Learning and Memory
Background. Learning involves "a change in the content or organization of long term memory and/or behavior." The first part of the definition focuses on what we know (and can thus put to use) while the second focuses on concrete behavior. For example, many people will avoid foods that they consumed shortly before becoming ill. Learning is not all knowledge based. For example, we may experience the sales people in one store being nicer to us than those in the other. We thus may develop a preference for the one store over the other; however, if pressed, we may not be able to give a conscious explanation as to the reason for our preference.
Much early work on learning was actually done on rats and other animals (and much of this research was unjustifiably cruel, but that is another matter).
Classical conditioning. Pavlov’s early work on dogs was known as classical conditioning. Pavlov discovered that when dogs were fed meat powder they salivated. Pavlov then discovered that if a bell were rung before the dogs were fed, the dogs would begin salivating in anticipation of being fed (this was efficient, since they could then begin digesting the meat powder immediately). Pavlov then found that after the meat had been "paired" with the meat powder enough times, Pavlov could ring the bell without feeding the dogs and they would still salivate.
In the jargon of classical conditioning, the meat powder was an unconditioned stimulus (US) and the salivation was, when preceded by the meat powder, an unconditioned response (UR). That is, it is a biologically "hard-wired" response to salivate when you are fed. By pairing the bell with the unconditioned stimulus, the bell became a conditioned stimulus (CS) and salivation in response to the bell (with no meat powder) became a conditioned response (CR).
Many modern day advertisers use classical conditioning in some way. Consider this sequence:
Operant conditioning. Instrumental, or operant, conditioning, involves a different series of events, and this what we usually think of as learning. The general pattern is:
There are three major forms of operant learning. In positive reinforcement, an individual does something and is rewarded. He or she is then more likely to repeat the behavior. For example, you eat a candy bar (behavior), it tastes good (consequence), and you are thus more likely to eat a similar candy bar in the future (behavioral change).
Punishment is the opposite. You eat what looks like a piece of candy (behavior), only to discover that it is a piece of soap with a foul taste (consequences), and subsequently you are less likely to eat anything that looks remotely like that thing ever again (changed behavior).
It should be noted that negative reinforcement is very different from punishment. An example of negative reinforcement is an obnoxious sales person who calls you up on the phone, pressuring you into buying something you don’t want to do (aversive stimulus). You eventually agree to buy it (changed behavior), and the sales person leaves you alone (the aversive stimulus is terminated as a result of consequences of your behavior).
In general, marketers usually have relatively little power to use punishment or negative reinforcement. However, parking meters are often used to discourage consumers from taking up valuable parking space, and manufacturers may void warranties if the consumers take their product to non-authorized repair facilities.
Several factors influence the effectiveness of operant learning. In general, the closer in time the consequences are to the behavior, the more effective the learning. That is, electric utilities would be more likely to influence consumers to use less electricity at peak hours if the consumers actually had to pay when they used electricity (e.g., through a coin-slot) rather than at the end of the month. Learning is also more likely to occur when the individual can understand a relationship between behavior and consequences (but learning may occur even if this relationship is not understood consciously).
Another issue is schedules of reinforcement and extinction. Extinction occurs when behavior stops having consequences and the behavior then eventually stops occurring. For example, if a passenger learns that yelling at check-in personnel no longer gets her upgraded to first class, she will probably stop that behavior. Sometimes, an individual is rewarded every time a behavior is performed (e.g., a consumer gets a soft drink every time coins are put into a vending machine). However, it is not necessary to reward a behavior every time for learning to occur. Even if a behavior is only rewarded some of the time, the behavior may be learned. Several different schedules of reinforcement are possible:
- Fixed interval: The consumer is given a free dessert on every Tuesday when he or she eats in a particular restaurant.
- Fixed ratio: Behavior is rewarded (or punished) on every nth occasion that it is performed. (E.g., every tenth time a frequent shopper card is presented, a free product is provided).
- Variable ratio: Every time an action is performed, there is a certain percentage chance that a reward will be given. For example, every time the consumer enters the store, he or she is given a lottery ticket. With each ticket, there is a 20% chance of getting a free hamburger. The consumer may get a free hamburger twice in a row, or he or she may go ten times without getting a hamburger even once.
Variable ratio reinforcement is least vulnerable to extinction.
Sometimes, shaping may be necessary to teach the consumer the desired behavior. That is, it may be impossible to teach the consumer to directly perform the desired behavior. For example, a consumer may first get a good product for free (the product itself, if good, is a reward), then buy it with a large cents off coupon, and finally buy it at full price. Thus, we reinforce approximations of the desired behavior. Rather than introducing Coca Cola directly in Indonesia, fruit flavored soft drinks were first introduced, since these were more similar to beverages already consumed.
Vicarious learning. The consumer does not always need to go through the learning process himself or herself—sometimes it is possible to learn from observing the consequences of others. For example, stores may make a big deal out of prosecuting shop lifters not so much because they want to stop that behavior in the those caught, but rather to deter the behavior in others. Similarly, viewers may empathize with characters in advertisements who experience (usually positive) results from using a product. The Head ‘n’ Shoulders advertisement, where a poor man is rejected by women until he treats his dandruff with an effective cure, is a good example of vicarious learning.
