Professional Documents
Culture Documents
BCG No Shortcuts Sep 2010 Tcm80-60004
BCG No Shortcuts Sep 2010 Tcm80-60004
Jens Harsaae
Ross Link
Neal Rich
Kevin Richardson
Rohan Sajdeh
September 2010
bcg.com
© The Boston Consulting Group, Inc. 2010. All rights reserved.
Appendix 19
N S
Executive Summary
C
ompanies invest staggering sums of money in ◊ In the face of such complexity, managers turn to rules
marketing with surprisingly little rigor. Because of thumb, such as spending as a percentage of revenues,
it is notoriously difficult to measure and opti- to guide their budget allocations for marketing.
mize the return on marketing investments
(ROMI), marketing executives oen rely on rules To test some of the common rules of thumb used to
of thumb—such as spending as a percentage of revenues— set spending on marketing, The Boston Consulting
to guide their decision-making. Our research and experience Group and Marketing Analytics, a leader in market-
suggest that these shortcuts can be imprecise, unreliable, response modeling, reviewed a data set of compara-
and just plain wrong. We recommend another approach— ble marketing-mix models for 75 consumer brands.
one that goes beyond marketing to encompass all commer-
cial investments. It integrates a top-down strategic perspec- ◊ In our joint research and analysis, we found a wide
tive and a rigorous bottom-up analysis. variance in marketing impact, efficiency, and return on
investment—even across a relatively similar collection
More than $1 trillion a year is spent on marketing. of brands.
◊ The figure is even higher if you include related com- ◊ We also found that five of the most commonly used
mercial investments, such as trade spending, price pro- rules of thumb for marketing investments had no clear,
motion, and sales force incentives. consistent relationship to marketing performance in
the sample we reviewed.
◊ Many companies spend as much—or more—on ad-
vertising alone as they invest in capital expenditures. ◊ Managers who rely on such rules, therefore, could be
using the wrong approaches to allocate their market-
As marketing budgets continue to escalate, managers ing budgets across business portfolios and the market-
are coming under increasing pressure to prove the ing mix.
value of their marketing investments over time.
Rather than look for shortcuts, managers should em-
◊ At the same time, marketing is becoming much more brace the complexity of optimizing returns on mar-
complex as digital and viral marketing vehicles prolif- keting. We recommend that they integrate a top-
erate and market segments fragment. down strategic perspective on commercial
investments with a rigorous bottom-up analysis.
◊ Another complicating factor is that marketing perfor-
mance depends heavily on elements that either are ◊ Such a comprehensive approach would incorporate all
difficult to measure precisely—such as the content of the tactics that managers generally consider when
an ad and its creative execution—or emerge over they are optimizing ROMI, as well as those that fall un-
many months and years, such as the impact of market- der the broader banner of return on commercial in-
ing on brand perception. vestments (ROCI).
N S
The Need for Rigor
in Spending
M
ore than $1 trillion is spent annually on investments one-tenth the size in other areas of their
marketing around the world, and that business.
sum doesn’t include related forms of
commercial investment, such as trade Many executives tell us that they are feeling increasing
spending, price promotion, and sales pressure to meet demands for accountability through
force incentives. Many companies spend as much—or better marketing metrics. Yet marketing performance
more—on advertising alone as they invest in capital ex- depends heavily on elements that are difficult to
penditures. (See Exhibit 1.) This situation led Jim Sten- measure precisely, such as the content of an ad and its
gel, former global marketing officer for Procter & Gam- creative execution. Furthermore, a change in one busi-
ble, to observe that companies spend millions on media ness activity can directly—and indirectly—affect the
today with less data and discipline than they bring to efficacy of many others. As a result, it can be extremely
20
10
0
10 20 30
Capital expenditures/net revenues
(median percentage, 2004–08)2
= $15 billion in median net revenues, 2004–08
Source: BCG ValueScience Center, BCG analysis.
Note: We analyzed 169 companies from the S&P 500.
1
We divided each company’s median annual advertising expenditures from 2004 through 2008 by its median annual net revenues.
2
We divided each company’s median annual capital expenditures from 2004 through 2008 by its median annual net revenues.
