Goal Gradient Hypothesis

Research Project Presented to

MPSTME, NMIMS


In Partial Fulfilment of the Requirement
For
the Degree MBA Tech (Information Technology)


By
Sanket Baxi
Roll no: 103

2008-13
ACKNOWLEDGEMENTS
I am deeply grateful to my mentor Prof. Prashant Kulkarni for his expert advises & guidance,
involvement and whole hearted cooperation, which inspired me to complete this research
project successfully.
I take this opportunity to express my sincere thanks to all my friends for helping me to carry
out my project work by participating in my experiments.
I would also like to thank Mr. Anil Kumar and Mr. Ashish Raj from RedSeer Consulting for
their advice on how to proceed with my research project.

Sanket Baxi
MBAtech IT
Contents
ACKNOWLEDGEMENTS ....................................................................................................... 2
ABSTRACT ............................................................................................................................... 4
INTRODUCTION ..................................................................................................................... 5
STATEMENT OF THE PROBLEM ..................................................................................... 5
PURPOSE OF THE STUDY ................................................................................................. 5
SIGNIFICANCE OF THE STUDY ....................................................................................... 5
LITERATURE REVIEW .......................................................................................................... 6
EXAMPLES OF GOAL GRADIENT HYPOTHESIS ............................................................. 7
INITIAL STUDIES ................................................................................................................... 8
METHOD ............................................................................................................................... 8
RESULTS............................................................................................................................... 9
RESEARCH STUDIES ........................................................................................................... 10
METHODOLOGY ............................................................................................................... 10
Experiment 1..................................................................................................................... 10
Experiment 2..................................................................................................................... 12
KEY FINDINGS AND THEIR IMPLICATIONS .................................................................. 14
GENERAL DISCUSSION ...................................................................................................... 15
LIMITATIONS AND FURTHER STUDIES ......................................................................... 17
CONCLUSION ........................................................................................................................ 18
REFERENCES ........................................................................................................................ 19


ABSTRACT
The goal-gradient hypothesis denotes the classic finding from behaviorism that animals
expend more effort as they approach a reward. Building on this hypothesis, we generate new
propositions for the human psychology of rewards. We test these propositions using field
experiments, secondary customer data, paper-and-pencil problems, and Tobit and logit
models. The key findings indicate that (1) participants in a real caf reward program purchase
coffee more frequently the closer they are to earning a free coffee; (2) Internet users who rate
songs in return for reward certificates visit the rating Web site more often, rate more songs
per visit, and persist longer in the rating effort as they approach the reward goal; (3) the
illusion of progress toward the goal induces purchase acceleration (e.g., customers who
receive a 12-stamp coffee card with 2 preexisting bonus stamps complete the 10 required
purchases faster than customers who receive a regular 10-stamp card); and (4) a stronger
tendency to accelerate toward the goal predicts greater retention and faster reengagement in
the program. The conceptualization and empirical findings are captured by a parsimonious
goaldistance model, in which effort investment is a function of the proportion of original
distance remaining to the goal. In addition, using statistical and experimental controls, we
rule out alternative explanations for the observed goal gradients. We discuss the theoretical
significance of our findings and the managerial implications for incentive systems,
promotions, and customer retention.

INTRODUCTION
STATEMENT OF THE PROBLEM
In a world where most airlines, big retailers and chains, including coffee shops and food
chains, offer reward point systems where customers accumulate points for shopping which
they can later expend for free rewards or discounts, it is necessary to understand the
consumer psychology towards such systems and how they can be tweaked to achieve the
most profitable outcomes. We need to understand how to structure the reward system, how
rewards should be handed out and how long or how expensive must it take/be for consumers
to achieve rewards.
PURPOSE OF THE STUDY
This study tries to understand the impact of goal gradient hypothesis in consumer reward
systems. We will try to figure out whether consumers tend to expend more in terms of time or
money as they get closer to achieving a reward and, if yes, how we can use this psychological
understanding to better structure rewards in a way that consumers are encouraged to spend
more on the brand.
SIGNIFICANCE OF THE STUDY
This study will help determine an efficient reward structure that will be beneficial for both
the brand and the consumer.

