Why Drip Pricing Still Works
A ticket site ran all-in prices and fees-at-checkout side by side on its own traffic for ten days, and the concealed version earned about a fifth more. Experience barely protected the buyers, and the two best-known books on consumer psychology do not mention the tactic at all.
Answer to the question
Why does the price go up at checkout?
Because the fee was withheld until you were nearly finished, a practice called drip pricing. In a randomised field experiment on StubHub in August 2015, revealing a 15 per cent fee only at checkout raised revenue by about a fifth against all-in prices. Roughly a seventh more people bought, and each buyer spent about 5 per cent more per ticket.
Key takeaways
- On StubHub in August 2015, revealing a 15 per cent fee only at checkout raised revenue by about a fifth against all-in prices, with roughly a seventh more buyers and about 5 per cent more spent per ticket.
- Experience barely helped: users with ten or more prior visits still showed a 15 per cent revenue effect against 21 per cent for everyone else, and buyers did not learn the correct fee level even after the site switched permanently.
- In six studies of airline and hotel booking, only 15.9 per cent of participants started over when explicitly invited to after seeing the total.
- A 2023 report for the UK Department for Business and Trade found just under half of 525 online providers dripped at least one fee excluding delivery charges, costing consumers an estimated £595 million to £3.5 billion a year.
You pick the cheap seats, click through two screens, and the total at checkout is nine pounds higher than the price you chose from. That is drip pricing, the practice of advertising part of a price and withholding the rest until you are nearly finished, and our position is that it works, and that it works on people who know perfectly well it is coming. Drip pricing is not a trick that stops working once you have seen it.
The ticket site that ran two prices at once
Between 19 and 31 August 2015, StubHub showed two pricing regimes to its own traffic at once. Roughly half of several million users, split by cookie, saw all-in prices with a notice that fees were included. The other half saw lower prices with the fees revealed at checkout. Tom Blake, then at eBay Research, published the experiment with three co-authors in Marketing Science in 2021.
The design is why it matters. Nobody had to guess what shoppers would have done under the other regime, because the other regime was running in the next browser along. The site had been an all-in seller and moved everyone to fees-at-checkout on 1 September 2015, days after the experiment closed.
What the fees-at-checkout group did is more interesting than what they spent. They scrolled through listings about 10 per cent more often and were 56 per cent more likely to compare several listings. Having reached the final review page they were 19 per cent more likely to pick a ticket and 45 per cent less likely to go through with it. They also bought better seats, the effect fading as the rows ran down the alphabet.
> Shoppers meeting the fee late searched harder, aimed higher, and bailed out more at the last screen.
Shoppers meeting the fee late searched harder, aimed higher, and bailed out more at the last screen. The ones who stayed were worth about a fifth more.
One disclosure the paper prints and so do we: the fourth author worked and consulted for eBay from August 2011 to February 2016, the research was done after that period and was not funded by eBay, and StubHub supplied the data.
The books that never mention it
Before the numbers, an absence. The two best-known practitioner books on consumer psychology in this vault, Richard Shotton's The Choice Factory and The Illusion of Choice, between them cover drip pricing once. A search of the complete reading notes for drip, partition, surcharge, booking fee and hidden fee returns a single hit in 190 pages, an uncited 2000 survey about how card surcharges are worded. The most heavily regulated pricing tactic of the past decade is not in them, and that gap in the consumer-behaviour shelf is itself worth reporting.
The vault book that does name drip pricing is Nudge, at page 146, and it names both halves of the pair.
> Drip pricing is "to advertise only part of a price up front and withhold other parts until customers move toward finalizing their choices (or even until they check out)". Partitioned pricing is "to divide a price into components without disclosing the total." Both, Thaler and Sunstein write, exploit consumer inattention.
That distinction does real work for us, because we have already written about the other half and ruled this mechanism out there. A free delivery threshold announces the fee, the threshold and the saving, usually in a coloured banner, so nothing is concealed and shrouding is the wrong frame for it. Drip pricing conceals. Xavier Gabaix and David Laibson's 2006 account of shrouded attributes, where a headline price understates true cost because the add-ons are hard to discover, is the right frame here for exactly the reason it was the wrong one there.
Why nobody starts over
Shelle Santana, Steven Dallas and Vicki Morwitz ran six studies of airline and hotel booking in 2020, four with real consequences for the chooser, and every one let people start again once the total appeared. Samples ran from 93 members of a business-school subject pool to 805 online workers.