N S
Testing the Rules
T
o validate some of the rules of thumb most during the same time period. (See Exhibit 2.) Across the
commonly used to justify marketing bud- brands in our sample, the incremental impact of mar-
gets, BCG and Marketing Analytics part- keting on unit sales volume ranged from 1 percent to
nered on a meta-analysis of 75 comparable more than 50 percent. In examining marketing efficien-
marketing-mix models for consumer pack- cy, we found that companies spent anywhere from less
aged goods in the United States. We used the models to than $100,000 to almost $18 million to capture an incre-
analyze each brand’s ability to drive incremental sales mental percentage point of unit sales volume for a
volume from marketing activities, including trade spend- brand. Thus, for every dollar spent on marketing, the
ing, and from advertising via television, print, radio, bill- companies in our data set were able to generate any-
boards, and online media. (See the sidebar “Building where from $0.04 to $2.75 worth of incremental sales.
Our Proprietary Data Set” and the Appendix for details And in terms of ROMI, brands realized anywhere from
about the study’s data set and analytical method- $0.03 to more than $2 of incremental gross margin for
ology.) every dollar their companies spent on marketing.
The first key finding from our research was a surprising- Such a wide variation in the outcomes of similar market-
ly wide variance in marketing impact, marketing ing-mix models suggests that there is no benchmark for
efficiency, and return on marketing investment (ROMI) marketing returns that managers can reliably apply
for brands with reasonably similar profiles—that is, for across brands when planning budgets. Historical bench-
consumer packaged goods sold in the same channels marks may be calculated for a specific brand, but they
The meta-analysis of marketing-mix models developed by sessed the amount of marketing spending it took to in-
The Boston Consulting Group and Marketing Analytics re- crease unit sales volume or retail sales by 1 percentage
sulted in a unique proprietary data set with associated point. Finally, we measured the incremental revenue and
benchmarks. The findings are powerful, but companies incremental gross margin generated by each dollar spent
must carefully consider the implications for their own sit- on marketing.
uations.
It is important to remember that the models in our data
We undertook this analysis because we were interested in set are focused on the shorter-term sales impact of com-
evaluating three measures of marketing performance: mercial activities. Consequently, they provide very little
marketing impact, marketing efficiency, and return on insight into the longer-term health of a brand or the im-
marketing investment. For impact, we sought to under- pact that longer-term commitments to branding have on
stand the incremental percentage of unit sales volume shorter-term sales.
that resulted from marketing. Our efficiency measure as-
The second key finding from our research was that many Maintain “Share of Voice” and
of the rules of thumb frequently used as shortcuts for Overinvest to Gain Market Share
allocating marketing budgets had little or no correlation
with enhanced performance in marketing. We eval- In the hopes of securing what they consider a fair share of
uated the following five commonly used rules for allocat- voice in the ongoing “conversations” with consumers, mar-
ing marketing investments. keters oen look to the competition when determining
their level of advertising exposure. However, we found
that brands with a relatively higher share of voice, which
Peg Marketing Budgets to Revenues we calculated as an estimated percentage of overall media
spending in the category, did not consistently drive greater
We found no consistent correlation between marketing unit-sales volume—and oen generated less gross margin
spending as a percentage of dollar sales and either mar- for every dollar they spent on marketing. Similarly, com-
keting impact or ROMI. Brands that spent more on mar- panies that overinvested to capture a share of voice that
keting as a percentage of dollar sales (that is, 30 to 35 per- exceeded their market share also failed to achieve greater
cent rather than 20 to 25 percent) did not drive more return on their investment. (See Exhibit 3.) In short, sim-
volume, but they did realize lower returns on their invest- ply shouting louder than the competition does not appear
ment. We therefore conclude that spending as a percent- to be the best path to achieving marketing success.
Exhibit 2. Marketing Impact, Efficiency, and ROMI Varied Widely Across the Modeled
Brands
Return on marketing
Marketing impact Marketing efficiency investment (ROMI)
80 2.00
15
60 1.50
10
40 1.00
5
20 0.50
0 0 0.00
Brands 1–75 Brands 1–75 Brands 1–75
N S
Exhibit 3. Capturing a Share of Voice in Excess of Market Share Failed to Consistently
Deliver a High Return on Investment
2.00
1.00
0.00
–100 –50 0 50 100
Share of voice – unit market share (%)
Integrate Marketing Campaigns to consider investing in lower-share brands, since they offer
Realize Synergies Across Vehicles greater marginal opportunity.
