Week 07
Consumer Behavior · Davood Wadi, PhD
A purchase is the last step of a decision that begins long before the checkout.
Consumer decision making is the process by which people identify a need, gather information, compare options, and choose.
Some decisions take weeks of research, such as choosing a university or a car.
Other decisions take less than a second, such as picking up the usual brand of milk.
The same consumer can make careful decisions in one category and habitual decisions in another.
Today, a consumer can also hand part of the process to an AI agent that searches and compares on the consumer's behalf.
The traditional model describes decision making as five stages.
Stage 1 is problem recognition. The consumer notices a gap between the current state and a desired state.
Stage 2 is information search. The consumer looks for information about possible solutions.
Stage 3 is evaluation of alternatives. The consumer compares the options on the attributes that matter.
Stage 4 is product choice. The consumer selects one option and buys it.
Stage 5 is postpurchase evaluation. The consumer uses the product and judges whether it met expectations.
The model assumes a rational consumer. Real consumers often skip stages, repeat them, or move through them in a different order.
Follow Ines's laptop purchase through the five stages of the decision process. Put her moments in the order they happen.
Tap a step, then the number where it belongs.
Problem recognition occurs when the consumer sees a significant difference between the actual state and the ideal state.
Need recognition happens when the actual state falls. A phone breaks, or the fridge is empty. Marketers can trigger need recognition by reminding consumers that a product runs out or wears out.
Opportunity recognition happens when the ideal state rises. A consumer sees a new phone with a better camera and wants it. Marketers can trigger opportunity recognition by showing a better state that the consumer had not imagined.
Each card describes a consumer who has noticed a gap between the actual state and the ideal state. Did the actual state fall (need recognition) or did the ideal state rise (opportunity recognition)?
Tap a card, then tap the group it belongs to.
Consumers do not put the same effort into every decision.
Routine response behavior is automatic. The consumer buys the same product with little or no search, such as a favorite toothpaste.
Limited problem solving uses some search and simple rules. The consumer compares a few options, such as when choosing a new cereal.
Extended problem solving uses extensive search and careful comparison. It occurs when the purchase is expensive, risky, or rare, such as buying a house.
The level of effort depends on involvement, perceived risk, and past experience with the category.
Each case describes one purchase. Which level of decision making is it?
Routine response: automatic, with little or no search.
Limited problem solving: some search and a simple rule across a few options.
Extended problem solving: extensive search and careful comparison for a costly, rare purchase.
Routine response: she buys the same product with no search.
Limited problem solving: she compares a few options with a simple rule.
Extended problem solving: the purchase is expensive and risky, so effort is high.
Which concept does this case show?
Internal search means looking into memory for what the consumer already knows about the options.
External search means gathering new information from ads, websites, stores, reviews, friends, and sales staff.
Consumers search more when the purchase is important, the risk is high, and they know little about the category.
Consumers search less when they are experts, when they are satisfied with their current brand, or when time is short.
Search has costs. Each extra comparison takes time, money, and mental effort, so consumers stop searching before they have seen every option.
Each card describes how a consumer looks for information. Is it internal search or external search?
Tap a card, then tap the group it belongs to.
Perceived risk is the belief that a purchase may have negative consequences.
Monetary risk is the chance of losing money on a poor choice.
Functional risk is the chance that the product will not perform as expected.
Physical risk is the chance of harm to the body or health.
Social risk is the chance that others will judge the consumer negatively.
Psychological risk is the chance that the choice will damage the consumer's self-image.
The higher the perceived risk, the more information consumers tend to gather before they buy.
Consumers can now delegate information search to an AI agent.
The consumer gives the agent a goal. The agent looks up product details, compares options, and makes a choice.
Like a human shopper, the agent often cannot see every detail at once. It has to look up each attribute, and each look-up can cost time or money.
Researchers can trace which details an agent looks up before it chooses, as consumer researchers have long traced which information human shoppers open before they choose (Wadi & Ma, 2026b).
When looking up information was free, AI agents gathered the details they needed and rarely chose a worse product (Wadi & Ma, 2026b).
When each look-up had a cost, agents stopped searching earlier and sometimes chose without the details they needed (Wadi & Ma, 2026b).
The quality of a delegated decision therefore depends on the information the agent gathers, not only on how well it reasons.
Dana asks an AI agent to find her running shoes. Each time the agent opens a product page, it is charged a small fee. Which outcome fits what the slide describes?
The instruction the consumer gives the agent defines what a good choice means.
A vague goal leaves the criterion open, for example "find the best deal on instant coffee."
A specific goal states the criterion, for example "find the coffee with the lowest price per ounce."
In one study, one coffee cost $4.99 for 10 ounces and another cost $5.00 for 11 ounces. The $5.00 coffee was cheaper per ounce (Wadi & Ma, 2026b).
When look-ups were costly and the goal was vague, AI agents often skipped the package weight and more often chose the $4.99 coffee (Wadi & Ma, 2026b).
When the goal was specific, the agents kept looking up the details needed to compare price per ounce and chose better (Wadi & Ma, 2026b).
A consumer who states the criterion that defines a good choice gives the agent a clear objective. An agent cannot pursue an objective the consumer did not state.
Each card is an instruction a consumer gives to an AI agent. Does it leave the criterion open (vague goal) or state it (specific goal)?
Tap a card, then tap the group it belongs to.
Consumers rarely evaluate every brand in a category.
The awareness set includes all the brands a consumer knows.
The evoked set includes the brands a consumer recalls from memory when making a decision.
The consideration set includes the brands the consumer seriously considers buying.
The inept set includes brands the consumer rejects, and the inert set includes brands the consumer feels indifferent toward.
A brand that does not enter the consideration set has almost no chance of being chosen. When an AI agent searches, the list it returns can become the consumer's consideration set.
Chen is about to buy coffee. Match each set of brands with the part of her thinking it describes.
Set
Chen's coffee brands
Tap a term, then the example it goes with.
Decision rules are the strategies consumers use to choose among alternatives.
Compensatory rules let a strength on one attribute make up for a weakness on another. A laptop with a weak battery can still win if its screen and price are excellent.
Noncompensatory rules do not allow this trade-off. A weakness on one important attribute rules the option out.
The lexicographic rule chooses the brand that is best on the most important attribute. If there is a tie, the consumer moves to the next attribute.
The elimination-by-aspects rule removes every option that fails a cut-off on an important attribute, such as any flight with more than one stop.
The conjunctive rule sets a minimum level for every attribute and rejects any option that falls below one of them.
Consumers often use noncompensatory rules to narrow the options and then a compensatory rule to choose among the few that remain.
Each case describes how a consumer chooses. Which decision rule is it?
Lexicographic: best on the most important attribute, with ties settled on the next attribute.
Elimination by aspects: options that fail a cut-off on an important attribute are removed, one attribute at a time.
Conjunctive: a minimum level on every attribute, and falling below one rejects the option.
Compensatory: a strength on some attributes makes up for a weakness on another.
Lexicographic: he chooses the option that is best on the most important attribute.
Compensatory: a high score on one attribute can offset a low score on another.
Which concept does this case show?
Think of a product you plan to buy soon. Write the instruction you would give an AI agent to buy it for you. Which attributes and decision rule does your instruction state, and which does it leave for the agent to decide?