I

After three successful years in thr: Irersonli Llare division of Unilever in Pakistan, Laercio Cardoso was contemplating an atiractive leaderslrip position in China when he received a phone call from the heaci of Unile.uer's Home Care division in Brazil, Iris native country. Roberl Davidson was ioohing for sorneone to explore growth opportunities in the rnarketing of detergents to lorv-incoffre ccrnsumers living in the Nortlreast of Brazil. An alumnus of INSEAD's Advancer-{ Marrege!-nerrt Programme, Laercio had.ioined Urrilever in 1986 after graduating in business aiininistration from Fundagdo Getulio Vargas in 56o Paulo. He thus had the sjniority and rriarketing skills that were necessary for the project. More irnpoftantly, he had ncver been irivolved in the traditional approach to marketing detergents and, having witnessed the s'.iccess of Nirrnal in India. he rvas acutely aware of the threat posed by local brands targc:teci at low- i ncolle coltstl ltters.

iior this prtrject, named "Everyrnan". Laercio assenrbled an interdisciplinary tearn irrcluding Marccs Diniz from Sales, Antonio Conde fi"om Finarrce, and Airlon Sinigaglia fionr Manufacturing. The first phase of the pro-iect involved extensive field studies to understand the lifestyle, aspirations, shopping and laundry habits of low-income consumers. It was durin-e one of these trips that Laercio met Maria ConceigSo, pictured on the cover page in lrer horne in Foftaleza. where she- lived with her dar-rghter. Elizangela. lg (shown on the right with trvo of her four children). Like almost everyone in Brazil, Maria told Laercio that althoLrgh she would love to buy Orno. Unilever's flagship brand, her tight budget meant that she could only
afford cheapel local brands.
Back at LJnilever's headquarters in Sdo Paulo, Laercio prepared foran important meetit.tg witlr Davidson to decide whether the company shoLrld change the way it rnarketed its detergent brands to low-incotne col.lsuurers in the Northeast. Increasittg detergent usage by Maria and the other 48 million predominantly low-income consLnners itt Brazil's Northeast was crucial for Urrilever. given that the cornpany already had an 810% share of the detergent powder category. However, many in the company believed that a large multinational like Unilever shoLrld not fi-eht in the lower-errd of the market, wlrere even small, local entrepreneLrrs with a lower cost stntcture struggled to break even. How could one .jLrsti[, diverting money frorn Ouro to invest in a lower-rnargin segrnent?

Deciding to target low-income consumers iu the Norlheast wor.rld throw up sorre rrore difficult questions: Should Unilever change its current rnarketing and branding strategy? For example, could Unilever extend or reposition its existing cheaper brands, Minerva and Campeiro, or rvould a new brand be necessary? What would be tlre ideal positionin;; anci rnarketing rnix of a Unilever braud targeted at low-incorre consluners? Finding tlte attsr^"er:; would not be easy as few at Unilever (or other multinational firnis) had any kno'wlecige of low-iucome consutlers or first-hand experience of the kind of rnarketing strategy'that ivouiti uork for this segnrcrr, t t,

li,

Nirma. a lorv-price cletergent dei'elopccl b1'a small Indian entrepreneLrt'. cluickl', gained 48uh of thc lnclian detcrgent ntarkct. leaving Ur-rilcver in a clistant second placc rr;th a 249o markct sharc. Iror more inlbrmation on Nirrna. see "Flir.rdustan l,cvel l.inritcd: L,evers lbl Change"" tri Char!otte Ilutlel and Sumantra Ghoshal
o

./\a I oo I

,4,

,00?\

Consumer Information
Brazil: Ovenriew and
Regriena.i.

Differences

Brazil is by far the iar.gest countrw irr Latin America. lt covers 8.5 rnillion kni'? (almost as big as the US and 35 trines big.ger than the UK) stretching 4,345krn from Noftli to South and 4,33Okm frorn East to Wsi,sL. Its 170 million people live predominantly in two clusters on tlre Atlantic coast: one ccricenrrated in the Southeast, home to Brazil's two largest cities, 56o Paulo and R-io de Jarreiro, and the other in the Norlheast. whose main cities are Salvador.
Reciie

ani Fortaieza.