W
the brands that had relatively higher market share—in e are not saying that rules of thumb don’t ex-
fact, the return on investment was actually lower for ist. No doubt, in certain cases, some could de-
these brands. Therefore, driving incremental unit-sales liver higher levels of marketing performance.
volume was expensive and inefficient for the highest- However, as our study reveals, such rules shouldn’t be fol-
share brands in our sample. Managers should therefore lowed blindly without analysis of the specific situation.
30
2.00
1.74
20
1.00 0.82
0.68 0.63 0.61 10
0.56 0.51
0.43
0.00 0
Average 1 2 3 4 5 6 7 Number of vehicles used
of all
vehicles
Number of
brands 75 3 12 11 29 11 6 3
in sample
Source: 2010 ROMI modeling meta-analysis by BCG and Marketing Analytics.
Note: The models for the 75 brands were based on at least two years of marketing activity and weekly sales in the U.S. food channel between October
2005 and July 2009.
N S
A Better Road Map
M
ost companies have tried to improve Strategy: Spend on the Right Things
discrete elements of their marketing
plans, but few have addressed all the At the strategic level, the key ROCI disciplines are precise
activities in their commercial mix. Few- allocation of investments across the portfolio, clear com-
er still combine the rigor of models and mercial objectives, and focused customer targeting.
research with the art of strategy and creativity for a bal-
anced approach to marketing. Many also fail to suffi- Allocate spending on the basis of both strategic po-
ciently balance traditional marketing-mix measures of tential and commercial response. Commercial budgets
near-term sales impact with longer-term measures of should be directed to businesses that deliver high ROCI
brand strength and customer retention. Finally, few and offer long-term strategic potential. Our research in
managers are able to adjust and optimize their commer- consumer packaged goods suggests that U.S. manufactur-
cial activities continuously. ers should adopt a strategy of “go big or go home.” The
largest absolute budgets allocated to the largest brands
The approach we recommend for achieving higher ROMI (determined by unit sales volume rather than market
employs a proven and comprehensive set of tools. It in- share) in the largest categories yielded a disproportion-
corporates all the tactics that are generally considered ately higher return on investment. (See Exhibit 6.) By
when optimizing ROMI, as well as those that fall under a contrast, the smallest brands and the smallest categories
broader banner that we call return on commercial invest- had lower returns regardless of their budget range. Not
ment, or ROCI. In addition to the traditional ROMI levers only is direct investment being wasted on these “long-
of media spending and consumer promotions, ROCI cap- tail” opportunities, but managerial time and overhead
tures a wide range of other business drivers, including are also being squandered. Commercial budgets can’t be
trade marketing, product quality, pricing, distribution, allocated democratically—not every brand deserves an
and sales force activities, as well as the external context equal share of the budget.
of the overall brand category and economy. Through this
approach, we evaluate commercial investments not only Set clear, discrete, and measurable commercial objec-
for their ability to drive short-term sales, but also for their tives. Commercial objectives and success are too oen
potential to build brand equity over the longer term and defined aer the fact by choosing from a grab bag of the
to expand the purchase funnel—that is, the mental deci- most favorable metrics at hand. Even when commercial
sions a customer makes from gaining initial awareness of objectives are determined ahead of time, they are not al-
a product to purchasing it. ways aligned with corporate objectives or across the com-
mercial organization. We’ve found that a return to classic
The ultimate goal of this approach—which we outline be- purchase-funnel analysis can help companies greatly im-
low and in Exhibit 5 at the strategic, tactical, and opera- prove ROCI by encouraging them to select a few specific
tional levels—is to establish a lasting measurement and objectives (for example, brand awareness, trial purchase,
optimization capability within the commercial organi- or repeat purchase) and support them with a strong busi-
zation. ness case and associated metrics.
Strategic Precise allocation of the portfolio ◊ Allocate resources to brands with high
potential and high commercial response
Clear commercial objectives ◊ Align on a few specific objectives
Tactical Level of spending ◊ Find your sweet spot: above the minimum
and below the maximum
Commercial mix ◊ Optimize on the basis of driver efficiency
Target the subset of customers who really matter. spans. Others are too large given the marginal impact
Managers need to first determine which customers drive they deliver. Our research clearly showed that increased
value for a category and a brand, and then understand absolute spending drove incremental volume—but with
their behaviors, demographics, and attitudes. For exam- declining efficiency. The challenge of finding the limits of
ple, one of our clients discovered that a large, highly valu- minimum and maximum investment is all the more rea-
able group of buyers tended to make purchase decisions son to adopt a statistical, model-based approach, rather
as early as six months before a product’s launch. There- than relying on rules of thumb.
