What kind of shopper are you? Do you
always buy the most popular items? Do you look to see what your
friends are buying? Do you look at the newspaper reviews before
buying a new book?
Questions like these lie at the heart
of a new set of techniques that are just starting to be applied to
e-commerce, lifestyle changes and almost any facet of your
Such "persuasion profiles" differ from
the personal recommendations we know and, sometimes love, from
Amazon and iTunes, which are based on purchase patterns and
behaviour on site, in that they are based on an understanding of
our psychology rather than the relevance of a particular book or
They are also a step beyond online
behavioural targeting in that they detail not what we have been
looking at but rather what messages we are likely to respond
What "persuasion profiles" potentially
tell brands and governments is whether you are more likely to
respond to a message based on authority, whether you are more
likely to respond to scarcity or whether you are more likely to
respond positively to someone you like?
They can be applied across all aspects
of your behaviour and ultimately could become tradable data valued
on the basis of their ability to boost conversion rates and improve
take-up of other changes desired by governments, service providers
In essence, they are the digital
equivalent of the salesperson's skill in sizing up a customer and
working out what message about his or her product is most likely to
convince them to buy. But while their use on a one to one basis is
as old as the human race, their widespread adoption and automated
inclusion in communications raises a number of wider issues.
With us now
The use of psychological strategies is
already part of the world we live in. They are used in DirectLife,
Philips's health and weight management tool, which guides consumers
towards a healthier lifestyle.
They can be seen in apps like MyZEO,
which is designed to help people sleep better. Devices and tools
such as these rely on what are called influences strategies, which
are designed to increase their effectiveness.
Some researchers specify more than 100
influence strategies or ways to convince people to adopt a certain
pattern of behaviour but there are six core principles of
6 core principles of
People feel obligated to return a
favour, thus when a persuasive request is made by a person the
receiver is in debt to, the receiver is more inclined to adhere to
When something is scarce, people will
value it more. Announcing that a product or service is scarce will
favor the evaluation and increase the chance of purchase.
When a request or statement is made by
a legitimate authority, people are more inclined to comply or find
the information credible.
People do as they said they would.
People try to be consistent with previous or reported behaviour,
resolving cognitive dissonance by changing their attitudes or
behaviours to achieve consistency. If a persuasive request aligns
with previous behaviour people are more inclined to comply.
People do as other people do. When a
persuasive request is made people are more inclined to comply when
they are aware that others have complied as well.
We say "yes" to people we like. When a
request is made by someone we like, we are more inclined to act
In fact, each of these six influence
strategies can be seen in an e-commerce domain, so reciprocity can
be delivered when online stores offer a small gift to consumers;
scarcity includes limited time offers; authority is
recommendations; commitment strategies could include a wish list;
consensus would include showing how many other people had also
bought these products in question; while liking could be the use of
social media to promote what your friends had bought. While such
techniques are a common part of the e-commerce palette, what's new
is the application of them to you as an individual. So that the
sales tactic used to make you buy is intentionally different from
the pitch I get and based on your (and my) previous behaviour.
Having developed the concept of
persuasion profiles with Dean Eckles at Stanford University, I
wanted to see how they might be applied in a real world context. In
partnership with a new Dutch children's clothing store,
kinder-kleetjes. com, I tested two of these influence strategies
while monitoring store performance before and during our
Kinder-kleertjes.com offers a
selection of more than 1.200 products, via two affiliate programmes
and the website aims to attract traffic via search. Running since
July 2010, it is a small site with an average of nearly 400
visitors each month during the six-month experiment. I monitored
click-throughs and average purchase per visitor from July until
October 21 to provide a baseline for our month-long experiment.
The homepage of the online store
presents a random collection of 40 products together with pictures
and a single sentence description. Once a visitor clicks on one of
the products (or enters the site using a search term directly
pointing at a product page) a product is displayed which shows a
large image and a textual description of the product.
We decided to offer a three-strategy
choice - no influence strategy, scarcity ("special offer") and
consensus ("best seller"). Every other factor including price
Consumers who clicked through would
either see no text or a message that read "This clothing item is
available today at a special discount rate" for the scarcity
strategy and "This is one of our best-selling clothing items" to
appeal to consensus seekers.
Consumers would get a particular
message, initially most often based on what was proving most
successful, but as users looked at multiple items we were able to
apply learnings about them as individuals to see if we could
We attracted 831 unique visitors
during our trial and, while our adaptive persuasion took time to
learn what might work best, both the scarcity and consensus
strategies significantly outperformed the "No Strategy"
implementation within two weeks.
During our baseline period, 14,4% of
the users of the site eventually clicked on one of the products and
were taken to the vendor's home page. With our adaptive persuasion
techniques in use this increased to 18,3%.
We've since repeated the experiment
using only half the customer base as our test and the other half as
our base level. The aim was to remove any element of seasonality
from our results. Once again click-through rates increased, up from
9,4% to 13,5% using our adaptive persuasion algorithm.
Because such profiles and strategies
can deliver improvements in performance, they clearly have a value
for anyone who seeks to change consumer behaviour. Companies or
organisations that we visit infrequently might rely on probability
tools such as algorithms to decide which strategy to apply, but for
a retailer such as Amazon that we might visit frequently, it could
be worth building up a complex individual profile.
And, unlike recommendations for books
and music, persuasion profiles could be applied across product
sectors. For example, the proﬁle constructed from observing a
user's online shopping behaviour could be of use in increasing
compliance in saving energy.
Not only could persuasion proﬁles be
used across different contexts within a single organisation but
there is the option of exchanging the persuasion proﬁles between
corporations, governments, other institutions, and individuals.
As with other consumer data, it's
possible to envisage a market for these profiles developing subject
to existing constraints on data usage. Such potential transfers
create ethical issues because, once constructed, the proﬁles can be
used for ends not anticipated by its designers.
Then there's the issue of ownership.
Do individuals have access to their complete persuasion proﬁles or
other indicators of the contents of the proﬁles? Are individuals
compensated for this valuable information?
What if an individual wants to use the
hypothetical persuasion proﬁle created by Amazon to jump-start and
improve the effectiveness of a mobile exercise coach, would they be
able to obtain and transfer this proﬁle?
It's early days and the answers
haven't been resolved.
Maurits Kaptein is a Ph.D.
candidate at the Technical University of Eindhoven in The
THE VIEWS AND OPINIONS
EXPRESSED IN THIS DOCUMENT ARE THOSE OF THE AUTHOR.