive brand name facilitates initial posi-tioning of a product, but (2) a nonsuggestive brand namefacilitales laler repositioning if necessary. To address theseissues, Ihe article is organized as follows: First, we reviewrelevant prior re.search to develop hypotheses; after describ-ing the experimental methodology and results, we discussimplications of the research findings.

Journal of MarketingVol. 62 (January 1998). 48-57

Conceptual BackgroundIndependcnl ol ihc decisions made about the product andhow il is marketed, brand names can be chosen to build asnuicb brand equity as possible. That is, brand names can bechosen (1) that are inherently memorable and thereforefacilitate recall and/or recognilit)n in purchase and/or con-sumption settings, and/or (2) whose inherent meaningsenhance the formation of strong, favorable, and uniquebrand associations consistent with that meaning.

A memorable and meaningful brand name offers manyadvantages. Because consumers often do not examine muchinloriiiatit)n in making product decisions, brand names mustbe recognized and recalled easily and be inherently descrip-tive and persuasive. Moreover, memorable or meaningfulbrand names can reduce the burden on marketing communi-cations to build awareness and link brand associations. Thedifferent associations tbal arise from the likability andappeal of brand names also can play a critical role in theequity of a brand, especially when few other brand associa-tions exist in memory.

Suggestive Brand NamesTlius. one key consideration in choosing a brand name is theexlcnl to which it conveys descriptive or persuasive infor-mation. First, in terms of descriptive meaning, to whatexient docs the brand name suggest something about theproduct category? How likely would it be that a consumercould identify correctly the corresponding product categoryor categories for the brand on the basis of the brand namealone,' Rclatedly, does ihe brand name seem credible in tbeproduct category.' hi other words, is the content of the brandname consistent with what consumers would expect or wantto see from a brand in that product category,' Second, interms of persuasive meaning, to what extent does the brandname suggest something about the particular kind of prod-uct ihe brand is—for example, in terms of key attributes orbenefits? Does it suggest something about a product ingre-dient or the type of person wbo might use the brand?

A ,vi(i.',t,'f .v/(iT brand name can be defined as a brand namethai conveys relevant attribute or benclil informalion in aparticular product context. For a brand name to be deemedsuggestive. It must bave well-defined associations or mean-ings thai could be seen as relevant in a product selling. Inotber wt)rds, associalions for a suggestive brand name mustbe botb salient and relevant in a particular product context.A meaning-hiden name might be suggestive in dilTerentways depending on tbe product context involved and theparticular associations tbat are evoked. For example, though"Chier" mighl convey stain removal when used as a brandname i)f detergent, it niiglil convey heritage and leadershipif used as a brand name of gasoline. Thus, in some sense,there is a conlinuuni of ihe suggestiveness ofa brand name,ranging trom ct^nipletely nonsuggesiivc lo highly sugges-tive, depending on tbe latent and cvocable meaning of thename as well as the product category contexts involved.

A brand fuinic thai is suggestive in a pariicular productcategory should offer two important benefits. First, even inIhc absence of any marketing activity, the semantic meaningof a suggestive brand name should enable consumers toinlcr certain attributes or benefits. For example, consumers

might assume on the basis of their names alone that Day-break cereal is wholesome and natural. Chief laundry deter-gent removes tough stains, and Diamond toothpaste whitensand brightens teeth. Second, the suggestiveness or meaning-fulness ofa hrand name should facilitate marketing commu-nication efforts designed to link corresponding attributes orbenefits to tbe brand (e.g., claims related to product attrib-utes or benefits). Research has shown that meaningful brandnames tbat are visually represented are easier to remember(cf. Childers and Houston 1984; Lutz and Lutz 1977) andcan enhance development of memory structures for brand-related infonnation communicated in advertising. Further-more, in a series of studies, Keller (1987, 1991a. b) showsthat nonsuggestive brand names C4)ntaining no productmeaning often serve as poor reminders for communicationeffects stored in memory. In sucb cases, supptLMiieiitaiyretrieval cues containing information more specificallyrelated to the advertisements themselves are necessary forsuccessful recall of tbe corresponding brand claims.

These research findings are consistent with the associa-tive strength theory of information recall, which maintainsthat the effectiveness of a retrieval cue depends on howstrongly associated it is with target information (iillis andHunt 1983). According to this theory, a brand name thatsemantically suggests a product benefit might be associatedmore strongly in memory antl facilitate recall of that bene-fit, especially in those situations in which tbe brand name isused as a retrieval cue by consumers (e.g.. when making anin-store decision). This reasoning suggests the followinghypothesis:

H|: Suggestive brand names will lead in greater recall ofadveiiised henelll claims that are eun.sistent in incaTiingthan nonsuggcslivc brand names.

Brand Repositioning

Because ol changing consumer needs, competitive actions,or any other changes in the marketing environmenl overtime, managers may need to reposition their brands throughnew niarkeling ct)mmunication campaigns. A brand mayneed to advertise new claims to link associations that func-tion either as additional points of difference for the brandversus competitors or as points of parity designed to negatecompetitors' intended points of difference. One problern thatpotentially can arise wben advertising attempts to link anew, unrelated brand association is thai consumers stillmighl coiiiinue to ihink of the brand in the "old way"because of strong associations already in tnemory. As aresult, con.sumers might fail to incorporate new advertisinginformalion into iheir brand knowledge siruclures or tail toretrieve new advertising infomiation when making laterbrand-related decisions.