fore, virtually all the commercial activity timed to coincide
with the launch itself was wasted on a critical segment of Reallocate the commercial mix toward the highest-
“early deciders.” This discovery led the company to invest return vehicles. Companies intuitively know that there
a portion of its commercial budget much earlier in the will be varying levels of impact, efficiency, and ROCI for
product launch cycle. different commercial activities. But they do not always
quantify these important differences. In our research, we
identified clear patterns of inefficient allocation. Specifi-
Tactics: Spend in the Right Ways cally, most of the brands in our sample overinvested in
trade spending and television ads that generated high in-
Tactics address a company’s level of commercial spend- cremental volume but low efficiency and low return on
ing, allocation of the commercial mix, and timing of com- investment. This type of behavior oen reflects corporate
mercial activities. cultures and key performance indicators that reward vol-
ume over value.
Spend above the minimum threshold but not past the
point of diminishing returns. Many commercial invest- By contrast, less than half of the brands in our sample
ments are too small to break through the clutter of over- invested in online advertising during the period of our
crowded markets and capture consumers’ short attention study (and only in relatively small amounts, at that),
N S
Exhibit 6. Go Big or Go Home
Large Budgets Spent on Large Brands and Categories Yielded the Highest ROMI
Marketing expenditures versus brand size Marketing expenditures versus category size
0.33 0.85
≥100 0.89 0.89 ≥100 0.89 1.74
0.42 0.62
50–99 0.62 1.25 50–99 0.84
0.42 0.64
0.42
30–49 0.55 30–49
0.22 0.42 0.35
10–29 10–29
0.30 0.38 0.27 0.27 0.48 0.55 0.27
<10 <10
0.44 0.26 0.71 0.70
0 0
<100 100– 250– 500– >1,000 <0.5 0.5–<2 2–<5 5–<10 >10
249 499 999
Brand size ($millions) Category size ($billions)
0.62 = Incremental gross margin per $1 spent on marketing ($)
Source: 2010 ROMI modeling meta-analysis by BCG and Marketing Analytics.
Note: The models for the 75 brands were based on at least two years of marketing activity and weekly sales in the U.S. food channel between October
2005 and July 2009.
despite our finding that online ads delivered the greatest Managers should have a system for evaluating message
efficiency and return on investment. No doubt, findings complexity to ensure sufficient frequency of exposure,
like ours are prompting the recent expansion in digital and they also should consider seasonality and the pur-
marketing by consumer-packaged-goods manufacturers. chase cycle when trying to reach a specific audience. (See
Print advertising was also characterized by relatively Exhibit 8.)
smaller volumes and higher economic returns. Both print
and online media are highly targetable vehicles that can
be effective and efficient with discrete audiences. Over- Operations: Analyze, Execute, Measure,
all, it is clear that impact equals neither efficiency nor Adjust, and Repeat
return on investment—and that volume does not ensure
value. (See Exhibit 7.) The final and perhaps greatest barrier to making com-
mercial spending more accountable is developing the ca-
Balance message complexity, reach and frequency of pability to measure and optimize ROCI consistently and
marketing exposures, and product purchase cycles. continuously. This capability requires companies to em-
Consumers must be exposed to ad messages with suffi- bed new tools into their commercial-planning processes
cient frequency if investments in them are to pay off— and new thinking into the culture of their commercial or-
just how oen depends on message timing and complex- ganizations. (See Exhibit 9.)
ity. Complex messages require more exposures, but that
doesn’t mean they cannot be as effective as simpler ones. Although it is positioned as the last of our three sets of
Complex messages also require a minimum threshold of activities, operations must be more than an aerthought
initial investment in order to achieve efficiency and pay- when optimizing ROCI. Take the example of a telecom-
back (an S-shaped curve). By contrast, simple messages munications company that recently embedded a ROCI
have no minimum threshold: the first impression is the capability within its organization. The company divided
best, and it’s all downhill from there (a C-shaped curve). its initiative into two basic phases: analysis, which consist-
Return on marketing
Marketing impact Marketing efficiency investment (ROMI)
Spending per percentage point
Percentage of unit sales volume of unit sales volume attributed Incremental gross margin
attributed to marketing1 to marketing2 ($millions) per $1 spent on marketing ($)
20 19 5 3.00
4 3.8
15
2.00 1.85
3
10 2.4
2 1.7
1.00 0.84
5 0.9 0.62
4 1 0.39
1 0
0 0 0.00
Trade TV Print Online Trade TV Print Online Trade TV Print Online
Number of
brands in 75 64 43 29 75 64 42 29 75 64 43 29
sample
Source: 2010 ROMI modeling meta-analysis by BCG and Marketing Analytics.
