insight Predictably Irrational Customers
Optimizing Choices for How People Really Buy,
Not How We Think They Buy
Bill Abbott, Alex Mannella, Kyle McNamara, and Amaresh Tripathy
The Internet has a huge influence on consumer behavior worldwide.
Millions of customers use the Web, as well as cell phones and in-store
kiosks, to compare, contrast, and purchase products. Once a nascent
“marketspace,” the Internet is now an established shopping environment
in which even the smallest features can cause shifts in buyer behavior.
In spite of the new opportunity arising from online channels, most
companies do not test their e-commerce interfaces for revenue
maximization. Those that do, often fail to leverage the science of
behavioral economics, which documents how people really think, as
opposed to how they are supposed to think (as suggested by rational
economic theory). Companies that deploy behavioral economics to
improve their e-commerce interfaces can garner short-term profits and
long-term competitive advantage, because they are in a position to learn
faster than their competition about what really works.
Diamond thanks Dan Ariely, the Alfred P. Sloan Professor of Behavioral Economics, Massachusetts
Institute of Technology, for his contributions to this paper.
Diamond Partners Paul Blase, Anand Rao, and John Sviokla, and Senior Associate Eric Siegmann, also
contributed to this report.
Executive Targeting consumers with behavioral shopping cart abandonment rates remain
Summary economics principles clearly has wide- high and conversion rates remain low.
ranging impacts beyond online channels. One reason for this is that few companies
The Internet, however, provides us with carefully address all the subtle levers in
the necessary flexibility for testing purposes the sales cycle—from offer presentation
as well as the ability to mine an abundance through order execution—that can influence
of data. Once analyzed, the results can consumer purchase behavior. As a result,
be scaled to optimize online marketing efforts consumers often face a confusing array
and also be cascaded through brick-and- of product combinations and uncertainty
Measuring Value Through a mortar storefronts and other channels. around the online purchase process,
Structured Testing Approach causing them to turn to higher-cost channels,
Based on our marketing and e-commerce
migrate to a competitor, or delay the
project experience, and our partnership with
5. Prof. Dan Ariely, a highly respected behavioral
Tracking and 1. economist from MIT and a Diamond Valuable insights into consumers’ decision-
Generation Fellow, we believe firms can begin to drive up making behavior can be gained from
conversion rates by designing and testing new the field of behavioral economics. Behavioral
4. 6. models fast and frequently (Figure 1). economics brings together insights from
Execution 2. psychology and economics, as they relate to
Targeting Simple factors can have profound impact
human judgment and decision-making.
on success rates. In Europe, countries that
3. The central premise is that when confronted
Offer and had an “opt-out” organ donor strategy saw
Creative Design with limited resources (e.g., ability, time,
a seven-fold increase in donor participation;
information), human beings are not “rational”
firms that offered fewer 401(k) investment
and do not make the same economic
Figure 1 options radically improved participation
decisions as if complete and certain information
in retirement plans; a magazine that framed
purchase choice with three options rather
than two more than doubled the uptake of In collaboration with Prof. Ariely and
the pricier—and more profitable—option. through independent research, our findings
Details matter. It is necessary to employ have led us to propose a set of ideas and
a scientific approach to truly know what an approach to apply behavioral economics
to offer: If you don’t test your strategy, you principles to improve online marketing.
won’t learn what really works. We refer to the use of behavioral economics
to improve online marketing and sales as
While companies continue to spend heavily
table of contents on online marketing and sales, online
Executive Summary . . . . . . . . . 2
Current State of Online Marketing . . . . 4
Common Pitfalls Companies Encounter . . 6
Applying Behavioral Economic
Concepts Online . . . . . . . . . . 9
Getting Started . . . . . . . . . . 12
About the Authors . . . . . . . . . 15
For more information contact:
Bill Abbott, Partner
Key Behavioral Economics Principles
The field of behavioral economics addresses how consumers • Risk Transfer. People are willing to pay premiums to insure
use heuristics, or personal rules, to assist their decision- themselves from risk and uncertain outcomes.
making process. It further examines how heuristics combine with • Likelihood. Consumers often have a difficult time assessing risk.
alternatives in information presentation to affect the ultimate They tend to purchase warranties to protect against long-shot
choices of consumers. An online marketing strategy that maximizes failures while underestimating the likelihood of near-certainties.
consumers’ willingness to buy will carefully consider how consumers
• Happiness. Segregating windfalls (e.g., annuities) and pooling
approach the decision-making process. Further, understanding how
setbacks (e.g., credit card balances) minimizes unhappiness and
the framing and placement of products influence that process allows
companies to tailor their strategies to reap the most benefits.