InU'/fercncf effecis. One important cause of forgettingof information is "interference" as a result of additional,related informalion in memory (Muniane and ShilTrin 1991;Postman and Underwood 1973). Two types of inierlerenceexist: Proactive interference ari.ses from existing inlorma-tion in memory encountered before exposure to target infor-malion; rcinniciive interference arises trom new intbrnia-tion encountered after exposure to target information. A

Brand Name Suggestiveness / 49

variety of mechanisms help explain interference effects. Forexample, related information already existing in memorycan result in weaker associations to target information dur-ing initial encoding (Melton and Irwin 1940) and/or theinability to access target information during later retrieval{Tulving 1974). Retrieval-based explanations are based onthe notion of response competition (McGeoch 1942). Forexample, the associative network memory model predictsthat the more pieces of information linked to a particularnode, the more likely it is that the "spreading activation"elicited from that node when it is cued will fail to reach thethreshold level necessary for recall of any one specific pieceof information.

In marketing, interference effects due to competitiveadvertising have been demonstrated (Burke and Srull 1988;Keller 1987. 1991b): The more competing brands advertis-ing in a product category, the greater the likelihood that thetarget brand and its advertised claims will be either confusedwith other information or inaccessible. In other words, ifmultiple brands advertise in a product category, overlappingadvertising memory traces might confuse the correspon-dence of advertisements and brands in the product categoryor result in sufficiently weak associations between brandsand their advertising memory traces so that some communi-cation effects cannot be recalled at all. Besides advertise-ments for other brands. Burke and Srull (1988) also showtbat anotber possible source of interference effects is fromrelated advertising for different products for the same brand;that is, advertising for different models in a brand's productline might interfere with one another.

Effects of prior advertising. Along these lines, previousadvertising for a brand could inhibit recall of a newly adver-tised claim for tbe brand that is unrelated in meaning. Exist-ing brand associations created by old advertising campaignscould create proactive interference effects, so that con-sumers fail to form strong brand associations to newlyadvertised benefit claims during encoding or to access anynewly stored communication effects during later retrieval.In terms of the latter effects, research has shown that factorsthat increase recall of some infomiation from an advertise-ment can reduce recall of other infomiation from that adver-tisement (Alba and Cbattopadhyay 1985; Keller 1991a).These inhibition effects might result because people eitber"fixate" on information that they can recall, which makes itmore difficult to recall other information not yet recalled, orfail even to try to search memory for other infonnation.

Thus, strong associations to the original benetit claimsmight make it difficult to link new associations at encodingor to access new benefit claims at retrieval. The.se interfer-ence or inhibition effects may be particularly likely withsuggestive brand names. As stated in H|. initially advertis-ing a benefit claim consistent in meaning with a suggestivebrand name should lead to strong brand associations andfacilitate later recall of that claim. If a suggestive brandname serves as a strong cue in part because of its semanticmeaning, bowever. it may continue to facilitate recall of pre-viously advertised benefit claims and inhibit recall of newbenefit claims tbat are unrelated in meaning. A nonsugges-tive brand name, conversely, will not provide the samesemantic cue, whicb suggests that consumers might be less

likely to overlook newly learned product information. Thus,a nonsuggestive brand name might be more tlexiblc andaccommodate more easily subsequently advertised benefitclaims. Thus, in repositioning a brand, the followinghypothesis can be made;

H : Suggestive brand names will lead to lower recall of new.unrelaied benefit claims than nonsuggestive brand names.

Methodology

ProcedureOne hundred sixty adult subjects enlisted from the l<jcaluniversity community participated in the study in return for$10 and a chance to win cash prizes. Eight-five percent ofthe sample were female. 30% had college degrees, andapproximately half of the sample were more than 35 yearsof age. Subjects were given the cover story that an electronicinfonnation and shopping service was being developed thatthey were being asked to evaluate (Keller 1987; Ray andSawyer 1971). With one feature of this service, subjectswere told, consumers would be able to examine magazine-type advertisements on their television screens by means ol"a computer hookup. These advertisements could berequested for a particular brand, for all brands within a prod-uct category, or in a random order, as determined by tbe ser-vice providers. Thus, consumers might see an advertisementmore than once or see more than one advertisement for abrand, just as they would on television or in a magu/ine.

The experiment was conducted in two sessions over thecourse of three days. Subjects were told in the first sessionthat they would view some illustrative advertisements forihe electronic inlbrmation and shopping service and beasked to evaluate the advertised brands. After viewing eachof nine advertisements for 30 seconds, subjects providedtheir overall evaluations of the brands as a whole and theirgeneral reactions to tbe proposed new service (in terms ofperceived advantages and disadvantages). Subjects thenwere dismissed and asked to return two days later to answeradditional questions about the service. At this second ses-sion, subjects viewed eacb of nine advertisements for 30seconds, provided additiona! general evaluations, and com-pleted a five-minute filler task regarding otber aspects of theservice. Memory measures then were collected for five tar-get brands.