Note: The models for the 75 brands were based on at least two years of marketing activity and weekly sales in the U.S. food channel between October
2005 and July 2009.
1
This percentage reflects the amount of sales volume that can be attributed specifically to the impact of marketing expenditures.
2
This calculation reflects the marketing expenditures that can be linked specifically to an incremental increase in unit sales volume.
Sources: Marketing Analytics experience and analysis; Jim Surmanek, Media Planning: A Practical Guide, NTC Business Books, 1996.
1
Media penetration is the percentage of people exposed to a given media vehicle within a given period; it is a key factor affecting an ad’s maximum
potential audience.
2
Media dispersion reflects the coverage of a given media daypart based on the total number of ad placements and the number of programs
(publications, for print ads) in the rotation. It is a key factor affecting the rate of audience accumulation. For example, 20 ads appearing in 15 to 20
programs constitute high media dispersion; 20 ads appearing in 4 to 8 programs constitute low media dispersion.
N S
Exhibit 9. Companies Must Embed ROCI Capabilities into the Commercial Organization
Communications planning
Demo- Awareness Budgeting
graphics Fulfillment
Consideration Commercial Advertising $
calendar
Preference t
Messaging Media
strategy buying Production
Trial
Behaviors Attitudes
Commercial plan
$
ed of developing and executing a ROCI model, and indus- effort on all fronts, the model was accurate and robust—
trialization, which called for the company to hard-wire the but also a bit untidy. To “industrialize” the analysis, this
ROCI approach into its day-to-day business activities. Al- disorderly plate of spaghetti needed to be transformed
though the analysis was tough, the industrialization effort into a replicable jigsaw puzzle. The pieces of the puzzle
took just as long—and was in many ways the more chal- became systematic, preformatted templates for each
lenging of the two phases. function to use in order to contribute to efficient updates
of the ROCI models. Then the company embedded the
In the first phase, the company developed a set of models insights from those models into the commercial organiza-
and analyses addressing the entire commercial budget. tion’s planning. It also created a full-time position to man-
Every one of the commercial functions—including tradi- age the ROCI process, and it assigned clear roles and re-
tional brand communications, retail marketing, distribu- sponsibilities across each function. The result is a more
tion, and new media—worked together on this effort, in cohesive and objective planning process across the entire
some cases for the first time. In the past, each function commercial organization.
had relied on completely different data, metrics, and tools
to make decisions about commercial investments. Since As this example illustrates, most successful companies
this phase was completed, they’ve gained a common cur- spend a good deal of time addressing organizational and
rency that allows them all to compare and optimize the cultural hurdles in addition to analytics and models. Spe-
impact and efficiency of—and the return on—their col- cifically, they focus on five objectives.
lective investments.
Organizational Alignment. A typical commercial or-
Whereas the analysis phase delivered immediate finan- ganization houses many disciplines, including advertis-
cial impact, the industrialization phase resulted in an on- ing, sales, loyalty marketing, and pricing. These groups
going capability. Dozens of independent data sets were must be brought together in order to develop a common
pulled in to feed the initial analytical model. Aer great understanding of the accountability imperative, a
N S
Stepping Up
to the Challenge
R
OCI isn’t just for consumer-packaged-goods ◊ Identify the minimum and maximum thresholds of
companies. In fact, some of the most ad- spending
vanced ROCI work being done today can
be seen in industries such as telecommuni- ◊ Shi their commercial mix to the vehicles with the
cations and financial services. Similarly, highest ROCI
the approach we are describing is not only for large, so-
phisticated companies with sizable commercial budgets. ◊ Deploy a single ROCI definition and toolkit across
Brands of all sizes around the world are using these commercial disciplines
tools and analytics.