Memory ranges in duration on a continuum from extremely short to very long term. Sensory memory includes storage of stimuli that one might not actually notice (e.g., the color of an advertisement some distance away). For slightly longer duration, when you see an ad on TV for a mail order product you might like to buy, you only keep the phone number in memory until you have dialed it. This is known as short term memory. In order for something to enter into long term memory, which is more permanent, you must usually “rehearse” it several times. For example, when you move and get a new phone number, you will probably repeat it to yourself many times. Alternatively, you get to learn your driver’s license or social security numbers with time, not because you deliberately memorize them, but instead because you encounter them numerous times as you look them up.
Several techniques can be used to enhance the memorability of information. “Chunking” involves rearranging information so that fewer parts need to be remembered. For example, consider the phone number (800) 444-1000. The eight digits can be more economically remembered as an 800 number (1 piece), four repeated 3 times (2 pieces), and 1000 (1-2 pieces). “Rehearsal” involves the consumer repeating the information over and over so that it can be remembered; this is often done so that a phone number can be remembered while the “memoree” moves to the phone to dial it. “Recirculation” involves repeated exposure to the same information; the information is not learned deliberately, but is gradually absorbed through repetition. Thus, it is to the advantage to a marketer to have an advertisement repeated extensively—especially the brand name. “Elaboration” involves the consumer thinking about the object—e.g., the product in an advertisement—and thinking about as many related issues as possible. For example, when seeing an ad for Dole bananas, the person may think of the color yellow, going to the zoo seeing a monkey eating a banana, and her grandmother’s banana-but bread. The Dole brand name may then be activated when any of those stimuli are encountered.
Memories are not always easily retrievable. This could be because the information was given lower priority than something else—e.g., we have done a lot of things since last buying a replacement furnace filter and cannot remember where this was bought last. Other times, the information can be retrieved but is not readily “available”—e.g., we will be able to remember the location of a restaurant we tried last time we were in Paris, but it may take some thinking before the information emerges.
“Spreading activation” involves the idea of one memory “triggering” another one. For example, one might think of Coke every time one remembers a favorite (and very wise) professor who frequently brought one to class. Coke might also be tied a particular supermarket that always stacked a lot of these beverages by the entrance, and to baseball where this beverage was consumed after the game. It is useful for firms to have their product be activated by as many other stimuli as possible.
There are numerous reasons why retrieval can fail or, in less fancy terms, how we come to forget. One is decay. Here, information that is not accessed frequently essentially “rusts” away. For example, we may not remember the phone number of a friend to whom we have not spoken for several months and may forget what brand of bullets an aunt prefers if we have not gone ammunition shopping with her lately. Other times, the problem may rest in interference. Proactive interference involves something we have learned interfering with what we will late later. Thus, if we remember that everyone in our family always used Tide, we may have more difficulty later remembering what other brands are available. You may be unable to remember what a new, and less important, friend’s last name is if that person shares a first name with an old friend. For example, if your best friend for many years has been Jennifer Smith, you may have difficulty remembering that your new friend Jennifer’s last name is Silverman. In retroactive interference, the problem is the reverse—learning something new blocks out something old. For example, if you once used WordPerfect than then switched to Microsoft Word, you may have trouble remembering how to use WordPerfect at a friend’s house—more so than if you had merely not used any word processing program for some time.
Memorability can be enhanced under certain conditions. One is more likely to remember favorable—or likable stimuli (all other things being equal). Salience—or the extent to which something is highly emphasized or very clearly evident—facilitates memory. Thus, a product which is very visible in an ad, and handled and given attention by the actors, will more likely be remembered. Prototypicality involves the extent to which a stimulus is a “perfect” example of a category. Therefore, people will more likely remember Coke or Kleenex than competing brands. Congruence involves the “fit” with a situation. Since memory is often reconstructed based on what seems plausible, something featured in an appropriate setting—e.g., charcoal on a porch next to a grill rather than in a garage or kitchen—is more likely to be remembered (unless the incongruence triggers an elaboration—life is complicated!) Redundancies involve showing the stimulus several times. Thus, if a given product is shown several places in a house—and if the brand name is repeated—it is more likely to be remembered.
Priming involves tying a stimulus with something so that if “that something” is encountered, the stimulus is more likely to be retrieved. Thus, for example, when one thinks of anniversaries, the Hallmark brand name is more likely to be activated. (This is a special case of spreading activation discussed earlier).
A special issue in memory are so called “scripts,” or procedures we remember for doing things. Scripts involve a series of steps for doing various things (e.g., how to send a package). In general, it is useful for firms to have their brand names incorporated into scripts (e.g., to have the consumer reflexively ask the pharmacist for Bayer rather than an unspecified brand of aspirin).
Positioning involves implementing our targeting. For example, Apple Computer has chosen to position itself as a maker of user-friendly computers. Thus, Apple has done a lot through its advertising to promote itself, through its unintimidating icons, as a computer for “non-geeks.” The Visual C software programming language, in contrast, is aimed a “techies.”
Repositioning involves an attempt to change consumer perceptions of a brand, usually because the existing position that the brand holds has become less attractive. Sears, for example, attempted to reposition itself from a place that offered great sales but unattractive prices the rest of the time to a store that consistently offered “everyday low prices.” Repositioning in practice is very difficult to accomplish. A great deal of money is often needed for advertising and other promotional efforts, and in many cases, the repositioning fails.