Dr,rring the ia.st three decades Brazil has experienced cycles of deep recessiorr and strottg rccovery. GDP grew by 8.1% per year during the "economic uriracle" of the 1970s, but crniy'by 2.6%o per year during the 1980s, the so-called "lost decade" clraracterized by stagnation and hyperinflation. In 1994. the Plano Reul initiated by the Finance Mirrister (and later President) Fernanclo Henrique Cardoso introduced a new currency (the Reais, R$) and succeeded in corrtrolling inflation, wlrich lecl to a strong economic recovery in 1995-1996. The boom was particlrlarly beneficial to lower-incorne cousLlr.ners and the purchasing power of tlre poorest 10% of the populatiott grew by 27% per year during this period.
ecorromie-:

ln l996,Brazil'spercapitaincomewas$4,420,o11aparwithcor-rntrieslikeHungary($4,370)
and Malaysia ($4,310), and well above other developing cor"rntries like Indonesia ($1,050) and lndia($380). As shown in Exhibit l, however, this average hid large regional differences. Per

capita income was around $6,600 in the Southeast (comparable to Uruguay or Saudi Arabia) and only around $2,250 in the Norlheast (comparable to Peru or Jarnaica). More generally, the 48 million people living in the Northeast lagged their SoLrtheastern colrnterpafts on -iust abor-rt every development indicator. For example.40% of the population in the Nortlieast (NE) are illiterate. a level comparable to India (52%). whereas only l5oA are illiterate in tlte Southeast (SE). As showr.r in Exhibit 2" 53o of the population in the Northeast lives on less than two rrinirnum wages (social classes E+ and E-) vs. 2l'h in the Southeast. During the 1990s, federal and local governlnents started providing tax incentives to companies irrvesting in the NE region, yet the econorry in the NE, was predominantly rural and remained heavily dependent on agriculture.

The Noftheastern states of Brazil also have a distinct culture and Iristory. It rvas the first region of Brazil to be coloriizedby Europeans. who brought large rrumbers of West Af icans to work as slaves ou sugar cane and cocoa plantations as early as the sixteenth centurl'. ln 1996:,650/0 of the population irr theNE was of rnixed African and Er.rropean origins (vs. i0% in the SE). Lifestyle. culture and religion all share African influences. Music and huincur arc key elements of their culture and many of Brazil's best-known artists come fi'crn the region. Popular parties like Carniva[, "Forr6 Festivals" altd "Maracatu" bring millions of people oirto the streets and are nrajor events irr the region. Iu contrast, tlre Southeast was developc<.| later, -flte mainly by ELrropeans who rnigrated in the 1880s to work on the ccfiee plantations. econornic and political power of modern Brazil is firmly rooted in ihe Southeast region.

Clothes Washing in the Southeast and Northeast of Brazil

The way clothes are washed in the Northeast and Soutllc;ast of Bra;:il is very different. In Recife QrlE), only 28Yo of households own a washiitg i.naclrine aridl3o/o of women think that
z 1,,

bleach is necessary to remove fet starns. Il Sio PaLrlo (SE), 67%o of families own a washing nrachine and only 18oh of won-reil tlrink tlrat bleach is necessary to remove fat stains. In general, womeu in the l.ioi'ilreast scir"ii; clothes r"rsing bars of laLrndry soap, a process which requires interrse and srrstzrined effort. Ihey then add bleach to remove tough stains and only

add a little detergeltt p'c',rrder ill the end, prirnarily to make the clothes srnell good. Irr tlre Southeast, tlre process is sirniiar to European or Norlh American habits: wolnen mix powder detergent ancr so\fltener in lr u,ashing machine and use laundry soap and bleach only to remove
tlre toughe:sr s:ains.

As a resu'it of fhc-se differeuces, the penetration of detergent powder and laundry soap is aimost the same in tlre NE and tlre SE, but Northeastemers use a lot Inore soap and less powJsl than Soutlreasterners (see Exhibit 3). Another difference is that clothes are washed irore frequently in the NE than the SE (5 tirnes a week in Recife versus 3.9 in S5o PaLrlo). 'fhis rs because low-ilrcome consumers own fewer clothes and have more free tir-ne (because
fewer wornen work outside the home) than higher-ir.rcome consumers. Interestingly, many worren in the NE view washiug clothes as one of the more pleasurable activities of their rveek. This is because they ofterr do their washing in a public laundry. river or porrd where they meet and chat with their friends. Irr the SE, irr corrtrast, rnost women wash clothes at home alone. They peiceive doirrg laundry as a chore and are prirnarily interested in ways to
nrake the task easier.