◊ Develop a culture of commercial accountability
Of course, managers must take into account a company’s
individual context when anticipating results. Still, we be- To measure and improve ROCI continuously, companies
lieve our approach to be widely applicable and the ben- must first acknowledge and embrace the complexity of
efits it delivers significant. But it will require most com- the challenge. That means resisting simplistic rules and
panies to make some major changes. Companies that questioning traditional formulas for allocating budgets. It
adopt a more comprehensive ROCI approach will do the also means being ready and able to take a more compre-
following: hensive approach to commercial investments.
◊ Allocate spending to the businesses that are most re- With advertising budgets at many companies beginning
sponsive to commercial investment to approach or exceed capital expenses, there is a tre-
mendous amount to be gained by optimizing ROMI and
◊ Align commercial spending with commercial and cor- ROCI, and the need to do so has never been more ur-
porate objectives gent. Today, when growth is hard to come by and bud-
gets remain tight, the opportunity to reclaim or reinvest
◊ Focus their efforts on motivating the most valuable 15 to 30 percent of commercial spending is simply too
customers good to pass up.
Our meta-analysis summarizes results from 75 marketing- In all cases, we were careful to protect the confidentiality
mix models developed by Marketing Analytics for con- of our clients. Individual, anonymous identifiers were
sumer-packaged-goods brands. used for all brands in our sample (for example, BR000001).
Furthermore, all analytics were conducted on the sum-
mary input and output of individual models, rather than
Constructing the Data Set on content specific to the model itself.
Trade spending, however, required a different approach, total sales volume attributed to marketing
since it is notorious for not being governed by a consis- total sales volume of the brand
tent price list. We therefore estimated trade spending at
a constant 17.5 percent of dollar sales volume. Marketing Efficiency: Lower Is Better. We calculated
marketing efficiency as the dollar amount of marketing
Finally, we applied third-party estimates of gross margin expenditures per percentage point of unit sales volume
for manufacturers and retailers, at the category level, en- attributed to marketing (measured in either category-
abling us to calculate return on investment. equivalent units or retail sales):
N S
average annualized marketing expenditures ◊ We focused on consumer-packaged-goods brands in
percentage of total sales volume attributed to marketing the United States. The primary reason for this focus
was to leverage the relative similarities found across
Return on Marketing Investment: Higher Is Better. these brands and models in terms of inputs, outputs,
We calculated ROMI as incremental manufacturer reve- assumptions, and methodologies. When BCG and Mar-
nue in dollars per $1 spent on marketing: keting Analytics have worked in other industries, such
as entertainment, media, telecommunications, and
total sales volume in dollars total unit-sales volume
total unit-sales volume
× attributed to marketing
× 1 − retailer gross margin consumer electronics, we have found those brands to
total marketing expenditures be very different from brands in the consumer-pack-
aged-goods industry—and from one another.
We also calculated the incremental manufacturer gross
margin in dollars generated by $1 spent on marketing: ◊ The brands in our data set are predominantly well-es-
tablished, mature, and widely distributed, as are the
total sales volume in dollars total unit-sales volume
× attributed to marketing overall product categories in which they compete. On
total unit-sales volume 1 − retailer gross margin
total marketing expenditures
×
the basis of our experience, we anticipate that many
total trade spending younger, niche brands in rapidly growing categories
+ × manufacturer gross margin
total unit-sales volume
will encounter different results.
N S
Note to the Reader
Acknowledgments For Further Contact
The authors would like to thank BCG’s Marketing & Sales and Con-
Lasse Jensen, Dan Krohm, and Eric sumer practices cosponsored this
Stuckey for their role in developing publication. For inquiries about
our study and Megan Findley, Sally these practices or our approach to
Seymour, Mary DeVience, Gina Gold- achieving better returns on market-
stein, and Angela DiBattista for their ing investments, please contact one
contributions to the writing, editing, of the following BCG partners:
design, and production of this report.
Asia
Miki Tsusaka
BCG Tokyo
+81 3 5211 7939
tsusaka.miki@bcg.com
Europe
Patrick Ducasse
BCG Paris
+33 1 4017 1023
ducasse.patrick@bcg.com
Jens Harsaae
BCG Copenhagen
+45 7732 3400
harsaae.jens@bcg.com
The Americas
Neal Rich
BCG Chicago
+1 312 993 3300
rich.neal@bcg.com
Rohan Sajdeh
BCG Chicago
+1 312 993 3300
sajdeh.rohan@bcg.com
To receive future publications in electronic form about this topic or others, please visit our subscription website at
www.bcg.com/subscribe.
9/10