LITERATURE REVIEW
We look at reward systems like loyalty programmes where customers accumulate points for
shopping and are later allowed to use them for free shopping or discounted shopping. An
example is a coffee shop that gives its customers a free coffee after they have purchased ten
cups of coffee. A close example is MODs (Mad Over Donuts) reward scheme where you
buy 12 boxes of half a dozen donuts, you get a box free.
The effect of goal gradient hypothesis was first observed in rats in the 1930s. It was observed
that rats who were left in one corner of a box with the food in another corner tended to run
faster towards the food as they got closer to it. In other words, as the distance to the goal
decreased, their efforts in achieving it increased. The impact of this effect on humans is not
well studied, and a recent study was done by Ran Kivetz, Oleg Urminsky, and Yuhuang Zhen
in two ways:
1. They studied users of a website called JaBoom where they were offered a $25
Amazon gift card for every 51 songs they rated.
2. They studied two groups of caf visitors, one of which got a free coffee after
purchasing 10, and another which did not get anything. The first group was further
divided into two groups, one which started with a blank card with ten areas for
stamps, another which got cards with twelve stamp areas, but two of which were
already filled out as a bonus. This was done to study the effect of illusion of
progress.
Using the results, they understood whether humans, too, tended to accelerate towards to goal
as they near it. In both cases, the answer was in the affirmative. It was realised that users of
JaBoom started spending more time on the site rating songs as they neared their goal of 51.
Also, visitors of the caf purchased coffee at lesser and lesser intervals as they neared their
goal of ten coffees. Two interesting phenomena were noticed:
1. Those who started with cards with ten blank spots tended to take more time to
purchase ten coffees then those who started with cards with twelve spots of which two
were pre-stamped. It was realised that giving out bonus stamps created an illusion of
having progressed towards the goal, even though the total number of purchases
remained the same as in the earlier case.
2. Once the reward was achieved, their speed was reset to the original speed. i.e. once
they got their Amazon gift card or the free coffee, their rate of purchase decreased
again and the acceleration cycle was repeated, meaning that the effect of goal gradient
is temporary.
This leads us to ask:
1. How much effort should users be asked to put in so that the goal gradient cycle is
short as well as profitable?
2. Should users be given some bonus while first signing up to give them an illusion of
progress while simultaneously increasing the number of reward points needed for the
reward?
EXAMPLES OF GOAL GRADIENT HYPOTHESIS
Customer loyalty programmes
Loyalty programmes are now commonplace. Customers are given cards when they sign up.
Every time they shop at particular locations, points are accumulated. These points can be
redeemed after a certain number is accumulated.
Stamp cards
Stamp cards are handed out in places like MOD. Unlike loyalty programmes which are
continual, these stamp cards only last till a certain number of purchases. After every
purchase, a stamp is made and once the card is full, the next purchase is free.
Frequent Flier miles
Airlines are known to offer a reward system similar to loyalty programmes where customers
accumulate miles based on how often they fly and how many miles they cover. These
miles can be redeemed for free tickets.
Websites
Certain websites depend on repeat visitors for revenue. Such sites offer rewards for users who
visit often, or perform certain actions. Eg. An online radio might offer free song downloads to
users who listen for a certain duration.
INITIAL STUDIES
To facilitate a strong and realistic test of intertemporal behavior, we conducted a field study
in which customers made real coffee purchases in the context of an actual caf RP. By
tracking purchases, we were able to test for purchase acceleration toward the reward goal
(i.e., H1). The study included two control groups: (1) members from whom we bought back
incomplete cards and (2) customers participating in an experimental control program that was
identical to the actual RP, except that purchasing coffee did not earn rewards. The inclusion
of these control groups enabled us to compare the intertemporal purchase behavior of
redeemers and nonredeemers (i.e., loyals and defectors) and to examine differences
between reinforced and nonreinforced behavior. We also investigate alternative explanations
using various methodologies, including testing a key corollary termed postreward resetting,
exploring the behavior of the two aforementioned control groups, and incorporating
unobserved heterogeneity in the tendency to accelerate.
METHOD
The participants in the field study were customers of a caf located within the campus of a
large East Coast university. At the time of data collection, the caf had several oncampus
locations. Customers were offered free enrolment in a caf RP, in which they needed to make
ten coffee purchases to earn a reward. To enable the tracking of their purchases, members
were required to carry a frequent-coffeebuyer card . They received one stamp on the card for
each coffee purchase they made (only one stamp per visit was permitted). Stamps were
printed with six-wheel automatic numbering machines that, unbeknownst to customers,
sequentially numbered each stamp issued (these numbers did not resemble dates). After
members accumulated ten stamps from any combination of the caf locations, they were
eligible for a free coffee redeemable on their next visit to one of the caf locations. Members
were asked to indicate their name and e-mail address on the back of the card, which enabled
us to match cards redeemed by the same member. Overall, we obtained 949 completed (i.e.,
redeemed) ten-stamp cards, recording approximately 10,000 coffee purchases. Buyback of
incomplete cards. The design of the caf RP enabled us to collect only those cards that were
completed and redeemed for a reward. Therefore, to sample from the broader member
population (i.e., including the members that would otherwise fail to complete or redeem their
card), we instituted a card buyback offer. Research assistants posing as caf employees
approached individual card-holding members and offered them the opportunity to return their
cards to the caf (regardless of the number of stamps on them) for a cash award of $4 per
card and a 1% chance to win $100. Members were told that the cards were needed for the
cafs customer research. Overall, we acquired 73 buyback cards. Recruitment of
experimental control group with transparent cards. We recruited 42 customers for an
experimental control condition in which they carried transparent cards. These cards were
similar to the regular ten-stamp card but were marked on the back so that they could not be
redeemed for a reward. The control customers were randomly sampled from the population of
program members. Research assistants (posing as caf employees) intercepted customers who
requested a regular program card and offered instead to enroll them in a purchase-habit
tudy designed to help the caf management better understand its customers. Participants were
asked to carry a transparent card and have it stamped every time they made a qualifying
purchase at the caf. They received $5 when they agreed to participate in the study, and they
were told that they would receive $15 more when they surrendered their cards six weeks
later, regardless of how many coffee purchases they made during that time. We verified that
control participants did not use the regular RP cards during the study.
RESULTS
As previously noted, the results showed that visitors of the caf purchased coffee at lesser and
lesser intervals as they neared their goal of ten coffees.

