Stefano Bussolon Interactions and Decision Making

Designing Interactions that Help Customers in Decision Making

Stefano Bussolon

Click F11 for full screen, and to move, home to go to the first, end to go to the last.
About me
Stefano Bussolon portrait

Stefano Bussolon


Ph.D. in Cognitive Sciences

Contract professor in Data Analysis

Information architect



@sweetdreamerit on Twitter

Web site:


Google plus

Would you eat those mushrooms?

Yes - No - Maybe
Do you need more information?

We all are informavors

Every animal needs some information to survive.
But gathering information has a cost (information foraging)

The goal of InfoArc (first hypothesis)

The goal of information architecture is to maximize the quantity and quality of information, and to minimize the cost and effort to find it.

Information is not knowledge

Information is not knowledge,
Knowledge is not wisdom,
Wisdom is not truth,
Truth is not beauty,
Beauty is not love,
Love is not music,
and Music is the best.
Frank Zappa, "Packard Goose"

Knowledge and decisions

Information becomes knowledge if it helps an agent to take a decision.


Finding the information is the first step in decision making.
Making a choice is the second step.
Information architecture should help user not just to find things, but also to make the right choice. From findability to choosability

Let's buy some chocolate

You fancy some chocolate, so you decide to go and buy it.
Would you prefer to go to a shop with a huge number of options, or just few types of chocolate?

Less products

Eliminating the low selling products (up to 54%)

  • same perception of variety
  • increased sales (+10%)

Is this the end of the long tail?

Users as decision makers

Chernev 2003 distinguishes 3 types of consumers

  • Consumer A already chose what she wants, and she just needs to find it out.
  • Consumer B knows her preferences, but she need to make the choice.
  • Consumer C is not aware of which attributes are important in the decision process.
image/svg+xml Pick the product Identify the attributes Get suggestions A B C weight every attribute for every alternative calcolate the value choose the best
Less is more is more

The consumer A, who already knows what she is looking for, is more satisfied of her purchase if she can choose from a greater assortment of products.

Consumers B and C prefer to choose among a more limited selection of products.

How can we interpret this seemingly paradoxical outcome?

The effort of choice

A greater assortment is considered, per se, positive.

Choosing from many alternatives is a demanding cognitive task

Effort vs accuracy
image/svg+xml Accuracy Effort Random Rational(ist) Heuristics Irrational External cognition
Rationalist choice: weighted sum
  • identify all the pertinent attributes
  • assess a weight (importance) to every attribute
  • calculate the weighted sum for every alternative
  • choose the alternative with the higher weighted sum
Bounding rationality to the world
  • identify all the attributes: designer, experts, other users.
  • assess the weights: the user (with designed defaults)
  • calculate the expected value for every alternative: the application
  • sort the alternatives by their value: the application
Heuristics: a less effort approach

The goal of an heuristic is to improve the ratio between decision accuracy and costs (time, memory, cognition, computation)

Ecological rationality

Effort-reduction methods
  • Examining fewer attributes
  • Reducing the cost of retrieving and storing cue values
  • Simplifying the weighting process
  • Integrating less information
  • Examining fewer alternatives
Elimination by Aspects (EBA)

Identify a cut-off for every attribute, and eliminate each alternative that has an attribute that is outside the cut-off

Elimination by Aspects: example

Ebay offers a filter that works in a similar way to the Elimination by Aspects: it allows the user to decide a upper and lower cut-off for some attributes, and filters out the alternatives outside.

Majority of Confirming Dimensions (MCD)

Compare a couple of alternatives for all attributes, keep the best one, and compare it to another alternative.

Majority of Confirming Dimensions: example

I have no knowledge of the implementation of this heuristic. Users often use tabbed browsing to perform this strategy: open a new tab, compare the two alternatives, close the tab with the poorer one.

Satisficing Heuristic (SAT)

Take the first satisfactory alternative

Satisficing: example

Google can be seen as an example of satisficing: we don't process all the 217.000 results, but pick the first that satisfies our needs.

Lexicographic Heuristic (LEX)

The Lexicographic procedure determines the most important attribute and then sorts the alternatives on that attribute.

Lexicographic: example

Venere uses, among others, a lexicographic interface: it allows the user to sort the results by best sellers, price, stars, guest rating.

Equal Weight Heuristic (EQW)

all attributes have equal weight

Equal Weight: example

Venere calculates an overall Guest Rating based on the mean rates of the different attributes (Cleanliness, Quietness, Spaciousness, Service, Surroundings)

A user: direct path

The user A wants to go straight to the product she wants, but likes to know she is choosing among a great selection.
A good search engine would be the good tool for her.

B user: narrow the choice

The B user has a (clear) representation of the pertinent dimensions.

She needs to narrow the choice to the item that better fulfills her preferences.


The B user can be happy to use a faceted navigation and selection interface.

But see Peter Boersma, 2010 for its limits.

Discover the dimensions

Help the C user to discover the pertinent dimensions


Whenever you can, categorize the items

The mere categorization effect: people are more satisfied by their choice when the alternatives are presented in categories, even if those are arbitrary and meaningless.

Avoid the decision

Allow the C user to avoid the decision process.

Example: suggestions

Amazon gives the users some suggestions: the customer can just pick one of this, without carrying the process of decision making.

How to improve choosability
  • Facilitate both the "rational" and the heuristic decision strategies.
  • Divide the process of decision making
  • Design for the different users (A, B, C)
  • Give the users some external aid (external cognition, suggestions)
  • Categorize
Thank you

Information is not knowledge, knowledge is not tango.
And tango is the best. - @sweetdreamerit