What Six Experiments Reveal After the Fees Arrive
This chapter tests the central subject directly: how drip pricing changes consumer choice after a lower base price has already been selected. The important comparison is not simply whether people notice a fee, but whether they change their initial choice once add-ons and a total price are visible. The experiments are useful because they separate the first choice, the decision to restart, the final choice, the price consequence, and satisfaction rather than treating “the purchase” as one indivisible act.
What the experiments actually compared
The researchers used a common purchase sequence across six studies. Participants first chose between two offers. The lower-base option charged separately for optional add-ons; the higher-base option included those same services. Participants then selected any add-ons, saw the total price, and could complete the transaction or start over. Those who restarted returned to the initial choice and repeated the process. Finally, participants reported their satisfaction with the option they ultimately kept.
The experimental difference was the timing of the surcharge information. In the non-drip condition, the prices of the lower-base option’s add-ons appeared before the initial choice. In the drip condition, participants saw the base-price comparison first and encountered those charges after selecting the lower-base option. Random assignment to these conditions makes it possible to attribute differences within each scenario to the presentation sequence, rather than to a pre-existing difference between two groups of shoppers—at least within the limits of the experimental setting.
That sequence creates several outcomes that should not be collapsed:
- Initial choice: the option selected before the add-on prices are fully available in the drip condition.
- Restart decision: whether the participant abandons the current path and begins again after seeing charges and the total.
- Ultimate choice: the option held at the end, including any participant who changed after restarting.
- Financial mistake: selecting the option that is more expensive than the alternative given the add-ons the participant chose.
- Satisfaction: how positively the participant evaluates the final choice.
The “financial mistake” measure needs particular care. Choosing a lower base price is not inherently wrong. In these experiments, however, the prices were set so that commonly selected add-ons could make the lower-base option more expensive than the alternative. The measure asks whether the participant’s final option cost more under their own selected configuration. It does not claim that everyone should choose the higher base price, or that the lower option would always be worse under a different set of add-ons.
The researchers also varied how consequential the choice was. Study 1a gave participants a budget and a bonus for spending less. Studies 2 and 4 used a similar bonus structure. Study 1c used a lottery in which one participant would actually receive the selected hotel and the value remaining in their budget. Study 1b removed the incentive-compatible budget and allowed participants to choose add-ons freely, asking what they normally did. Study 3 changed the setup to compare a participant making a choice with a participant observing someone else make the same choice; that study was designed to examine a psychological explanation rather than to repeat every step of the restart process.
The common design deliberately omits some complications found in real transactions. Participants generally faced only two options and received limited information about attributes other than price. The focal add-ons were presented and paid for during the purchase process, not after the base product had already been paid for. The surcharges were disclosed during the drip sequence rather than hidden in tiny type or buried in unrelated material. These choices make the timing comparison cleaner, but they also define the evidence’s reach.
The diagram’s two paths merge at the add-on stage, but they do not begin from the same information state. That is the experiment’s central leverage: by the time both groups see the total, one group has already made a base-price choice without the surcharge information that the other group had when comparing.
What changed, what replicated, and what it means
The clearest evidence comes from study 1a. Participants chose between a $194 base-fare airline that charged for baggage and seat selection and a $239 airline that included those services. In the drip condition, 54.5% initially selected the lower base-fare airline, compared with 11.7% in the non-drip condition. The opening numbers therefore changed the first choice substantially.
Seeing the later charges did not make the initial difference disappear. Only 15.9% of all participants decided to start over when given the opportunity. Restarting was more common among drip-exposed participants—25.2% versus 6.8%—but that did not erase the effect because most people in both conditions stayed with the current path. After restarts were included, 36.5% of the drip group ultimately selected the lower-base airline, compared with 10.2% of the non-drip group.
The price consequence followed the choice path. Given the add-ons participants selected, 24.5% of the drip group made a financial mistake by ending with the more expensive option, compared with 7.8% of the non-drip group. Satisfaction also differed: the average was 5.39 in the drip condition and 5.83 in the non-drip condition on the study’s seven-point index. The study linked this difference in satisfaction to ultimate airline choice; participants who ultimately selected the lower-base airline were less satisfied than those who selected the higher-base airline.
The first study used an unusual incentive-compatible design. Participants who avoided spending from a $300 airfare-and-add-ons budget received a small bonus, and the procedure introduced contingencies for getting clothing to the destination when someone selected no baggage. That could raise a fair question: were the results produced by the incentive or by the artificial clothing problem? Study 1b addresses much of that concern. It removed the budget incentive, let everyone freely choose or reject add-ons, and still found the same direction. Initial lower-base selection was 49.3% under drip versus 15.3% without drip. Ultimate lower-base selection was 41.8% versus 15.3%; financial mistakes were 17.9% versus 8.7%; and satisfaction averaged 5.24 versus 5.57.