To compensate for this behavior, companies must help lower
The three key behavioral economics principles that should inform
the risk associated with consumers giving up what they have in
any online marketing strategy are purchase paralysis, risk aversion,
exchange for the product being sold.
1. Purchase Paralysis 3. Framing
People draw inferences about a decision based on the way they frame
When consumers are offered an exhaustive variety of options, they are
the question or their response to it. They also evaluate options on relative
often paralyzed by their inability to choose. If they ultimately choose one
rather than absolute terms. The implication is that framing comparisons
of the options, they will be left less satisfied with their selection relative
can influence purchase decisions. Thus, configuring product or service
to the others (Schwartz, 2004). Causes of purchase paralysis include:
offerings online can guide a consumer’s current and future perception
• Regret. Too many choices might cause consumers to regret or and decisions. Actual reversals of preference are demonstrated in
anticipate regret for the possibility of having made the wrong choice. choices regarding money, both hypothetical and real, and in questions
• Opportunity Cost. It is difficult for consumers to assess trade-offs pertaining to chance (Tversky, 1981).
between what would be received and foregone.
• Escalation of Expectations. Comparing options relative to each
• Faulty Discounting. Customers are impatient with decisions that
other increases expectations, which can raise doubt that the
involve benefits received in the future. This leads to inconsistent
choice was correct post-purchase.
and often incorrect discounting of the future benefits. Instead, they
• Blame. Consumers will blame themselves and retailers for errors prefer to be presented with near-term savings, even if long-term
in judgment and selection. benefits are considerably more valuable.
When deciding on the number and variety of products to offer • Loss Aversion. Consumers tend to prefer gains over losses, even
consumers, companies must optimize the number of available if the values of the gain and loss are equal.
choices to encourage browsing and purchasing, and to minimize • Irrelevant Benchmarks. Consumers’ perceptions of a product
consumer regret that could damage the brand. can be influenced by providing comparisons to more expensive
2. Risk Aversion or premium-priced products.
Consumers tend to think differently about benefits and disadvantages • Priming. Consumers remember an item best in the form and
when considering alternatives. Using mental accounting, they keep context in which they first learned about it. This can be shaped
track of and evaluate transactions individually. They also tend to through advertising and product placement.
assess specific risks in isolation, rather than consider the broader
Companies can improve acceptance and purchase rates by
perspective of overall or long-term gains (Rabin, 2001).
carefully designing product value propositions that focus on the
• Mental Budgets. Consumers behave differently with respect to product’s immediate benefits. Adding frames of reference also
money that is allocated to accounts (e.g., savings for a car). Once helps consumers understand the relative value of the product, and
consumers mentally allocate money to such accounts, it is no shapes consumer perceptions of the product.
longer “free” for consumption.
Current State of Online channels continue to grow in importance. functionality and effectiveness of their Web
Online Marketing Sales through online channels accounted sites. More than half of large companies
for almost 3% of overall sales in 2006, and are increasing their Web site operations
this rate is growing at more than 25 percent budgets by more than 5 percent annually
annually (Figure 2). Forrester Research (Figure 4). Increasing usability, strengthening
estimates that almost $400 billion of store customer satisfaction, and demonstrating
sales—or 16 percent of total retail sales— ROI are among the top five challenges these
are directly influenced by the Web, as companies report.
consumers research products online and
Despite all this investment, a significant
purchase them offline (Forrester Research, Inc.,
gap exists between the share of consumers
The Web’s Impact on In-Store Sales, May 2007).
who want to shop online and those who
These “cross-channel” shoppers migrate from actually follow through with purchases. The
online channels for several reasons, including: percentage of customers who click through
a site all the way to conversion remains steady
• Desire for immediate gratification
at 3 percent, and e-commerce shopping cart
• Need to physically inspect or try on the abandonment rates range between 15 percent
item before buying and 90 percent, with an average of 60 percent
• Concern about online security and (Brinker, 2007; Baxter, 2007).
The keys to improving the success rate of
Behavioral economics is These will likely remain significant factors, but these investments are: 1) employ a rigorous
concerned with the ways over the next five years, the value of retail sales approach to designing product presentation
influenced by the Web will increase by 63 percent, and the end-to-end purchase process, and
in which the actual decision- 2) evaluate and refine product offers over time.
and cross-channel sales will experience a
making process influences 17% CAGR to over $1 trillion in 2012 (Figure 3). Unfortunately, today we see a daunting and
loosely connected array of offers from companies
decisions that are made As this market grows, companies continue that are not consistently applying controlled
in practice. to spend significantly on improving the tests to their offers or the purchase process.