RESEARCH STUDIES
My research builds on the goal gradient hypothesis by not just verifying its validity in task-
reward situations, but also checking whether personal benefit is a must to observe goal-
gradient. For my research, I used two experiments. In the first, participants were rewarded at
the end of completing a series of tasks, while in the second there was no personal reward for
them. Both experiments are mentioned in detail below.
METHODOLOGY
I performed research on the goal gradient hypothesis using two experiments. Both
experiments are mentioned below with their procedures, participants, and results.
Experiment 1
For this experiment, I used my position as the secretary of the Editorial Board of MPSTME
to my advantage to get members of the board to participate in an experiment. Personal
rewards were set for achieving goals. The procedure and the results are mentioned below.
Procedure
16 Editorial Board members participated in the study. None of the participants were informed
of the study, or that they were even a part of it, to avoid bias. They were told that the tasks
given to them were for Verve, the college magazine.
8 members formed the test group and 8 formed the control group. Each group was given a set
of 6 articles to edit. However, these articles were not mailed across all at once. Instead, each
participant was sent a new article as soon as s/he completed the previous one and mailed it
back to me.
Participants of the test group were informed in advanced that they had to complete 6 articles.
In each mail that I sent to them with their new article to be edited, I mentioned the number of
articles already edited and the number of articles left.
Participants of the control group were never informed of the number of articles they would
have to edit. Only at the end of 6 articles were they told their task was over.
To ensure that any time variance found in the editing time was not a result of the difficulty
level of the article itself, only 6 articles were used and sent to the participants in the same
order. The difficulty level was personally modified to be more or less equal. I verified the
difficulty level with the Verve Editor Kashyap Swaroop. Note that Kashyap Swaroop was
also not informed of the experiment to avoid information leak.
The time taken by each member to complete each article in both groups was taken as the
average time for that article for that group. The mean and variances of all articles was
calculated for both groups on the basis of these average times.
An individual reward of a Rs 40 canteen coupon was set for completing the task.
Results and Analysis
The results of this experiment were interesting in their revelations. To start with, we look at
the control group, which exhibited little to no change in its speed as they ran through the
articles.
For the First two sets of articles, t < 1 and p was at 0.9. This meant that there no notable
relation at all in the amount of time spent by the participants on Articles 1 and 2. Things did
not change when we compared the last two articles, with t remaining less than 1 and p at a
very high number of 0.86. The mean did not vary much, going from 92.5 minutes to 91.9
minutes. An overall difference of less than a minute for editing the first and the last article.
Looking at the test group, we find that there is no initial relation between the times taken for
the first and second article, with t < 1 and p =0.43. Things start to change as we reach the
sixth article. T takes a significant value of 2.25 and p = 0.05. We can thus say with 95%
confidence that there is a change in the amount of time a participant took to complete the first
article and the sixth article. Indeed the mean goes down from 93.5 minutes to only 75.1
minutes as the participants realised they were nearing their goal.
This is a huge difference and is seen only in the test group. This supports the goal gradient
hypothesis and shows that as participants neared their goals, they accelerated towards it.