Measures

The primary dependent measure was an aided recall mea-sure of main benefit claims. Subjects listed the appropriateclaims for the target brands when given product categoryand brand name cues. Subjects were asked to indicate allclaims recalled for a brand if more than one claim had beenadvertised. Research assistants unaware of the experimentaldesign and manipulations coded subjects' responses on tbebasis of the "gist" of the protocol in terms of correct recallof none. one. or both (if appropriate) of the advertised brandclaims.

Covariate measures also were collected as proxies lorsubjects' inherent product category involvement and atti-

50 / Journal of Marketing, January 1998

[udes thill potenlially could affect tbeir responses to adver-tising. Tliose t'ivt: measures, a.ssessing subjects' purcbuseand usage frequency, knowledge, importance of brand selec-tion, and perceived quality differences among brands ineach product category, were combined to fomi a scale ofcategory involvement witb a satisfactory coefficient alpbareliability estimate of .72.

Stimuli

I'ive target product categories were cboscn among durablegoods witb wbich subjects could have bad some involve-ment and of sufficient complexity to justify advertising mul-tiple benefits: cameras, luggage, personal computers, televi-sion, and tennis racquets. New ad vcrti.se men is were ereatedfor these products, eacb witb an identical format. The tophalf contained a col or-re produced photograph of the product(with no brand identification). The bottom half of ihe adver-tisement contained a headline, one paragraph de.scribing tbemain benetil of the advertisement (cither a consistent orunrelated product benetit), and a concluding tag line sum-marizing the main claim. The brand name was mentionedfour times in tbe advertisement (see Figure I).

Suggestive brand names were created by explicitlyembedding key product benefits into the name (e.g.,PicturePcrfect televisions). Thus, suggestive brand nameswere botb salient and relevant in the product category. Non-suggestive brand names were chosen to contain no productmeaning (e.g.. Emporium televisions). One set of advertisedbenefits was eboscn to be consistent in meaning with sug-gestive brand names (e.g., superior picture for televisions).Another set of benefits was ebosen to be unrelated in mean-ing to suggestive brand names (e.g.. superior sound for tele-visions) and. by definition, the nonsuggestive brand names.

Pretests confirmed these differences. Specifically, 53subjeets rated tbe likelihood that products given differentbrand names would possess various benefits on nine-point.scales (1 = extremely unIikciy/9 = extremely likely). Thetesults indicated that suggestive brand names (M = 7.75)were rated as significantly more likely (p < .05) to possessthe consistent henelit tban nonsuggestive brand names (M =5.00). There were no significant differences (F < 1), how-ever, between suggestive brand names (M = 5.25) and non-suggestive brand names (M = 5.00) for the likelihood ofpossessing unrelated benefits. The pretests also indicated(bat tbe conditional probabilities (i.e., tbe likelihood that anadvertisement for benefit 1 would mention benefit 2) wereroughly average (at tbe scale midpoint) tor the two sets ofconsistent and unrelated benefits, which suggests that tbebenetits were, as desired, essentially uneorrelated.

For each of tbe five target product categories, fouradvertisements were created. Two advertisements containedIhe consistent benetit claim with either the suggestive ornonsuggestive brand name, and two advertisements cun-lained the unrelated benefit claim with either the suggestiveor nonsuggestive brand natne. Seven filler advertisementswere created witb the .same basic format for golf clubs,leather handbags, computer software, stereo receivers,stereo speakers, tennis balls, and watcbes. In each session,two filler advertisements were seen first to control for pri-macy effects, and either two or three filler advertisements.

depending on tbe session and tbe ad exposure group (asdescribed subsequently), were seen last to control forrecency effects. Table I contains a summary of the adver-tisement and brand .stimuli.

Research Design and Manipulations

The objective behind tbe experimental design was to maxi-mize external as well as intemal validity. As a result, greatcare went into creating experimental conditions tbat wouldpermit fair and representative tests of the bypotbeses.Specifically, two factors were manipulated: (I) lirandnamesugjiestiveness in terms of whether a brand name explicitlyconveyed a product benefit and (2) ad exposure sequence interms of tbe number, nature, order, and liming of advertise-ments. Specifically, tbe design was a 2 (brand name sugges-liveness: suggestive versus nonsuggestive) x 12 (ad expo-sure sequence) incomplete block design. The ad exposuresequence depended on the number of advertisements seenfor tbe brand (1, 2, or 3), tbe panicular bcnclit claims thatwere advertised (consistent or unrelated in meaning with tbesuggestive brand name), and tbe timing and order in wbichtbe advertisements were seen during the two sessions.

Brand name suggesliveness was manipulated betweensubjects, so that balf of the sample viewed advertisementsfor products thai used suggestive brand names and tbe otherhalf viewed advertisements for products that used nonsug-gestive brand names. Brand names for filler advertisementswere .selected to correspond in suggestiveness with thebrand names of tbe target advertisements to avoid undueattention to tbe nature of target brand names.'

Tbe ad exposure sequence was a mixed witbin- andbetween-subjects manipulation in whieb each subject wasassigned to one of four ad exposure groups so tbat tbey wereexposed to and provided measures for 5 of tbe 12 differentpossible ad exposure .sequences. Each sequence Ineludedadvertisements for a brand in one of the five different prod-uct categories. The 12 ad exposure sequences are shown inTable 2.