Almost nobody started again. In the first study only 15.9 per cent restarted when invited to: a quarter of those who had met dripped fees, against fewer than one in fourteen of those who had not. The dripped group mostly kept the option with the lower headline price and the higher total, and reported being less satisfied with it.
Three mechanisms are tied to three separate studies: people expect searching again to cost too much, they justify the choice already made, and they wrongly believe every rival charges much the same fees. That last is the quietest and the most useful, because the belief that everyone does it is part of why it pays. The same checkout that hides a fee on the way in often makes a subscription hard to leave on the way out.
The term was coined earlier. Vicki Morwitz, with Eric Greenleaf and Eric Johnson, named partitioned pricing in the Journal of Marketing Research in 1998, and their abstract carries an admission: firms presumably use it to raise demand, but there was then "little clear empirical support" that it did, and no theory of why it should. They measured recalled cost and purchase intention, not purchases. Twenty-three years separate that sentence from a field experiment on a live ticket site.
Where splitting a price backfires
The strongest case against is not a failed replication. It is a published mechanism pointing the other way. Marco Bertini and Luc Wathieu argued in 2008, in the same journal as the StubHub paper, that splitting a price draws attention to whatever each component is tagged to. Across four experiments, when the tagged extra was valuable and easy to judge, partitioning helped demand; when it was not, it hurt. The tactic is not one-directional and they theorised the conditions in advance.
It also has a ceiling. Tanjim Hossain and John Morgan sold matched pairs of CDs and video games on eBay and found that a low opening price with a high shipping charge drew more bidders and more revenue, but only when the shipping charge was not excessive. Their own explanation is mental accounting, shoppers keeping separate books for the two parts of the price, which is a different variable again from how the money leaves your hand at the till.
Alexander Rasch and colleagues built the cleanest comparison in 2020: laboratory markets of two sellers and one buyer, 20 rounds, 54 subjects in each of four treatments, where buyers had to pay to uncover the dripped part. One treatment was a drip pricing ban. Consumer surplus was lower and seller profits higher when drip pricing was allowed. The subtler result is what shoppers got wrong. With a predictable maximum fee the headline price still ranked the offers correctly, so people overpaid without picking the wrong seller. Make the fee size uncertain and they picked the wrong seller too.
Both of those cases are strong, and we think the resolution sits in the boundary rather than the verdict. Bertini and Wathieu are right that a split price can repel a buyer when the component is worthless or unreadable. Drip pricing announces no component at all, which is how it sidesteps their mechanism.
### What the fee was worth
Now the number. Obscuring a 15 per cent fee raised StubHub's revenue by about a fifth, which read backwards is what a seller gives up by quoting an honest total. Most of that came from more people buying, around a seventh more over the ten days, and about a quarter of it from each buyer spending about 5 per cent more per ticket by moving up the rows. A separate 2012 experiment on the same platform, randomised across 99 football fixtures rather than across users, pointed the same way.
Experience helped, and not much. Among users with ten or more previous visits the revenue effect was 15 per cent rather than 21 per cent. It did not vanish, and after the site moved everyone to fees-at-checkout the authors found that users still did not learn to anticipate the right fee level. Nor did they leave: the fees-at-checkout group were 3.3 per cent less likely to churn over the following year, which is the opposite of what an anger story predicts, and the authors flag the confound and leave it unresolved.
The scale beyond one platform comes from a 2023 report for the UK Department for Business and Trade, which bought something from each of 525 online and app providers. Just under half dripped at least one fee once delivery charges were set aside, rising to 72 per cent in transport and communication and falling to 15 per cent in retail, and fees met more than one of the report's five harm criteria at 41 per cent of providers. Its estimate of the annual cost to UK consumers runs from £595 million to £3.5 billion, a range wide enough to be honest about.
What this evidence cannot establish
Delivery fees are excluded from those figures, deliberately, because they are the commonest dripped fee, shoppers expect them, and an optional one can be worth having. Every number above sits on that exclusion.
The revenue result cannot be generalised. It is one platform, one product, a percentage fee and ten days in August 2015, and the percentage structure is what created the upgrade to better seats. A flat fee would not produce it. The authors read the switch as expanding the market rather than taking share, on the evidence of search-engine interest in two competitors, which is a proxy and not sales data.
Nor does any of this settle welfare. Blake and his co-authors decline a verdict and note that if concealment raises what sellers earn, more and better sellers may list, which complicates the sum. Regulators assert consumer harm; the field experiment measured revenue, and we are not merging those two claims.
A price you cannot see is a price you cannot compare.
The whole tactic lives in that gap.
Common questions
Why does the price go up at checkout?