People in the NE and SE differ in the syrnbolic value they attach to cleanlirress. Many poor Noftheasterners are proud of the fact that they keep themselves and their farnilies spotlessly clean despite their low income. Because it is so labour intensive, many worren see the cleanliness of clothes as an indication of the dedication of the mother to her family. Personal and home cleanliness is a rnain subject of gossip. In tlre Southeast, where most wonren own a wasl-ring machirre, it is rnuch less important for self-esteem and social status.

How do Northeastern Consumers Evaluate Detergents?

Along with price, the prirnarily low-incorne consulrers of the Noftheast evaluate detergents on six key attributes (Exhibit 5 provides importance ratings, the range of consumer expectations. and the perceived positioning of key detergent brands on each attribute). The rnost impoftant attribute is the perceived power of the detergent (its ability to clean and rvhiten clothes with a small quantity of prodLrct), which is often.iLrdged by tlie quantity ot foam it prodr-rces. Second is the srnell of the detergent: consllnrers often associate a strong, pleasant srnell with softening power and gentleness to fabric and hands. Third is the abi!ity rc rer.nove stains without the need for laundry soap and bleach. Next is the ease,uvith llhicir ihe powder dissolves in water and the absence of residue on the firbric after rinsing, two elernenls that are evaluated by the consistency and granularity of the powder. Packagiilg cLrrnes ncxt: low-income consumers (who are often barely literate) prefer distinctive, sirriple and e:is;-iorecognize packages that are also easy to open and protect against humidity. Impact ori colours (fading) is the least irnpoftant attribLrte fbr tlrese consumers.

lr

_]

The Brazilian Fabric Wash Ma:vltet

Key Industry Players in Brazi!.
Unilever cornpalry. headquarlered in London (UK) ancl Rotterdanr (Netherlands). It has abc;,rt 300,000 employees itr urore than 150 countries. In 1996 it had a porlfolio of 1,600 biands worldwide, including 45 key detergent brands (see Exhibit 6). Unilcver is a nioneer of the consunler goods industry in Brazil. Lever Brothers stafted opera.tiorrs ir Y,\razil in 1929 and opened their first plant in 56o Paulo in 1 93 0 to tnanufacttlre 3uniight soap. Omo, Unilever's most successful brand, was launched in 1957 and was the first ietergurt powder in the country. Unilever acquired Cia Gessy Industrial and its rich portfolio of pci's..xal care brands in the 1960s and started its food operations irr the 1970s rvitlr the launch of Doriana. the first rnargarine in Brazil. ln 1996 it operated witlr three divisions: Lever for home care, Elida Gibbs for personal care. and Vatt den.Bergh for foods. Yet detergents remain the cash cow of Unilever Brazil, providing tuel for growth irr the food and personal care categories. In 1996. Unilever was a clear leader in the detergent porvder category irr Brazil, with an 810/o market share achieved r,vith three brarrds: Orno (one of Brazil's favorite brands across all categories). Minerva (the only brand to be sold as both

Unilever is

o US$56 biiiion

cletergent powder and laLrndry soap). and Campeiro (LJnilever's clteapest brand).2

Procter & Gamble

Procter & Garnble is a US$40 billion company, headquaftered in Cirrcinnati (USA). with 98,000 employees and operations in 80 countries. P&G started operations in Brazil only irr 1988. In 1996 they acquired the detergents business of Bornbril, a Brazilian cornpany, and its three brands: Quanto, Odd Fases and Pop. After spending a large atnount on tnanttfacttrrirrg irnproverrrents, P&G rnigrated Quanto towards Ace and Odd Fases towards Bold. two of its global brands, but kept the low-price brand Pop. P&G is a distarrt second player rvith only a 15% share of tlie Brazilian detergent market. However. the realthreat is largerthan its current rnarket share suggests because P&G Brazil can draw on tlre fornridable R&D and marketing expeftise of tlre cornpany worldu,ide.