E-Commerce Sales as a Percentage of Total Sales
E-Commerce Sales ($B)
2000 2001 2002 2003 2004 2005 2006 2007
Source: U.S. Census Bureau
We believe companies can increase the make. Traditional economics suggests that their chances of obtaining this goal. In reality,
effectiveness of their online marketing efforts consumers make rational, consistent however, too much choice can become
and sales channels by applying behavioral choices to obtain products or services that a barrier in the consumer’s decision-making
economics concepts. Behavioral economics best meet their needs at the least possible process, along with other obstacles such as
provides insight into how the decision-making cost. This suggests that presenting consumers increased complexity, uncertainty, and limited
process influences the decisions consumers with a wide range of choices increases information around a selection’s future value.
Forecast: U.S. Cross-Channel Sales, 2006–2012
Total U.S. Retail Sales
Total Cross-Channel Sales
Total eCommerce Sales
2006 2007 2008 2009 2010 2011 2012
Source: Forrester Research, Inc., The Web’s Impact on In-Store Sales, 2007
Change in Site Operations Budgets from 2006 to 2007
than 20% Stay the Same
Increase by 11%
Increase by 1%
Increase by 6%
Source: JupiterResearch, Web Site Spending, 2007
Common Pitfalls Companies can drive higher conversion motivation to purchase the product
Companies Encounter rates by applying Choice Optimization. But (Iyengar, 2000). These studies imply that
before tapping into these strategies, companies can determine an ideal
companies need to understand how offers number of choices that will optimize the
might negatively impact consumers. behavior of certain segments, and we
We believe there are three main reasons believe these core concepts can be applied
for suboptimal online sales performance: effectively to online marketing.
1. Too many choices. If an e-commerce To examine the impact of choice on product
site requires that customers choose satisfaction, researchers conducted
between a myriad of product options or a study where participants selected from
configurations, customers might fail either a limited assortment or an extensive
to complete the decision process. Adding array of chocolates. Participants then
options also increases the risk that sampled their selections and were given
consumers will regret their purchases or the option to receive cash compensation
have unreasonably high expectations about or an equivalent value of the chocolate they
the value of their purchases. chose. Participants with the extensive
array found the decision-making process more
Research has found that increasing the
Choice Optimization is the enjoyable than those with the limited
number of mutual funds available to
array, but were later more dissatisfied
application of behavioral employees in their 401(k) plans can result
and regretful of their choice and
in lower participation rates. One
economics principles in online were considerably less likely to choose
employer found that for every 10 mutual
chocolates over money as compensation.
marketing to simplify the funds included in its program, participation
(Iyengar, 2000). Having a greater number
dropped 2 percent, even though opting out
decision-making process and of options to choose from ultimately
of the plan meant passing up the opportunity
forced these consumers to abandon the
increase conversion rates. to receive up to $5,000 in employer
matching funds (Schwartz, 2004). In
another study, researchers set up
tables in an upscale grocery store that
Too much choice leads
featured a line of exotic jams. Consumers
to purchase paralysis
who stopped could taste samples and
they received a $1 coupon if they bought
Conversion Rates from Jam Presentation
a jar. Consumers were presented
with either six or 24 jam varieties. When 6 Jams
24 varieties were available, more 80
people stopped at the table than when 60%
six were available, but only 3 percent 60
of people examining the 24 varieties made 40%
a purchase, compared with 30 percent 40 30%
of those exposed to the six varieties
(Figure 5). The researchers concluded that
an extensive array of options can be
appealing at first glance, but a wide variety Approach Buy Conversion
also can subsequently reduce consumers’ Source: Iyengar, 2000; Diamond analysis
2. Process uncertainty. Customers may Companies are starting to incorporate reaffirmed this strategy, citing the free
be dissuaded from completing a transaction tools to make the purchase process more shipping offers “as an effective worldwide
if a Web site does not clearly help them predictable and secure. Google Checkout, marketing tool” that the company intends
place an order, determine when they will for example, integrates with merchant Web to continue offer indefinitely.
receive it, understand when and how sites, allowing customers to use their Google
Each of these three factors represents an
much they will be billed, or feel that their accounts to complete a purchase instead of
area of increased risk for not acquiring
information is secure. creating a separate account at each merchant.
potential online customers. Ignoring these
Other companies are letting consumers use
The average e-commerce shopping cart factors has serious consequences:
PayPal in addition to credit and debit cards.
abandonment rate is around 60 percent.