Because the seller gains by holding part of it back until you are close to finishing. A randomised experiment on StubHub in August 2015, published in Marketing Science in 2021, found that revealing a 15 per cent fee only at checkout raised revenue by about a fifth against all-in prices, with roughly a seventh more people buying and each buyer spending about 5 per cent more per ticket.
Why do people not start the booking again?
Because almost nobody does. In Shelle Santana and colleagues' 2020 studies, only 15.9 per cent restarted when explicitly invited to after seeing the total. The researchers tie that to expected search costs, to justifying a choice already made, and to a mistaken belief that competing sellers charge similar fees.
How common are hidden fees online?
A 2023 report for the UK Department for Business and Trade bought something from 525 online and app providers and found just under half dripped at least one fee, excluding delivery charges. Transport and communication was highest at 72 per cent, retail lowest at 15 per cent, and the estimated annual cost to UK consumers is £595 million to £3.5 billion.
Does drip pricing stop working once you know about it?
Not entirely. StubHub users with ten or more prior visits still gave up about 15 per cent more revenue under fees-at-checkout than under all-in pricing, against 21 per cent for everyone else, and after the site switched permanently the researchers found buyers still failed to anticipate the correct fee level.
Sources
Blake, T., Moshary, S., Sweeney, K., & Tadelis, S. (2021). 'Price Salience and Product Choice.' Marketing Science, 40(4), 619-636. doi:10.1287/mksc.2020.1261. [Full text read, the accepted version at faculty.haas.berkeley.edu, plus the published Crossref record and the NBER w25186 abstract. Disclosure as the paper prints it: Tadelis worked and consulted for eBay from August 2011 to February 2016, the research postdates that period and was not funded by eBay, and StubHub supplied the data]
Morwitz, V. G., Greenleaf, E. A., & Johnson, E. J. (1998). 'Divide and Prosper: Consumers' Reactions to Partitioned Prices.' Journal of Marketing Research, 35(4), 453-463. doi:10.1177/002224379803500404. [Abstract read in full; the full text was not reached and per-study sample sizes are NOT VERIFIED]
Santana, S., Dallas, S. K., & Morwitz, V. G. (2020). 'Consumer Reactions to Drip Pricing.' Marketing Science, 39(1), 188-210. doi:10.1287/mksc.2019.1207. [Full abstract and methods sections read, from the author copy at business.columbia.edu, plus the Crossref record]
Bertini, M., & Wathieu, L. (2008). 'Research Note: Attention Arousal Through Price Partitioning.' Marketing Science, 27(2), 236-246. doi:10.1287/mksc.1070.0295. [Abstract read in full; the four sample sizes were not reached and are NOT VERIFIED]
Hossain, T., & Morgan, J. (2006). '…Plus Shipping and Handling: Revenue (Non)Equivalence in Field Experiments on eBay.' The B.E. Journal of Economic Analysis & Policy, 6(2). doi:10.2202/1538-0637.1429. [Abstract read in full; the number of auction pairs was not reached and is NOT VERIFIED]
Gabaix, X., & Laibson, D. (2006). 'Shrouded Attributes, Consumer Myopia, and Information Suppression in Competitive Markets.' The Quarterly Journal of Economics, 121(2), 505-540. doi:10.1162/qjec.2006.121.2.505. [Crossref item record read in full including the 60-entry reference list; no abstract deposited and the article itself was not reached]
Rasch, A., Thöne, M., & Wenzel, T. (2020). 'Drip pricing and its regulation: Experimental evidence.' Journal of Economic Behavior & Organization, 176, 353-370. doi:10.1016/j.jebo.2020.04.007. [Full abstract read, from the accepted version at eprints.whiterose.ac.uk, CC-BY-NC-ND]
Alma Economics for the UK Department for Business and Trade (2023). 'Estimating the prevalence and impact of online drip pricing.' Published 4 September 2023. [Executive summary and methodology read at source, assets.publishing.service.gov.uk attachment 1182208. Delivery fees are excluded from every figure, by the report's own choice]
Thaler, R. H., & Sunstein, C. R. Nudge: The Final Edition, pp. 145 to 146 and pp. 160 to 161. [Read in the vault's reading digests; the two definitions at p. 146 carry no citation and are the authors' own statement of practice]
Shotton, R. (2023). The Illusion of Choice. Harriman House. [Read in the vault, audited from 110 photographs across three batches; a search of the complete reading digests for drip, partition, surcharge, booking fee and hidden fee returns one hit in 190 pages]
Shotton, R. (2018). The Choice Factory. Harriman House. [Read in the vault; the same search across the complete digests of all 25 biases returns no hits]