Experiment 2
For this experiment, I used a fund raiser started by Abhishek Malhotra in MPSTME college
for St Catherines Home. It is an institution that provides education to young girl students. A
target of Rs. 10,000 was set for each group (total target of Rs 20,000). The personal reward
mentioned in experiment 1 was eliminated.
Procedure
30 participants were selected and divided into two groups. Group 1 was the test group and
Group 2 was the control group.
Group 1 participants were informed of how much money was already contributed overall
towards the target, and the amount that was left to be collected. Note that the participants
werent actually provided real values, but randomly chosen values of Rs 1000, Rs 5000 or Rs
9000. Out of the 15 members of Group 1, 10 were provided a value of Rs 1000 collected on
Day 1, while 5 were provided a value of Rs 5000 collected on the same day. On day 3, 5 out
of the 10 members that were informed of Rs 1000 collected on day 1 were told that Rs 5000
had now been collected. The other 5 as well as all 5 who were informed of Rs 5000 collected
on day 1 were informed that Rs 9000 had been collected. This allowed us to note the change
in behaviour of members.
Group 2 was not informed of the current level of contribution. Each group was asked twice,
on days 1 and 3 of collection, to contribute money towards the charity. They were free to
choose any amount between Rs 10 and Rs 100. In total, the experiment lasted 3 days.
The money collected was, indeed, given to St Catherines Home as charity.
Participants
The participants included 30 students of MPSTME college. These students were chosen at
random based on their interest in participating in the study. The ages of these students ranged
from 20 to 22.
Results and Analysis
A Students T Test was used to analyse the results. The results were indeed a surprising
testament to the goal-gradient hypothesis.
For group 1, there was no significance difference when they were told the collection was at
its initial stage of Rs 1000 ((M = 27, SD = 8.72) and when it was at its middle stage of Rs
5000 (M = 29, SD = 4.36). We obtained t < 1 and p = 0.55. This is to say that contribution
levels did not increase significantly over the two values. However, when comparing
contributions at the Rs 5000 and Rs 9000 (M = 35, SD = 7.07) level, the change is
significant. The new t and p values were 2.28 and 0.05 respectively. This means that we can
say with 95% confidence that there is a significant increase in contributions towards the end.
Contrasted to the control group, we find they did not show any notable increase in
contribution over the two days. On day 1 we had M = 34 and SD = 7.86 compared to M = 33
and SD = 7.68 of day 3. T < 1 and p = 0.67 suggests that there is no significant increase in
contribution in a situation where the participant is not aware of how much money is left to be
collected to reach the target. Interestingly, the mean contribution is greater on both days as
compared to the contribution of the test group at Rs 1000 collection. This finding leads us to
question whether the control group is making an implicit assumption in their heads about the
money already collected and contributing accordingly, or whether informing the participants
that the money collected is quite low compared to the target changes their mind about the
amount to be donated.
This finding of the goal gradient hypothesis being supported is significant for two reasons:
One, we find that goal gradient hypothesis is not false even when there is no persona
motivation involved for the participant. Two, we find there is little increase in the initial and
middle parts of the experiment which means that there the acceleration towards the goal
occurs very late in the experiment. Further study must be performed to find out when exactly
an acceleration happens.

KEY FINDINGS AND THEIR IMPLICATIONS
We found that members in the Editorial Board experiment, when provided a personal reward
for achieving a goal, tend to accelerate their editing as they neared their goal. As the goal
distance grew shorter, they accelerated faster.
We surprisingly discovered that a personal reward is not always important for acceleration
however. Results of the Charity experiment show that members can try to accelerate towards
a goal by increasing their contributions even when they did not have a financial reward for it.
Even a reward of personal satisfaction is enough.
Results of these experiments are significant for corporations as well as NGOs.
One, they verify the importance of loyalty cards. However, we generally see that members
are not aware of their current number of points on their cards. Retailers must make it a point
to inform customers of how many points they must earn to achieve their next reward. The
literature review combined with my research also provides insight into the goal distance that
corporations must keep in order to ensure fastest acceleration towards the goal.
Two, they show that a personal reward is not always necessary. An NGO looking for funds
can use this to its benefits by informing donators of their target and their current level of
contributions. This might help them achieve their target faster when compared to not
informing people.