Combining the first two studies produced the same pattern. Initial lower-base choice was 51.8% under drip and 13.6% under non-drip. Ultimate lower-base choice was 39.0% versus 13.0%, and financial mistakes were 21.7% versus 7.4%. The absence of a meaningful interaction between study design and surcharge presentation increases confidence that the first result was not simply an artifact of the bonus or the baggage contingency. It does not turn the rates into a forecast for all travelers; it shows that the treatment contrast survived reasonable changes in the setup.
Study 1c moves the choice from airlines to hotels and changes the participant pool and incentive. Participants selected between two descriptions of the same hotel: one room rate was $227 with charges for optional amenities, while the other was $239 with those amenities included. In a pre-test, most participants selected at least one optional amenity, making the price reversal relevant to the scenario. The drip condition produced lower-room-rate selection of 31.9%, compared with 6.5% in the non-drip condition. Because no participant restarted, initial and ultimate choices were identical. Financial mistakes were 27.7% under drip versus 6.5% without it.
The overall satisfaction main effect in the hotel study was not statistically significant: the averages were 5.73 and 5.84. That is an important qualification, not a contradiction of the broader pattern. The lower-base hotel was itself associated with lower satisfaction, and the analysis found an indirect relationship from surcharge presentation through ultimate hotel choice to satisfaction. The study therefore supports a path from disclosure timing to choice to satisfaction, while not supporting the claim that every setting produces the same simple main effect on satisfaction.
Study 2 tests whether the alternative’s price can make restarting easier. Before the initial choice, both airline prices were available, but participants might not remember the unchosen price after moving through the purchase. At the moment of the restart decision, some participants were shown the alternative airline’s base fare alongside the total for the airline they had selected. Among drip-exposed participants who initially chose the lower-base airline, 63.1% restarted when the alternative price was shown, compared with 47.0% when it was not. Their financial-mistake rate fell from 39.4% to 16.0%.
This result is informative but not perfectly isolated. Showing the comparator may have reduced the time and memory cost of searching, or it may have increased the participant’s perception of how much could be saved. The study supports the practical value of making the alternative easy to evaluate; it does not prove that search cost alone caused the change.
Study 4 returns to hotels and makes the inclusion information more explicit. The higher-base hotel was labeled as including amenities such as pool, gym, spa, breakfast, Wi-Fi, and self-parking. Participants also received information suggesting either that firms in an industry tend to charge similar surcharges or that their surcharges differ. Even with those clarifications, drip exposure increased ultimate selection of the lower-base hotel: 21.1% versus 8.6% under non-drip presentation. Financial mistakes were 11.2% versus 5.1%.
The belief manipulation changed what participants reported believing in the drip condition, but it did not by itself significantly change their choices or their likelihood of restarting. Participants who ultimately selected the lower-base hotel were more likely to believe that competitors charged similar surcharges, and that belief statistically mediated the relationship between surcharge presentation and ultimate choice. The result points to a plausible benefit-side mechanism—switching may seem unlikely to save money—while also showing that a clear warning about fee differences is not guaranteed to overcome the established choice path.
Study 3 provides a different kind of replication and a clue about process. In the self-perspective condition, participants made the airline choice themselves. In the other-perspective condition, they watched Alex make the same choices. Among people exposed to drip pricing, those who personally made the lower-base choice reported higher satisfaction with it, averaging 3.48, than those who observed Alex make the same choice, averaging 2.55. The study was not designed to measure actual restarting; its contribution is evidence that the psychological meaning of owning the decision may help sustain a favorable view of it.
Taken together, the data support three conclusions with different levels of confidence. First, within the constructed scenarios, drip exposure reliably shifted initial selection toward the lower base price and often shifted ultimate selection in the same direction. Second, because the lower base price was paired with add-on charges, this trajectory increased the chance of selecting a higher total and, in several studies, a less satisfying option. Third, the repeated result survives changes in incentives, add-on freedom, domain, and participant pool, but it remains a study-specific effect rather than a population estimate.
The comparison with the non-drip condition also matters. Non-drip participants had the add-on prices before the initial choice, not necessarily one undivided all-in total. Some still selected the option that later cost more, which is consistent with anchoring and partitioned-price effects. Upfront disclosure is therefore better than delayed disclosure in these experiments, but it is not a magic shield against inattention or arithmetic error.
Nor do these studies show that experience never helps. The authors report that prior flying experience did not moderate the drip effect in study 1a, and that repeated exposure did not fully remove negative consequences in another decision. But the experiments were not field observations of actual booking histories, and their participants were relatively inexperienced in the relevant choices. The responsible conclusion is narrower: awareness of fees and an eventual total did not reliably neutralize the first-price effect under these conditions.
That leaves the most important unanswered part of the evidence. If consumers can see the total and can sometimes restart, why do they so often remain on the original path? The next explanation is not a single hidden calculation. It is the perceived cost and benefit of switching—how much time the restart seems to require, what admitting an error feels like, and whether another seller is expected to offer a meaningfully better total.