To illustrate, subjects wbo saw sequence 7 as one oftheir five ad exposure sequences would have seen an adver-tisement for a brand in one product category thai promotedthe benefit claim consistent in meaning witb tbe suggestivebrand name in tbe first session and in the second sessionwould have seen an advertisement for the same brand tbatpromoted tbe benefit claim that was unrelated in meaning tothe suggestive brand name.

Single and double exposures were employed for botb theoriginally advertised and newly advertised benefit claims.Incorporating ad repetition levels pemiils a wider range ofbrand association strength, so tbat more informative tests ofbypotheses are possible. Employing a variety tit' exposuresequences also better approximates ibc realities of consumerexperiences in tbe marketplace, wbere campaigns for ini-

'The suggestive and nonsuggestive brand names for the fillerproduct categories were, respectively. Hori/on and Lanford (golfclubs). Guardian and Gallery (leather handhag.s). UniSolution andTennant (e.g.. computer .software). FlexiSoimd and Sovereign(stereo receivers). Harmony and Glaser (stereo speakers). Sure-Bounce and Coopers (lennis balls), and Eternal and Medallion(watches).

Brand Name Suggestiveness / 51

FIGURE 1Advertising Stimuii Example

r .

LifeLong Luggage is so durablethat it will be your travel

companion—for life!

Heavy-duty Duktex fabric and specialwebbing make LifeLong Luggage tough

enough for all kinds of travel.This durable construction ensures that

LifeLong Luggage w ill withstandthe wear and tear of repeated use.

Count on LifeLong Luggage to protectall of your personal possessions

when you travel!

52 / Journal of Marketing. January 1998

TABLE 1Advertising Stimuli Summary

ProductCategory

Suggestive andNonsuggestiveBrand Names

ConsistentBenefit Claim

UnrelatedBenefit Claim

Camera

Luggage

PersonalComputer

Television

TennisRacquets

EasyProWatson

LifeLongOcean

CompuQuickCriterion

Picture PerfectEmporium

PowerStrokeCrown

Ease of use

Durability

Speed andease of use

Superiorpicture

Powerfulperformance

Flexibility andattachments

Fashionableappearance

Compatibility

Superiorsound

Special designfor men and women

tiaily or subsequently advertised benefit claims may bemore or less salient, depending on the particular budget,media buy. and so fortb. Conducting tbe study over two ses-sions allows for some separation between tbe initial posi-tioning and the later repositioning, which increases the like-lihood that the initially advertised claims become sutTi-ciently strongly encoded in metiiory.

Single and double exposures to advertisements withconsistent or unrelated benefit claims wore used as bench-marks or points of reference to compare with the effects onmemory of also baving advertised the other claim eitheronce or twice. Kach sequence provides informalion relatedto the research hypotheses. Sequences 1-6 permit tests ofH] regarding the effects of brand name suggestiveness onthe recall of initially advertised benefit claims. Sequences7-12 pcnnit tcsis of Hi regarding the effects of brand namesuggestiveness oti tbe recall of subsequently advertised ben-efit claims.

TABLE 2Experimental Design: Ad Exposure Sequences

Sequence Session 1 Session 2

(1)(2)(3)(4)(5)(6)(7)(8)(9)

(10)(11)(12)

CU

uc

OCcu

uuu

c•$

u

uu

uucc

CC

Nole. C represents exposure to an advertisement promoting a ben-efit consistent in meaning with the suggestive brand name; Urepresents exposure to an advertisement promoting a benefitunrelated in meaning with the suggestive brand name.

To itiipletnent the ad exposure sequence manipulation,subjects were assigned to otie of four ad exposure groups sothat they saw two one-advertisement sequences (i.e., two ofsequences 1. 2, 4, or 5), two two-advcrtiscment sequences(i.e.. two of sequences 3. 6, 7, or 10). and one three-advertisement sequence (i.e., one of sequences 8, 9. II. or12). Let Cf and Uf represent single ad exposures to consis-tent and unrelated benefits, respectively, in ibe first sessiont)nly; let Cs and Us represent single ad exposures to consis-tent and unrelated benefits, respectively, in the second ses-sion only; and let "\" separate exposures in the first sessionfrom those in the second session. Schematically, ihe four udexposure groups saw different ad expt)surc sequences acrossfive differetit product categories, as shown in Table 3.

Thus, subjects in the first ad exposure group saw anadvertisement lor a consistent benefit claim in the first prod-uct category (Cs). an advertisement for an unrelated benefitclaim in the second product category (Ul"). (wo udvcrtise-nicnls for a consistent benefit claim in the thirJ product cat-egory (C|C), an adverti.sement for an unrelated benefit claimfollowed by an advertisement for a consistent benefil claimin the fourth product category (U|C), and an advertisementfor a consistent benefit claim followed by two advertise-ments for an unrelated benefit claim in the fifth product cat-egory (C|UU).

Product category also was counterbalanced betweensubjects by rotating the different products and brands

TABLE 3Experimental Design: Ad Exposure Groups

Groups

1234

1

CsUsCfUf

Product

2

UfCfUsCs

Category

3

C|CU|UC|CU|U

4

U|CC|UC|Uuic

5

CjUUU|CCuuicCC|U

Brand Name Suggestiveness / 53

through the ad exposure sequences in each ad exposuregroup so that each product category was associated witheach ad exposure sequence an equal number of times. As aresult, responses to any one sequence were ba.sed on reac-tions by different groups of subjects to each of the five dif-ferent brands and product categories, which increased thegeneralizability of the findings.