Offering a third-party payment mechanism • Lost or delayed sales due to consumer
Of this figure, 48 percent of carts are
with built-in security features and consumer indecisiveness or anxiety;
abandoned during the checkout process.
familiarity reduces risk and uncertainty in
Many of the factors contributing to • Migration to higher-cost channels due
the purchase process.
abandonment are related to uncertainty to customer impatience or uncertainty; and
in the process, including hidden charges 3. Faulty discounting. If a Web site over-
at the checkout, lack of clear delivery emphasizes the long-term benefits of • Inconsistent expectations of the product’s
details, and discomfort with the buying a product relative to its short-term appeal, future value which may lead to increased
process (Baxter, 2007). consumers might be dissuaded from downstream support, churn, and damage
purchasing because they tend to incorrectly to the brand.
To reduce uncertainty, companies can
and inconsistently discount future
add cues within the checkout process that
benefits. Consumers value current over future
guide customers through the process. For Companies with Complex
consumption, even when that future
example, Esurance, an insurance company Products at Risk
consumption is considerably more valuable.
with a heavy online presence, developed Companies in several key industries are
a simple, three-step process to purchase In one study, the subjects were told that vulnerable to these risks, due to the
insurance, and adopted it as their slogan they had won $15 in a lottery and that they complexity and breadth of their product
(“Quote. Buy. Print.”). This process could take the money now or wait until and service offerings:
is prominently displayed as the consumer later. Participants were asked how much
• Wireless Carriers. Wireless companies
progresses from obtaining a quote to they would require to make waiting just as
are offering a wider variety of handsets
printing a policy, providing a simple visual cue attractive as getting the money now. The
(e.g., clamshell, candy bar, camera,
as to progress through the transaction. median responses were $20 in one month,
3G, smart phones) along with an array
Additionally, companies such as Dell and $50 in one year, and $100 in 10 years.
of voice and data plans (e.g., local,
Amazon provide estimated shipment These figures imply a compound discount
national, 3G data) and additional services
and delivery dates throughout the process rate of 345% over the one-month period,
(e.g., text and picture messaging).
to properly set consumer expectations. 120% over the one-year period, and 19%
Many of these options also have variants,
over the 10-year period (Thaler, 1981).
The risks of credit card theft and identity theft including handset color; inclusion of a
The large discount rate seen in the short-
also deter some customers from completing camera; and number of allowable minutes,
term periods illustrates consumers’
online purchases. Among consumers who messages, or kilobytes of data per
preferences for short-term gains.
have never shopped online, the top reason month. When choosing a plan, Web sites
they cite for not doing so is the fear of Offers of short-term gains can also influence do not effectively segment customers
providing their financial information online. consumers into buying products. For example, or provide historical usage data that might
Among consumers who have shopped online, Amazon.com offers free shipping on orders be available from a representative in
17 percent have stopped buying online over $25, which may encourage consumers the store or on the customer support line.
altogether because of fears of identity theft to buy more than they intended to order Potential areas of segmentation could
(Forrester Research, Inc., Checkout Tools to meet the threshold for free shipping. In include average minutes per month, roaming
That Boost eBusiness, January 2007). a July 25, 2006, press release, Amazon patterns, and international calls.
• Cable and Internet Providers. Cable • Financial Services Firms. Credit card • Insurance Companies. Insurance policies
and high-speed Internet providers are cross- companies offer a complex array of are based on a complex set of factors (e.g.,
selling tiers of core products along with services, including cards linked to checking the item being insured, the coverage levels
telephone service, packaging them into accounts, cards with branded rewards, sought, and the discounts for which the
bundles to save customers money. However, and cards with unique payment options. To insured is eligible). Many companies allow
each of these products has a number of prevent the risk of overwhelming customers consumers to obtain rate quotes online,
variants, including a wide range of premium with options, many companies leverage but there are many options from which to
channels and Internet connection speeds. self-segmentation questioning to narrow select. The initial quote is generally based
This diversity of choice places a large the list of choices. on a default set of coverage options and
burden on consumers to research and find discounts, but it is left to consumers to adjust
the ideal plan. coverage based on their preferences.
Testing Online Marketing Efforts for Maximum Impact
Diamond uses the following structured testing approach
to generate insights, provide value, and improve customer
1. Hypotheses Generation. Review background data to develop Tracking and 1.
hypotheses around the types of offers desired by each segment.
Leverage segmentation schemes to ensure targeting of the
2. Targeting. Identify the targeting criteria based on hypotheses. 4.
Determine to whom and to which channel the offer will be Execution 2.
presented. Size both target and control groups. Targeting
3. Offer and Creative Design. Develop a compelling offer based
on hypotheses, channels, and segments. Offer and
4. Test Execution. Test multiple options at the same time, using
a multivariate test design, and execute the marketing campaign.
5. Tracking & Analysis. Analyze offer effectiveness by collecting
responder data. Calculate response rates for target and
control cells and evaluate the test economics to decide whether
the test offer should be rolled out or not. Finally, gather
learnings from the test cycle to refine hypotheses.