Market Structure
The Brazilian fabric wash market cousists of two categories: detergent powder and laur-rdty soap (sales ol liqtrid laLrrrdry deterger)ts are rregligiblel.

Detergent Powder

ln I996, detergent powderwas a $106 million (42,000 tons) market in thr;l\crnheast, gt'cwine at the remarkable annual rate of 17% thanks to the econonric upturn of the ['lanr-t iiectl.The barriers to entry in this market are high because the manuf'acturing pfocess rs ca.pitai itttensive. Detergerrt powder is ntade by mixing sulfonic acid, sodiurn sulphate anij l<,-,lp. Premiuln products. like Unilever's three detereents, also contain specific ol'rZ)'lnaS anti builders which 2
l1nilever also sells Urilhante" a brand o1'launclry soap anrl Ctergertt po'r'dct'. ilorvever. it had alllnost zero market share in the NE in 1996 and is therefble not mentioireri ztl'ri' flrthtr rri this oase stucly (r 1,, tb

'1 i'
t.i

itnprove the whitening powerof tlre cletergerrr eylls'.t il is used irr a waslring rnachille. The rnix is then heated up to 400"C to form a licruid pulp which is then transfonned into powder when hot air is blow through it in a dry ton'cr. The drying process corrsllmes a great qr-rantitv of steam whicl-r is prod,.icecl by a local utility plant. Perfurne and otlrer heat-sensitive substarrces are added at tlre ertci of the process. Detergent designed for harrd washing is cheaper tcr prodLrce but perforins very pcorlv when used in a rvashing uraclrine.

At

75yo, Utrilever's slrare of the NE detergertt tnarket is below its national average (see Exhibit 7). Onro, its doininant brand, has a 52Vo share and is sold to retailers at $3 per kg. Minerva l'ras a 1'/u,/o share and its retail price is 82%o that of Omo. Campeiro has 6%o of the trrarket and is soid at 57o/o of Otno's price. In the NE, P&G's market share is slightly above its r-raticxala\,/erage (17.5%). Ace isthethird highest-selling brand with an llo% rnarketshare.

Lattndty Soap

In 1996, the NE market for laundry soap bars was as large as tlre detergent powder market ($102 nrillion for 81,250 tons), but growing at a slower raLe (60/o). The barriers to entrv were lower in the laundry soap market tlran in tlie detergent powder market because soap is relatively easy to produce from fats and oil. ln fact, the animal fat that is a primary component of soap is produced in large quantities by slaughterlrouses and meat processing plants. One of the lirnitations of laundry soaps is that animal fat tends to leave the clothes yellou,. They are also difficr-rlt to perfume because tlre base has a very strong srnell.3 l-aundry soap was sold at a tnuch lower price than lar"rndry detergent powders (average revenues of $1,250 per ton vs. $2,520 per ton for detergent powder).
Laundry soap is a multi-use product which has Inany home and persorral care Llses. People with washing machines prirnarily use it to remove tough stains (e.g., orr shirl collars): for tlrose without, laundry soap is used to wash allclothes. The popLrlarity of laundry soaps irr the NE is also due to the softness of the water in this region (i.e., its low calcium content), which helps the soap to dissolve and produces great quantities of fbarn, thus reducing one of the key advantages of powders. In comparison, rnost water in Europe. US and Ilrdia is hard Tlre NE rnarket for laundry soap was very fragmented. As shown in Exhibit 7, the top for-rr players lrave only 38Yo of the market. Unilever's Minerva brand is the leader with a lgok nrarket share, selling to retailers at $1.7 per kg (a41% discoLrrrt relative to Omo). P&G did not rnanufacture laundry soap. Hence Unilever's main competitors were local Braziiian companies. The biggest competitor was ASA. Its brand, Bem-te-vi. had I l% of the market and sold at $1.2 per kg. The other players were even srraller local cbmpanies with ito inore tlrarr loh of tlte rnarket each (except for Flora Fabril. r.vlrich had 60/o of the laundn, soai;
rnarket).