Moreover, to ensure that subjects saw a varied set ofadvertisements, the ad exposure sequences were combinedso that target advertisements for the different brands andproducts were intermingled and combined with filler adver-tisements. Specifically, subjects saw advertisements for theparticular ad exposure .sequences in their ad exposuregroups in the order shown in Table 4.

ResultsHI and Hi were tested through a series of two-sided plannedcomparisons conducted by subject and target advertisement.All contrasts were between subjects and used the within-subjeets error sum of squares from the overall ANCOVAanalysis—consistent with guidelines by Keppel (1982, pp,428-32)—with I and 766 degrees of freedom in the numer-ator and denominator, respectively. The ANCOVA analysis

included product category type and product categoryinvolvement as covariates.- Table 5 contains cell means forthe key dependent measure—aided recall of main benefitclaims.

Initially Advertised Benefit Claims

H| hypothesizes that suggestive brand names cue advertisedbenefit claims consistent in meaning more effectively thando nonsuggestive brand names. H| was tested by comparingrecall of consistent benefit claims for ihe appropriate single-and double-exposure conditions (i.e., the Cf, Cs. and C|Csequences) for suggestive (SBN) and nonsuggestive (NBN)brand names. As hypothesized, recall of the consistent ben-efit claim when it was advertised in the first session was sig-nificantly higher for a suggestive brand name than for a non-suggestive brand name (SBN = .33. NBN = .15: F = 2.74./?< .05). There were no significanl difterences. however, inrecall of the consistent benefit when it was adverti.sed in the

-Significant effect.s were observed for the recall measure forbolh the pmducl category invotvemenl j i iJ prcului;! category typefovariatcs. which indicates ihal the iidvcrtisemcnls ditfcrcd some-what in their memorability aiiiJ higher levels ul involvement werea.ssociated with greater recall.

TABLE 4Experimental Design: Ad Exposure Order

AdExposureGroups

1234

1

Li.

FFF

2

FFFF

3

U,C ,

Ul

Session 1

4

C,u.Us

5

u.

C3

U3

6

c.u.C4U4

7

FFu.

8

FFFF

9

FFFF

1

FFFF

2

FFFF

3

c,

u,C2

Session 2

4

Us

C3U3

5

C3U,U4C4

6

C4U4

Us

7

Us

FF

1

8

FFFF

9

FFFF

Note: 1, 2. 3, 4, and 5 represent the five product categories; F represents a filler advertisement; last exposure of each sequence in bold,

TABLE 5Aided Recall Measure of Main Benefit Claims:

Cell Means (Standard Deviations)

AdExposureSequence

c,CsCC

u,UsU|UC|UcciuciuuuicuuicU|CC

N

404040404040402020402020

Unrelated

NonsuggestiveBrand Name

———

.08 (.27)

.43 (.50)

.45 (.50)

.50 (51)

.50 (.51)

.35 (.49)

.15 (.36)

.20 (.41)

.16 (.37)

Benefit Claim

SuggestiveBrand Name

——

.05 (.22)

.38 (.49)

.35 (.48)

.35 (.48)

.20 (.41)

.30 (.47),13 (,33),15 (,37),05 (22)

Consistent

NonsuggestiveBrand Name

.15 (.36)

.35 (.48)

.45 (.50)———

.23 (.42)

.20 (.41)

.25 (.44)

.50 (.51)

.55 (.51)

.50 (51)

Benefit Claim

SuggestiveBrand Name

.33 (47)

.40 (.50)

.54 (.51)———

.45 (.50)

.40 (.50)

.50 (.51)

.48 (.51)

.60 (50)

.65 (,49)

Note: C represents an advertisement for a consistent benefit claim; U represents an advertisement for an unrelated benefit claim; Cf and Uf rep-resent single ad exposures to consistent and unrelated benefits, respectively, in the first session only; Cs and Us represent single ad expo-sures to consistent and unrelated benefits, respectively, in the second session only; "|" separates exposures in the first session from thosein the second session.

54 / Journal of Marketing, January 1998

second session (i.e.. the Cs or C|C sequences), perhapshecause ihe advertising inlormation was too sahent andaccessible in memory—the delay trom ad exposure to recallmeasurement for advertising claims from the second sessionwas only tive minutes or so compared with iwo days foradvertising claims from the first session. Nevertheless, theresults prtwide at least some support for Hi and the notionIhat a suggestive brand name serves as a more effective cueto advertised brand claims consistent in meaning with thathrand name.

There were no significant differences in recall of theunrelated benetlt claim between the suggestive and nonsug-gesllve brand names (/» > .20) for the appropriate single- anddouble-exposure conditions (i.e., the tjf, Us, and U|Usequences). Thus, as was expected, both types of brandnames were equally effective at cueing advertised brandclaims unrelated in moaning with the two types of brandnames.