6. Solution. Roll out successful offers to the broader population.
This new offer becomes the control for future campaigns.
Applying Behavioral Research in behavioral economics has shown But results also must be filtered to prevent
Economic Concepts that conversion rates can be affected by recommending items that the consumer
seemingly subtle presentation factors. For has already purchased. Netflix employs
Online example, an online publisher tested a collaborative filtering approach that
two offers on its Web site. The first offered leverages user ratings, movie genre
“7 Days Free,” and the second “7 Days frequency, and cross-sell likelihood. As a
Risk-Free.” The words “Free” and “Risk-Free” result, the online movie rental service is
were bolded to highlight the key benefit, and able to quickly filter through thousands of
the second example resulted in a 12 percent movie titles to show consumers one title
higher conversion rate than the first option. they can have collaborative faith in renting.
In a separate experiment, the same publisher
• Explicit filtering (also known as “self-
tested two different page layouts, the
segmentation”) requires consumers to
first containing two columns of text and the
select or rate their preferences in order
second containing only one. Simplifying the
to narrow the number of options presented.
layout by combining two columns of text into
Citibank and U.S. Bank allow customers
one produced an 88 percent improvement in
to self-segment by identifying the credit
conversion rates (McGlaughlin, 2006).
card features most important to them.
The application of Based on our work with Prof. Ariely, as well as U.S. Bank asks its customers to select up
our own research, we believe there are three to seven key features (e.g., traditional
seemingly subtle specific ways companies can use core behavioral cards, student cards, travel or cash rewards),
presentation factors economics principles to improve online sales: in order to present them with cards that
match those preferences. Since this data
can significantly 1. Choice Filter. Offering customers
directly reflects consumer choices, it can
excessive options can overwhelm them
affect conversion rates. be especially useful in understanding their
with a fear of making the wrong choice,
preferences. However, since providing this
or cause them to anticipate regret, which
information requires consumers to expend
can lead to postponing a purchase.
effort, companies often have less data
The ability to properly filter through an
than what they desire. In addition, this
exhaustive list of products or services
process indicates to customers that there
can help consumers complete the purchase
are additional options available to which
process. Companies can provide collaborative,
they are not being exposed. Also, explicit
explicit, or implicit filters—either
filtering still introduces the risk of regret.
individually or in combination—to help
customers narrow their choices. • Implicit filtering involves identifying
consumer preferences based on their
• Collaborative filtering examines the
actions (e.g., items viewed or purchased),
preferences of many users to make
or on the demographic information
predictions about the interests of any
they provide. Data on items viewed
one user. Web sites such as Amazon
and purchased during a session gives
and Netflix use this technique to make
insight into preferences, and information
recommendations to their users. The
such as customer billing zip codes can
relevance of recommendations can vary,
be matched to databases from vendors
but they tend to improve the collection of
such as Claritas and Acxiom to obtain
larger amounts of data. These techniques
socioeconomic characteristics. For the
are good for categories with large numbers
most part, data collection is automatic and
of products (e.g., books, music) because
largely independent from the consumer,
a single person cannot view all of the items.
Companies across industries can benefit from applying behavioral economics principles
to their online sales channels.
A Sampling of How Behavioral Economics Principles Can Impact Consumers
Industry Too Many Choices Process Uncertainty Faulty Discounting
Wireless 1,900 Combinations1 • A four-step process is presented to choose • A trial period allows customers to cancel
a phone and plan followed by a separate within a few weeks, but they must pay service
four-step process to complete the purchase. charges. A shorter, risk-free trial period
Presenting one process to guide the might attract more customers because it
consumer through the full transaction could reduce current consumption in favor
would reduce uncertainty about next steps. of long-term benefits.
• The activation fee is described in small print • Free 2–4 day shipping is offered for online
at the bottom of the page, where customers purchases, forcing customers to delay
might not see it. Displaying the fee more benefits. Free overnight shipping reduces
clearly as part of “one-time charges” helps the delay before benefits are realized and
avoid surprises. might avoid delayed decisions or migration
to a higher-cost sales channel.
Cable 64 Combinations2 • The site indicates that a limited package • Savings resulting from choosing a package
selection is available online, but full instead of individual services are not
range is available via phone. Customers consistently indicated. Displaying the specific
are likely to call Customer Service, a higher- savings will lead customers to focus on the
cost sales channel, to understand all options. product’s near-term value.
Credit Cards 34 Unique Options3 • Sites do not contain a timeline to help • The annual fee to earn rewards in effect
customers understand their progress promises customers future rewards in
through the application process. Including exchange for current consumption.
timing elements helps customers reduce Customers may incorrectly value the benefits
uncertainty about next steps. received over time.