GENERAL DISCUSSION
The goal gradient is one of the classic phenomena discovered in the animal-learning and
behaviorism literature of the early twentieth century. It has important implications for
achievement motivation and goal pursuit but, nevertheless, has been understudied in humans.
This is particularly surprising, given that the goal-gradient hypothesis provides considerable
insights into the psychology of rewards and the optimal design of customer, employee, and
sales force incentive systems. In this research, we extended the goal gradient hypothesis to
the domain of consumer behaviour and investigated its consequences for illusionary goal
progress and customer retention.
For marketers, the goal gradient may provide profitable opportunities. In addition to
facilitating segmentation, targeting, and promotions (we discuss this subsequently), the goal
gradient may lead to a sales lift that exceeds the cost of the reward. For example, the results
of the caf RP from literature review imply that to earn one free coffee, customers bought
two more coffees than they would have otherwise. At the same time, consumers may derive
pleasure from working toward future goals. This idea is consistent with the findings of an
observational study, in which research assistants unobtrusively recorded the behavior and
affect of the caf customers. The results indicate that customers who participated in the RP
(as opposed to customers who did not) were more likely to smile when buying coffee (3.8
versus 3.4 on a five-point scale; p < .05), chat for a few minutes with caf employees (26%
versus 7%; p < .05), say thank you (95% versus 87%), and leave a tip (21% versus 3%; p <
.01). Although these results should be interpreted with caution because customers self-select
into the RP, they suggest that goal striving is intrinsically motivating beyond extrinsic
rewards.
The illusion of goal progress and its boundary conditions merit further research. Beyond its
theoretical importance, this phenomenon has substantial managerial implications for the
design of RPs and other incentive systems. Currently, many RPs award bonus points to new
members (e.g., American Express Membership Rewards Program, Hyatt Gold Passport).
Given the rich, complex structure of such programs, it is easy for managers to increase the
point requirements of rewards by an amount equivalent to the bonus, effectively creating
illusionary goal progress.
Consistent with the notion that a steeper goal gradient implies a greater drive to achieve the
reward, we found that stronger accelerators reengaged in the program faster and were more
likely to earn a second reward. Relatedly, failure to persist in the effort stream and fulfill the
requirements was associated with weaker acceleration and even deceleration. The
relationship between the goal gradient and retention was also evident in the finding that just
after reward attainment (when goal distance regressed to 100%), customers exhibited a drop
in activity (postreward resetting) and were also most likely to defect. These findings have
important implications for customer segmentation and the design of marketing interventions
aimed at reducing churn. For example, it is particularly important to communicate with and
motivate customers immediately after they earn a reward.

LIMITATIONS AND FURTHER STUDIES
During this study, we discovered that there might be a tendency for participants to get stuck
in the middle. Further studies are recommended to study this phenomena. Indeed, Andrea
Bonezzi, C. Miguel Brendl, and Matteo De Angelis have already written a paper on Stuck in
the Middle: The Psychophysics of Goal Pursuit that studies the tendency of participants to
decelerate or give up towards the middle of achieving their goal. Deeper studies must be
performed to analyse the full extent of this and its implications on marketers.
Another limitation is that I was unable to perform true market research on field because of
various outlets refusal to allow me to view their loyalty cards or interact with their customers
in-store. As a result I was unable to study the true implications of goal gradient hypothesis in
the real world on customers.
Further studies must also be done to verify two things. One, the best goal distance a marketer
must keep to ensure fastest goal completion. A very short goal distance would result in the
marketer losing money, or the reward being made so small as to be uninteresting to the
customer. At the same time, a longer goal distance would result in slower acceleration
towards the goal, resulting in more time to reach it. An optimum goal distance must be
figured, and this will vary depending on the reward itself and the customer segment. Hence
the effects of these parameters on the goal distance must be calibrated.
Two, the effect of illusionary goal must be identified. Illusionary progress does not affect the
actual goal distance. Allowing the marketer to keep a longer goal distance while still ensuring
faster acceleration. However, various questions need to be answered. Some of them are
how much progress illusion must be provided to the customer in order to ensure fastest
completion? How much progress illusion would cause the customer to question the integrity
of the marketer?

CONCLUSION
Building on the behaviorist goal-gradient hypothesis, we proposed that people working
toward future rewards would accelerate their effort as they near the reward goal. Based on a
wide range of empirical and modeling approaches, the findings we report in this article
provide converging evidence for the impact and importance of goal gradients in the human
psychology of rewards. Not only do customers accelerate toward rewards (in terms of timing,
quantity, and persistence of effort), but their acceleration also predicts loyalty and future
engagement with similar goals. On the basis of this research, we propose that the goal
gradient and its modeling have important theoretical and practical implications for
achievement motivation and goal behavior and for incentive systems and marketing
promotions.

REFERENCES
Book:
Philip Kotler (2009); Marketing Management; Pearson Education

Papers:
The Goal-Gradient Hypothesis Resurrected: Purchase Acceleration, Illusionary Goal
Progress, and Customer Retention

Others:
MOD reward programme
Personal studies