Subsequently Advertised Benefit ClaimsHi hypothesizes that a suggestive brand name, comparedwith a nonsuggestive brand name, results in lower recall ofadvertised benefit claims unrelated in meaning to the brandname if benefit claims consistent in meaning with the sug-gestive hrand name already have become associated withthe brand through prior advertising. Given the indicationthat a suggestive brand name improved recall of a consistentbenefit claim that was advertised in the first session, it isajiprupriate to consider ihe effects on consumer memory ofsubsequently advertising a new benefit claim in the secondsession. Hi was tested hy comparing the recall of unrelatedand ctinsistent benefit ctaim.s for suggestive versus nonsug-gestive brand names in the case in which the consistent ben-efit claim was advertised in the firsl session, but the unre-lated benefit claim was advertised in ihe second session (i.e.,the CC|U, C|U. and C|UU sequences).

Consistent with Hi. recall ot the unrelated benefit claimwhen it was advertised in the second session was signifi-cantly lower for the suggesiive brand name than for ihe non-suggestive hrand name tor bolh ihe CC|U sequence (SBN =.20. NBN = .50; F = 7.02, p < .01) and the C|U sequence(SBN = .35. NBN = .50; F= 2.95./) < .00). Thus, though thetwo lypes of brand names served as equally etfective cues tounrelated benefit claims when there had been no prioradvertising, when a claim consistent in meaning with thesuggestive brand name already had been advertised in thefirst session, recall of the unrelated claim was significantlylower for the suggestive brand name compared with thenonsuggestive brand name. The difference in recall was nolsignificant in the C|UU sequence (SBN = .30. NBN = .35; F< 1), however, perhaps because ihe douhle exposure in thesecond session made the unrelated claim sufficiently salientand accessible.

Conversely, recall of the consislenl benefit claim thatwas adverti.sed previously in the first session was signifi-cantly higher for a suggestive brand name than for a non-suggestive brand name for the CCjU sequence (marginally.SBN = .40, NBN = .20; F = 2.51. / x . 11). the C[U sequence{SBN = .45. NBN = .23; F = 4.55, p < .05). and the C|UUsequence (SBN = .50, NBN = .25; F = 3.45, p < .Ob). Com-

bined, these analyses imply ihat the suggestive brand name,compared with the nonsuggeslive brand name, was morelikely to continue to access the original benefit claims alterrepositioning at the expense of recall of the new benetiiclaims. These results lend additional support for H| and thereasoning behind H2.

Finally, to gain additional insight into how brand namesuggestiveness affected consumer memor)' perfomiance. itis also instructive to look at those ad exposure sequences inwhich the unrelated benefit claim was advertised first fol-lowed by advertising for the consi.stent benefit claim (i.e.,the UU|C, U|C, and U|CC sequences). There were no signif-icant differences in recall for either the unrelated elaims orthe consistent claims between suggestive and nonsuggestivebrand names for any of the three sequences (F < 1).

There are several possible inicrpretations for the lack ofdifferences between suggestive and niuisuggestive brandnames for recall of the unrelated brand claims in thesesequences. It might indicate that Mule retroactive interfer-ence eftects exist for a suggestive brand name. In otherwords, if an unrelated claim already has hecome associatedwith a suggestive brand name, then subsequent advertisingof a consistent claim does not necessarily inhibit later recallofthe unrelated claim. Altematively, it might just be a resultof the fact that relatively low levels of recall occurred forunrelated claims when they were shown in the first session(i.e., "basement" effects were present).

Relatedly, the lack of differences between suggestiveand nonsuggestive brand names for recall of the consistentbrand claims in these sequences once again might reflect thefact that, because these claims had been shown in the secondsession, they were relatively salient and accessible duringthe recall task regardless of the nature of the correspondingbrand name.

DiscussionSummaryA suggestive brand name was defined as a brand name tiiatconveys relevant attribute or benefil intbrmaiion in a partic-ular product context. To examine the eltectiveness of sug-gestive and nonsuggestive brand names as cues to adver-tised product claims, we conducted an experiment thatmanipulated whether a brand name explicitly conveyed aparticular product benefit and the number, nature, order, andtiming of advertised benefit claims for llie brand. The resultsindicate that a suggestive brand name—compared with abrand name that contains no product meaning—can

1, facilitate recall of initi;illy iKivoriisod hcnetlt claims consis-tent in meaning with ihc bi'LiiRJ naim: but

2. inhibit recall of subsequcnily advenised benetli claims unre-Inted in meaning to the t>rand name.

The capability of suggestive brand names lo cue adver-tising infonnation efteetively extends the research of Keller(1987, 1991a. b), who showed that iioiisuggesiive names canbe inettective advertising retrieval cues. Conversely, the factthat strong associations in memory wiiti a henefii claim con-sistent in meaning with a suggestive brand name loweredthe recall of a subsequently advertised benefit claim that

Brand Name Suggestiveness / 55

was unrelated in meaning extends prior research by demon-strating interference and inhibition effects in a differentdomain than has been shown previously in consumer behav-ior research.

Implications

These study lindings have important implications for mar-keters regarding advertising strategies and the optimal useof meaningful brand names in building and managing brandequity. Re.search in this area has pointed out that one way tobuild brand equity is throtigh the initial choice of brand ele-ments—for example, the brand name, logo, or symbol(Aaker 1991, 1996; Keller 1993. 1998). This research pro-vides some guidance to the naming decision by consideringhow different brand name strategies affect consumer mem-ory for advertising effects.