Insurance 65 Combinations4 • Sites do not contain timing elements to • Many companies add a service charge for
help customers understand their progress the option to pay premiums monthly.
toward obtaining an online quote. Including Framing this fee instead as a discount
details about timing helps reduce customers’ for payment in full helps customers realize
uncertainty about next steps. the near-term value of their choice,
instead of having to value it into the future.
Examination of a national wireless carrier’s consumer Web site revealed 41 handset models available for purchase (excluding color variations), eight plan types with six levels
of minutes within each. Total number of phone/plan combinations available is 1,968 (41 x 8 x 6).
Examination of services available from a national cable provider’s Web site revealed seven choices for cable service, three for Internet, and one for phone. Consumers have the
option of only choosing one or two of the services (e.g., cable and Internet, cable and phone, Internet and phone), so each category also has an “opt-out” option. Total number of
combinations available is 64 (8 x 4 x 2).
Examination of consumer cards available on a major carrier’s Web site revealed 34 unique options.
Examination of options required to obtain an auto insurance quote on a national carrier’s Web site revealed that consumers must choose coverage levels within 10 categories,
along with up to nine discounts for which they are eligible. Each of the 10 coverage categories (e.g., property damage, collision deductible) contains between two and nine
coverage levels (e.g.,$25–50, $500–1000), for a total of 56 combinations. The nine discounts offered a “Yes/No” option. Total coverage/discount combinations is 65 (56 + 9).
meaning that much more data is or comparable products at other only two choices were presented
available. Behavioral data, however, does companies to help consumers choose (“Online Only” and “Online and Print,”
not necessarily provide an accurate to buy their products. at $59 and $125, respectively), the
representation of a user’s true opinion number of customers choosing the lower-
• Satisfaction risk often results when
of an item. For example, some purchases priced “Online Only” product increased
consumers are faced with too many
will be given as gifts and do not to 68 percent, while the percentage
choices. For many consumers, examining
necessarily reflect the consumer’s own of subscribers choosing “Online and Print”
a wider array of product options raises
preferences. Companies such as Best dropped to 32 percent. While no one
their expectations because they search
Buy and Target use both types of implicit chose the “Print Only” option, having it
for a product that contains many
filtering through a consumer’s account available made the more expensive
of the most desired attributes. This also
information, viewing history, and “Online and Print” option appear to be
increases the chance that consumers
purchase history. By using algorithms, a bargain, and this drove a higher
will regret their ultimate choice because
these retailers then present the percentage of customers to select it.
another option might have better met
customer with better-defined product Amazon.com effectively cross-sells
their needs. Verizon Wireless reduces
or service options. additional products to consumers by
this risk by offering a 30-day return
highlighting the remaining amount of a
2. Risk. In theory, people are willing to pay policy. When the fear of satisfaction is
balance required for free shipping. The
a premium to avoid risk. Eliminating any a deal-breaker in completing the
relative framing of additional purchases as
perceived risks in the purchase process can purchase of a phone and plan, consumers
savings in shipping costs helps consumers
help increase consumer conversion rates. can feel reassured that they have
overcome the reality of a higher total
30 days to return the phone and cancel
• Process risk is introduced when consumers price after additional items are added.
their contract at limited or no cost.
do not understand the steps in the
• Constructed framing allows companies
purchase process, such as when their 3. Framing. The way in which choices are
to position a preferred product for
items will be shipped, or when their presented can influence a consumer’s
maximum exposure. For example, T-Mobile
credit cards will be charged. Hewlett- decision to purchase a product. Effective
strategically highlights one of its most
Packard helps eliminate process risk frames give consumers points of
popular plans in a list of other options.
by presenting a clear purchase process, reference to understand the relative value
While the highlighted plan is not
providing estimated shipping and of a product compared to alternatives.
the most expensive, its price is above the
delivery dates, telling the consumer when Understanding the impact of framing can
average. Emphasizing this plan begins
their credit card will be charged, help companies cross-sell, up-sell, or steer
to set customer expectations about cost.
and allowing the customer to click on a customers toward more profitable options.