Brand Positioning Exhibit 8 provides information on brand awareltess, brand knowledqe arrci brancl penetratiorr of the major detergent powder brands in the NE in 1996. Exhibit 9 shorvs tite perceptiorr of ,. /^ these brands on two dimensions: perceived quality and perceived price. i-lxiribit l0 provicles )/y

soap uses the sante base as laundtl' soap but fhe rau tr,',r'eri{ is suLmi'.teci to a long and costl;'process of'filtcring. u,hich rerroves thc base snte ll and leaves i1 nr:t,tral

-foilet

,(r

J
key information on all detergent porvder and laundry soap brands (packaging, positioning, key historical facts, and financial and markei data).

Decision-makilrg Time
The results of'the Everyrnan project irrcreasecl Laercio's conviction that Unilever should also Siill, he was facirrg strong internal resistance frorn people like target low-income "onj,rnr"..r. Fernanijr t\4achado, ihe category mauager for detergents. A typical argument between Laercio and Fernaitda wculd run like this:
"Loerci.o, I think that we should stay au,ay Jrom lhe low-income segntent. These people just h.ave io money and I really don'l see vtfuy vtg should divert money i'om our premiunr brands to invest it in a low price bruncl! In the short term this would simply cannibalise our high-margin sales u'ith lower-margin ones. In the Ionger term this would certainly increase price contpetition in the category. Hov, will I be able to sustain Omo's price prentium if'people can buy nlmost the same product at ha('the price'? "

"Fernanda, I understand your concerns but v,e need to do somethingfor the low,income segment. We already have Blok of the nnrkel qnd I really see no other way to grow. Besides, if we don't clo anything, P&Gv,ill attackus in this segtnent where we are most vulnerahle. .Ju,st look al what happened lo us in India. " "But Laercio, caramba! Brozil is not India! Detergent penetrcttion
i,s

95% here vs.

55% in India, our products are cf much higher cluttlih,, and v,e hc;e been marketingpremiunt brands in Brazil ,since 1929. Think cthout the kind of message that the global investment contntunilry will hear: "Unilever has lost its marketing skills and is abandoning its premiunt brands." Renternber \vfnrlboro Friday?a Hovt do you think the stock nrurket will respond? l4that ubottt our corporate reputation? Hov, are we going to be able Io ctltract ancl retain the next generation oJ brand managers who only wcntt lo u,ork ctn Jtrentiurn brund|l "
"Que isso, Fernanda! You should spend more time getting to know your fellou, Brazilians and less time behindyour computer! If we get the right strutegy, lowincome consumers will be rendy to pay for otrr brand and Omo buyers won't ntove. Also, think about the experti.se that v,e would guitt, which v,e c,or.tld apply to our other categories. If we become a leoder in morketing to low-incotiie consumers I bet that financiul analysls will praise us and thol lop students it,iii line up to interview wilh us. "

GolNo Go Decision
Robert Davidson had lieard these arguments over and over, yet he .,vas still undet:ided. He was . , particularly concerned with the profitability of this consumer se.grnent. Certainiy, part of the b/ft

t4

On 2lApril 1993, Philip Morris USr\ cut tlic price of Marlboro by 20Y,. and in the proccss knocked almost ivlanl anal5'sis i;rtelplet:d t'iiiiip Morris' decision as a sign $10 tril'lion OtTthe market value of the company. that fig hrands were losing the battle against cheaper pr\.,t-.e \abels ani unbianded proclucts

*il

'0

new sales would come at the e;xpense of Unilever's existing blrnds. At wlrat cannibalization rate (percentage of new salcs comiirg fiorn other Unilever brands) would Unilever start losing money? More geuerally. ire wondereci n'ulrether Unilever had the right skills and organization to compete in this msrket: ln the lcrng run, what exactly wor-rld lJnilever gain and wlrat would it risk if things went wrong?
B

rand and'Slark6ting $trategy

Value propositiot't

Vias there s.,:mething wrong with the existing positioning

bc?

of Unilever's

!--rands? Wouid it be really necessary to develop a new value proposition?

three detergent [f so, what should it

Brant! Strategy Could Unilever deliver the desired value proposition with one of its three existing brands, or witl-r a brand extension? Would Unilever really have to develop a new brand from scratch? Could it use a brand from its large international porlfolio? This was a thorny issue, especially considering the rumor coming frorn headquarters that Unilever was abottt to embark on a large-scale effort to reduce its brand portlolio.5