On the one hand, choosing a brand name that is concreteand evokes imagery that suggests a certain product benefitcan, by producing strong brand associations, contribute tobrand equity by facilitating initial positioning. On the otherhand, choosing a brand name thut suggests a certain productbenefit can, by producing interference and inhibition effectsin memory, affect adversely the capability of advertising tolink new brand associations at a later time. Such processeshelp explain why a brand such as Jack-in-the-Box restau-rants has found it difficult to establish a more adult, product-focused image; why brands such as Old Spice after-shave,Oldsmobiie automobiles, and John Hancock financial ser-vices have struggled to create more youthful images; andwhy Hidden Vulley Ranch salad dressing has encounteredproblems in expanding beyond its flagship ranch flavors.

One implication of these research findings is that mar-keters may be better off adopting more flexible brandingstrategies when introducing new products by using nonsug-gestive brand names if they anticipate the possibility of laterneeding to advertise additional benefit claims. In otherwords, it is important in choosing a meaningful brand nameto consider the possible contingencies in later repositioningor other associations that may become relevant or desirable.Consumers might find it more difficult to accept—or justtoo easy to forget—the new posiiioning if the brand namecontinues to remind them of other product considerations.

Alternatively, if marketers choose suggestive brandnames to introduce new products, they must be willing tocommit enough time and resources to reposition the brand ifit laier becomes necessary or must be willing to introducenew brands or sub-brands to capture product positions thatwould be difficult to attain with existing meaning-ladenbrand names. Witb sufficienl time and properly designedand supported marketing programs, there is at least anecdo-tal evidence that tbe restrictive nature of suggestive brandnames can be overcome. For example. Compaq computersinitially was named to convey the fact that it was a smallcomputer. Through subsequent introductions of "bigger""personal computers, advertising campaigns, and other mar-keting activity, Compaq seems to have been able to tran-scend the initial positioning suggested by its name as amaker of only small computers. Nevertheless, sucb market-ing maneuvers could be a long and expensive process—imagine tbe difficulty of repositioning brands such as I

Can't Believe It's Not Butter margarine or Gee, Your HairSmells Terrific shampoo.

t-

Limitations and Further ResearchThe previous discussion ol research implications should beinterpreted in light of the limitations of the experiment ihatwas conducted. These considerations and others suggestseveral future research opportunities. One research priorityis to provide a more detailed theoretical account of exactlyhow the suggestiveness or any other properties of a brandname influences marketing communication effectiveness.A clearer understanding of the role of encoding andretrieval—and other possible mediating factors—mostlikely will require experimental manipulations and mea-sures carefully designed to affect and capture particularmemory processes.

In a more specific sense, it is also of interest to examineother aspects of the manipulatit>ns of brand name sugges-tiveness and the ad exposure sequences and other possiblemoderating factors to assess the robustness of ihese facili-tating and inhibiting effects. For example, il may be that asuggestive brand name still can serve as a better cue afterrepositioning, compared with a nonsuggestive brand name,if subsequently advertised benclit claims are more consis-tent ill meaning with brand knowledge from previouslyadvertised claims and the brand name itself. In other words,the intt^rference effects encountered for subsequently adver-tised benefit claims may be less prevalent with new benefitclaims that are more consistent in meaning with existingbrand knowledge structures, because consumers will be bel-ter able to use this knowledge.

It is also important to explore the generalizability ofthese effects to real-world situations in which consumersmight have strong prior associations built up in memory(due to numerous repeated exposures to advertisements andproduct usage) and/or might be targeted by concerted, well-executed marketing programs designed to rept)sition thebrand. In terms of the former consideration, Kent and Allen(1994) find that interference effects from competitive adver-tising are substantially reduced for familiar brands com-pared with unfamiliar brands. Similarly, to the extent that asuggestive brand name has built up rich brand knowledgestructures in consumers' minds, tbere may not be as muchdifficulty in linking new associations lrom repositioning, interms of tbe latter consideration, one key communicationdecision in repositioning a brand is the extent to which thebrands previous positioning and heritage is explicillyacknowledged and addressed in a new advertising cam-paign. Prior research on one-sided versus two-sided argu-ments in advertising might be illuminating there.

In this study, ad exposures were concentrated in two ses-sions witb short delays. The boundary conditions of the facil-itation and interference effecis of brand names should beexplored hy employing less compressed exposure and mea-surement sequences. Generalizability also should be assessedby examining the effects of suggestive brand names on othertypes of products (e.g., consumer nondurables) and adver-lisements (e.g., less infbrmation-ladcn television advertise-ments) in whicb the level of involvement may not be as highas was the case in this experiment. Because tbe effects that

56 / Journal of Marketing, January 1998

were observed in this siiidy. ihough significant, were nottarge (e.g., the omega-.squared eflects size estimates of .sig-nificant contrasts generally were only approximately .01).laciliiaiion and interference from brand name suggestivenesstiiiglu nol be present under some circumstances.