By steering the consumer selection
previous step in the process to review
• Relative framing highlights the positive this way, the carrier is positioned to sell
or revise the information they provided.
value of a decision compared with a more expensive plan than it might
• Price risk arises when consumers an alternative. The Economist magazine have sold without constructed framing.
delay or forgo a purchase because once offered three subscription options
• Priming allows companies to make
they think that a better price might on its Web site:
consumers more receptive to their
be available from another retailer or
1. Online Only for $59 products by creating associations with
channel. Bed Bath and Beyond
items that consumers value. Fidelity’s
has significantly reduced price risk 2. Print Only for $125
Web site allows consumers free access
for its customers through the use 3. Online and Print for $125 to tools for financial planning and
of price matching. Consumers feel they
investment research. As consumers use
are getting the lowest price available Sixteen percent of customers chose the
these tools, they increasingly get the
since the retailer advertises the price “Online Only” option and 84 percent
impression that Fidelity has a wealth of
matching policy. Other companies, selected the “Online and Print” option. No
knowledge and would make the financial
such as Progressive Insurance and customer chose the “Print Only” option,
planning experience easy.
Buy.com, display prices of the same so the company removed it. However, when
Getting Applying behavioral economics principles of improvements. Useful baseline metrics
Started to improve online marketing and sales requires for defining and measuring the progress
companies to take a careful look at where of improvements to online sales channels
they currently stand. It also requires creative include the following:
thinking about applying the principles, and
• Number of available product
a rigorous process to test and measure their
impact (Figure 6).
• Number of clicks to purchase;
1. Define the Focus. Commercial Web sites
• Cart-abandonment rate;
typically have multiple purposes, including
consumer education, prospect development, • Number of steps completed before
lifecycle management (acquisition, abandonment;
stimulation, cross-sell, up-sell, retention, • Conversion rates;
winback), and self-service. Key consumer
• Customer satisfaction scores; and
decision points will vary across companies
and industries, and each company has its • Cost per transaction.
own pain points. Choice Optimization efforts
As choice optimization efforts are being
should focus on areas likely to produce
implemented, monitoring progress
a high return on investment. For example,
against the baseline helps determine
consumers visit a car company’s Web site
whether the efforts are truly improving
primarily to learn about a car before buying
processes. Tools that track Web site
one at a dealer. The focus here should
activity, such as Omniture and Google
be providing a rich, targeted education on
Analytics, can also be useful in identifying
features, options, and pricing, then helping
pain points in the purchase process,
the customer find a dealer. On the other hand,
including the stage at which customers
consumers who visit Amazon.com are more
abandon their carts.
likely to complete an online purchase, so the
focus should be on helping customers 3. Identify and Prioritize Improvement
narrow down their options and quickly Opportunities. One of the most effective
complete the purchase. ways to determine where behavioral
economics principles can deliver superior
2. Establish Baseline and Improvement
results is to examine the process from
Targets. Establishing a baseline is crucial
the consumer perspective. Documenting
for identifying areas where behavioral
the processes used by customers in
economics principles can have the most
searching for product information uncovers
impact, for preparing a convincing business
opportunities for improvement. Such
case, and for measuring the effectiveness
opportunities may include framing options
Diamond’s Choice Optimization Approach
2. Establish 3. Identify 4. Test and
1. Define Baseline and and Prioritize Implement
the Focus Improvement Improvement Improvement
Targets Opportunities Opportunities
more strategically, articulating next steps analyze offers, such as the one outlined 1. Optimize the number of choices available
more clearly, or eliminating underperforming in “Testing Online Marketing Efforts to encourage browsing and purchasing,
choices. Further hypotheses can come for Maximum Impact” on page 8, helps and to minimize customer regret.
from brainstorming workshops, root-cause deliver the right products to the right
2. Help lower the risk consumers associate
analysis of Web data, customer surveys, customers. This works because it teaches
with completing an online transaction.
and industry research. Quantifying and companies to understand the profit drivers
prioritizing the costs and benefits of each for each segment, identify characteristics 3. Focus more heavily on their products’
improvement maintains the focus of the of effective offers within each segment, and immediate benefits (and discount future
Choice Optimization efforts. document lessons learned that can be benefits to compensate for consumers’
applied to the development of future offers. inability to correctly value them).
4. Test and Implement Improvement
Opportunities. Continuous hypothesis- Choice Optimization applies these behavioral
based testing, in which multiple versions Conclusion economics principles to improve product framing
of an offer are tested using statistical Incorporating key elements of the consumer with the goal of positively influencing consumer
techniques, improves profitability by decision-making process into the development decisions for improved online marketing.
allowing companies to choose the offers of an online sales and marketing strategy
with the highest success rates. Using can help maximize the effectiveness of that
an approach to segment customers and strategy. Specifically, companies must:
Additional Diamond insights on behavioral economics and marketing are available at www.diamondconsultants.com:
“Behavioral Economics Comes to the Rescue of Retirement Savings,” 2007
“Rapid Market Segmentation: Collaborate and Conquer,” 2007
“Word-of-Mouth Marketing: How to Harness the Power of Consumer Influence,” 2007
“Curing Customer Churn,” 2006
The following papers provide good insight into key behavioral economics principles and their application:
“Placebo Effects of Marketing Actions: Consumers May Get What they Pay For.” Dan Ariely et al, 2005.