Marketing Mix
Product
Unilever could produce a product comparable to Campeiro. its cheapest product, bLrt would it deliver the berrefits that low-incorlle col'rsulners wanted? Alternatively, Unilever could use Minerva's formulation, but it rnight be too expensive for low-income consumers. Unilever's scientists could develop a tliird fonnula priced half-way betrveen Minerva and Carnpeiro if they could eliminate some ingredients. The question was to determitre which attributes could be eliminated, which should be retained, and wltich, if anv, would actually need to be improved relative to both existing brands.
Selecting the right packaging size and type was another diffjcLrlt task. I arger packages would reduce the cost per kilo but could price the product out of the weekly budget range of tire poorest consumers. Unilever could use a plastic sachet, wliich would cost 30% of the price of traditional cardboard boxes, but rnarket research data showed that low-income consume!'s were attached to boxes and regarded anything else as good fbr only second-rate proCucts. One solution miglrt be to launch rnultiple types and sizes.

Price
Choosing the wholesale price (the price paid by retailers) was t!'re sing;le mor;t irnpoftant decision for Unilever. Priced too high, the product would be cut of reaclt for the target segment. Priced too low, it would increase the inevitable cannibali,ztttion ot e"".isting Unilever
For the purpose o1'the break-even analvsis. assunrc that devclcping a nc\,r,' brerr-ril wouicl add $0.10 per kg in incremental marketing costs. that launching a brartd er:tensiott wcru!(l iidd $0.05 per kg and that repositioning an existing brand would nttt lead to an)' increinental marlidtints costs

th tL\

brands. Should Unilever use couponS or other ineans to reduce the cost of the product for lowir.rcorne consumers? Should it change thi: price of Omo, Minerva and Carnpeiro?

Promotion
What would be the ob-ieotive-of the communication? What should be the key message? Lowincome consurne.rs might be. ie_luctant to buy a product advertised "for low-income people",

especially as products viith' that kind of message were typically of inferior quality. On the otlrer hand, usrng ruhe, ciassic aspirational conrmunication of most Brazilian brands could confuse consumers and lead to unwanted cannibalization. Wlrat aboLrt packaging and point-ofprrrchnse riisplair:r'l Should they use the same slogan as the television commercial? Finally, u'hat shoultl tinilever tell the owners of the small stores where most low-income consumers shopped? Getting buy-in from small store owners would l.e crucial because low-irrconre conslrrrers relied on them for advice and for financing (wlrich is rvidely used in Brazil, everr

for inexpensive consumer goods).

In regular detergent markets Unilever had established that the most effective allocatiorr of communication expenditure was 70o/o above-the-line (rnedia advertising) and 30o/o below-tlreline (trade promotions, events, point-of-purchase marketing). The advarrtages of using primarily media advertising were its low cost-per-contact and high reach because almost all Brazilians, irrespective of income, are avid television watclrers. One alternative would be to use 700h below-the-line communication. At $0.05 per kg. this plan r.vould require only one third of the cost of a traditional Unilever communication plan. On the other hand, it would
lower the reach and increase the cost-per-contact.

Distribution
Unilever did not have the ability to distribute to the approximately 75,000 small outlets spread over tlre Northeast (see photograph, Exhibit 12). Yet getting access to these stores was key because low-income consunlers' rarely shopped in large superrnarkets like Wal-Maft or

Carrefour. For distribution, Unilever could rely on its existing network of generalist wholesalers, which supplied Unilever's existing detergents and a wide variety of products and had rrational coverage, but whicli sometimes had to rely on secondary, smaller local wholesalers to reach all stores, which increased their cost. Alterrratively, it could contract witll dozens of specialized distributors who would get exclusive rights to sell all [Jnilever detergents in certain areas (see Exhibit l3 for a comparison of the tu,o distribr-rtion channels). Choosing the right distribution channel was impoftant because it was a large component of the product cost, would be hard to reverse, and ultirnately would have strong implicaiir-ins icr the ability to push sales and build brands at points of sale. t

'-rt