Finally, two broader research areas are suggested by thisresearch. First, it also might be instructive to explore otherpossible explanations and outcomes of positioning and repo-sitioning—for example, in terms of principles from socialcognition research, such as schema congruity theory andhow brand names or otlier marketing communications affectconsumer expectations and the brand evaluations that areformed. Understanding the persuasive impact of reposition-ing is an itnportatU research priority, l-or example, underwhat circumstances are repositioning efforts seen as lackingcrcdibiliiy? Mow flexible can brands be in the minds of con-sumers? What i.s the iipiimal balance between continuity andchange for a hrand?

Second, a better understanding of the choice criteria forbrand names is an important managerial priority. As notedpreviously, brand names can be chosen to enhance brandawareness and the formation of favorable, strong, andunique brand as.sociations. This research shows not onlyhow a suggestive brand name can facilitate the creation ofinitial brand associations, but also how a nonsuggestivebrand name can accommodate additional benefit claimsmore effectively. Relatedly, Meyers-Levy (1989) showshow a suggestive brand name that evokes many associationssometimes leads to lower brand name recall, because theseassociations cue competing concepts and prtxiuce interter-ence. Thus, under some circumstances, choosing a tnean-ingful name can facililale consumer inference tnaking andbrand evaluations but potentially at the expense of adver-tisement and brand recall. Oiher trade-ofts in brand namechoice criteria should be considered, particularly in terms ofconsutner metnory versus persuasion.

REFERENCESAaker. David A. (1991). Managing Brand Equity. New York: The

Free Press,(1996). Buildhi}; Strong Brands. New Y(irk: The Free

Press.AIha, Joseph W. and Amitava Chattopadhyay (1985), "The Effect

of PartlJst Cueing on Benctll Recall: Prohlcm Framing at thePoint of Retrieval," in Advances in Consumer Research. Vol.12. FJi/aheth C. Hirschman and Morris B. Hoibrook. eds.Provo. UT: Association tor Consumer Research, 410- i 3.

Burke, Raymond R. and Thomas K. Snill (!988). "CompetitiveInlerfcrence and Consumer Memory for Advertising," Journalof Cttnswm'r Research. 15 (June), 55-68.

Chililors. Terry L, and Michael J. Houston (1484). '"Conilitions fora Picture Superiorily Effect on Cotisuiner Memory." Journal ofCon.Mimt'r Re.search. II (September). 551-63,

Ellis. Henry C. and R, Reed Hunt (1983). Fundamentals ofHunutnMemory and Cognition. 3d ed, Dubuque. IA: William C.Brown.

Keik-r. Kevin Lane (1987). "Memory Factors in Advertising: TheEtiecl of Advertising Retrieval Cues on Brand Evaluations,"Journat of Consumer Research. 14 (December), 3 16-33,

(IWla), "Cue Compatibility and Framing in Advertising,"Journal of Markeiing Research. 28 (February). 42-57.

(1991b), "Memory and Evaluation Kttects in CompetitiveAdvertisitig Environments." 7()HrH«/ of Consumer Reseuivh, 16(March). 463-77.

(1993), "Conceptualizing, Measuring, and Managing Cus-tomer-Ba,sed Brand Equity," Journat of Markfiiny. 57 (Janu-ary). 1-22,

(t998). Strategic Brand Management. Upper Saddle River.NJ: Prentice-Hatl.

Kent, Robert J. and Chris T, Allen (1994). "Competitive Interfer-ence Effects in Consumer Memory tor Advertising: The Role ofBrand Familiarity," Journat of Marketing, 58 (July), 97-105,

Keppel, Geoffrey (1982), Design and Analysis: A Researcher'sHandtiook. 2d ed. EnglewotKi Cliffs, NJ: Prentice Hall.

Lut/., Kathy A. and Richard J. Lutz (1977), "Effects of ItiteraetiveImagery on Learning: Application tt» Advertising." Journat ofApptied Pnrtiotoiiy, 62 (4). 493-98.

McGeoch, John A, (1942), The Psychology of tiuman l^'arning.New York: Longmans, Green,

Melton. Arthur W. and J, M. Irwin (1940). "The Influence ofDegree o\' Interpolated Learning on Retroactive Inhibition andthe Overt Transfer of Specific Responses." American Journal ofRsyilwlogy. 5^ (April), 157-73-

Meyers-Levy. Joan (1989). "The Intluence of a Brand Name'sAssociation Set Siie and Word Frequency on Brand Memory."Journal of Consumer Research. 16 (SeplLMiiber), 197-207,

Murnane, Kevin and Richard M. Shiffrin (1991). "Interference andthe Representation of Events in Metnory," Journal of E.\peri-mentat Fsvchotogv: Learning. Memory, and Cognition. 17 (5).855-74.

Postman. Leo and Benton J, Underwood (197.3). "Critical Issues inInterference Theory," Memory and Cognition. I, 19—40.

Ray. Michael L, and Allan G, Sawyer (1971). "Repetition in MediaModels: A Laboratory Technique," Journat of MarketingResearch. 8 (February), 20-29.

Robertson, Kim R. (1987). "Strategically De.sirable Brand NameCharacteristics," Journal of Consumer Markeiing, 6 (Fall),61-71,

Tulving, Liidel (1974). "Cue-Dependent Forgetting." AmericanScientist. 62. 74-82.

Brand Name Suggestiveness / 57