“Amazon.com Recommendations. Item-to-Item Collaborative Filtering.” Greg Linden et al, 2003.
“The Paradox of Choice: Why More is Less.” Barry Schwartz, 2005.
“The Winner’s Curse.” Richard Thaler, 1994.
“Why the Soy Milk King Still Reigns.” G. Pascal Zachary, 2004.
Sources Baxter, Mike. Online Retail 2007. Special Report: Checkout. 2007. London: E-Consultancy.
Brinker, Scott. Minding the Gap Between Click and Conversion. 20 February 2007. Encino,
California: E-Commerce Times, 22 May 2007. <http://www.ecommercetimes.com/
Drèze, Xavier, Stephen Hoch, and Mary Purk. Shelf Management and Space Elasticity.
Diss. Graduate School of Business, The Univ. of Chicago, 1994. 25 May 2007
Forrester Research, Inc. The Checkout Tools That Boost eBusiness. January 2007.
Forrester Research, Inc. The Web’s Impact on In-Store Sales: US Cross-Channel Sales
Forecast, 2006 To 2012. May 2007.
Iyengar, Sheena S. and Mark R. Lepper. “When Choice is Demotivating: Can One Desire Too
Much of a Good Thing?” Journal of Personality and Social Psychology. 2000, Vol. 79, No. 6,
JupiterResearch, LLC. Web Site Spending, 2007. 17 April 2007.
McGlaughlin, Flint. “The Power of Small Changes Tested.” 21 March 2006. Marketing
Experiments Journal. 23 May 2007. <http://www.marketingexperiments.com/improving-
Rabin, Matthew, and Richard H. Thaler. “Anomalies: Risk Aversion.” Journal of Economic
Perspectives. 15 (2001): 219-232.
Rayport, Jeffrey F., and John Sviokla. “Managing in the Marketspace.” Harvard Business
Review. November–December, 1994, pp. 141–150.
Schwartz, Barry. The Paradox of Choice. New York: HarperCollins, 2004.
Thaler, Richard H. 1981. Some empirical evidence on dynamic inconsistency. Economic
Letters, 8, pp. 201-07.
Tversky, Amos, and Daniel Kahneman. “The Framing of Decisions and the Psychology
of Choice.” Science. 1981, Vol. 211, No. 4481, 453-458.
United States Census Bureau. Quarterly Retail E-Commerce Sales. 16 May 2007.
Washington, D.C. 18 July 2007 <http://www.census.gov/mrts/www/ecomm.html>.
About the Authors Bill Abbott is a Partner in Diamond’s Telecom & High Tech practice and co-leader of the firm’s
information and analytics services. Bill has significant experience in the communications and
media, consumer packaged goods, and insurance industries, collaborating with clients to
successfully address challenges in the areas of channel optimization and marketing strategy,
organization and process design, including system design and implementation, financial and
operational modeling, and activity-based process and product costing.
Dan Ariely, the Alfred P. Sloan Professor of Behavioral Economics at Massachusetts
Institute of Technology, is a Diamond Fellow whose projects include examinations of online
auction behaviors, personal health monitoring, the effects of different pricing mechanisms,
and the development of systems to overcome day-to-day irrationality. Prof. Ariely holds a
joint appointment between MIT’s Program in Media Arts and Sciences and the Sloan School
Alex Mannella, a partner in Diamond’s Telecom & High Tech practice, is an expert on
the issues of customer value management, response modeling, database marketing, and
data mining. He has extensive experience developing response and segmentation algorithms,
and using artificial intelligence for data mining studies. Mannella, who is also skilled in
database development, currently helps lead the Diamond Information and Analytics Center
located in Mumbai, India.
Kyle McNamara is a Manager in Diamond’s Telecom & High Tech practice who has
significant experience in the areas of organization and process design, operations
management, revenue assurance, and Web analytics. He has served a variety of clients in
the telecommunications, electric utility, financial services, and manufacturing industries.
Amaresh Tripathy is a Manager in Diamond’s Telecom & High Tech practice and a thought
leader in the area of information and analytics. He has a wealth of experience in the areas
of customer value management, channel optimization, competitive analysis, and operations
management. Amaresh has supported opportunity strategy and implementation initiatives
for companies in the telecom technology, logistics, education, and entertainment industries.
Diamond Partners Paul Blase, Anand Rao, and John Sviokla, and Senior Associate
Eric Siegmann, also contributed to this report.
Diamond thanks Dan Ariely for the title of this paper, also the title of his upcoming book
on behavioral economics.