Memory can be a funny thing. Sometimes we recall seeing things we’ve never actually seen. Other times, we complain that we have a terrible memory when we can’t remember a name or some piece trivia that we definitely knew at some point. It is no wonder, considering the vast amounts of data and information we are subjected to on a daily basis. Not to mention the use of technology – including social media – has made us forgetful.
Now, imagine you find yourself in an electronics shop, about to buy a laptop. You’ve spent some time comparing the characteristics of several computers. In fact, your brain is filled to the brim with specifications. Laptop A is a strong contender since it has great processing power and boasts high screen resolution. However, you can’t quite recall if it has a long battery life.
If you don’t have the facts, are you more likely to conclude that high processing power means that battery life will drain faster, or will you just assume that since the laptop has excellent specifications overall, it probably also has a great battery? According to our research, it depends. If you believe you have forgotten the facts about the battery life, you are more likely to assume it is long, based on your overall positive impression of the laptop. However, if you think you’ve never seen this specification, you are more likely to use logic and conclude that a laptop with high processing power typically drains its battery quickly.
In our recent paper, Hal E. Hershfield and I showed how consumers who are uncertain about an aspect of a product can suspend logic when they think their own memory is at fault. Across a field experiment, three multiday studies and five lab experiments, we found that consumers who believe (or were made to believe) that they have forgotten some product information often extend their overall impression of the product to that missing information, even when it goes against logical deduction.Perception is key
In one lab experiment, we showed participants a review of either a high-quality or a low-quality bicycle. Both reviews provided scores for a great many product attributes. Half of the participants were randomly selected to be shown the number of retail locations servicing the bike (the forgotten condition) and half were not (the unknown condition).
A day later, the participants in the forgotten condition admitted that they had forgotten the number of service locations and the participants in the unknown condition agreed they had not seen it. Participants were also told that that higher quality bikes tended to have fewer retail locations. In the unknown condition, participants used this logical evidence and said that the high-quality bike had fewer service locations (evaluated as 38.1 out of 100) than the low-quality bike (68.9/100).
In contrast, in the forgotten condition, participants ignored the logical evidence and misremembered that the high-quality bike had more service locations (57.6/100) and the low-quality bike had fewer (47.4/100).
In another experiment, we let a week elapse instead of a day. Predictably, participants’ memories of the bicycle characteristics were even fuzzier. They then relied even more on their overall impression of the bicycle when trying to estimate the score of one of its given attributes, despite the fact that they had viewed correlational evidence to the contrary (e.g. a lightweight bicycle would be less durable, all else being equal).
Our results held across various fields. We had other participants read detailed reviews of two hotels. Both establishments were rated on a scale of 1 to 5 on about 20 characteristics (e.g. location, concierge service). After just a short break, the majority of participants had forgotten the exact value of a given attribute. However, they inferred it to be similar to their overall impression of the hotel, even though we set the attribute to be the opposite of the other hotel attributes. We obtained similar results in other studies involving high- and low-performance smartphones and televisions, even when participants viewed evidence that should have prompted them to make guesses based on logic, not general impressions.
Our final experiment allowed us to generalise our lab findings to a field setting with consumers making real choices. Over three cold days, we recruited 182 passers-by at multiple outdoor locations in Paris by telling them they would receive a pair of hand warmers (“hot packs”) in exchange for their participation in a short study.
Using a tablet, participants reviewed the attributes of a high-heat warmer that could last 8 hours or a low-heat warmer that could last 12 hours. We then asked them to memorise a long random number to crowd out their short-term memory.
Then, we successfully persuaded half of the participants that they had never seen any information about the heating life of the warmers. The other half readily admitted to having forgotten it. The participants who believed they had forgotten (versus never seen) the attribute estimated that the high-temperature warmer also had the benefit of having a longer heating life and were 25 percent more likely to select it as their gift.
Retrieving memories “crowds out” attention for existing evidence
The amount of product information has increased tremendously in recent decades, but our memory capacity has largely remained the same. As consumers, we are more likely than ever before to learn and forget information.
Even if some memory failures can easily be resolved by an online search, many others cannot. Maybe the information came from someone we can no longer contact or from a site that’s now inaccessible. Many consumers may not feel the need to look up forgotten information before purchasing a small item. In fact, they may find it impractical to look up every bit of information when considering a large array of products.
For a marketer’s perspective, if a product has a large number of attributes, making it very likely that the consumers will forget most of them, it may be more critical to create a positive overall impression of the advertised product, rather than highlighting the worth of each attribute value.Daniel Walters is an Assistant Professor of Marketing at INSEAD.
[This article is republished courtesy of INSEAD Knowledge
http://knowledge.insead.edu Copyright INSEAD 2010]