Partitioned pricing

Morwitz, Greenleaf and Johnson (1998) named the practice and gave it its mechanism: splitting a price into a base and a mandatory surcharge lowers the total people recall, because the base reads as the price and the surcharge as a detail. Across their studies, as the authors' own later review reports, between 12.2% and 35.6% of participants left the surcharge out entirely when asked what the item cost, and stating it as a percentage of the base rather than in currency pushed recalled totals lower still. The interface decision is not whether to itemise but which number sits in the slot people read as the price.

The pooled record does not support the simple version of the claim. Abraham and Hamilton (2018) meta-analysed 149 observations from 27 papers, N = 12,878, and put the mean effect of partitioning on demand at d = 0.08, p > .10 — indistinguishable from zero — with I² = 83.72%: significant positive and negative effects both sit in this literature and cancel. The moderators set the sign. In the full coefficient table, reported in Abraham's (2015) dissertation version of the analysis, showing the total made partitioning less favourable, a gap of 0.32 d between total-absent and total-present; a surcharge seen as delivering high benefit made it more favourable by 0.56 d, and one that is typically partitioned in the category by 0.58 d. Absolute and relative surcharge magnitude were both null as main effects, and the effect shrinks year on year. Partitioning has no direction of its own; the total's visibility and the surcharge's plausibility do.

The large field numbers people quote for partitioned pricing are really about total salience. Chetty, Looney and Kroft (2009) added a second, tax-inclusive price tag to roughly 750 products at a supermarket where shelf prices excluded a 7.375% sales tax. Quantity sold and revenue in the treated group fell about 8% against both control groups, so against a category price elasticity of 1 to 1.5, simply showing the total cost about what a 7.375% price rise would have: most shoppers had not been counting the tax. They knew it was there — the median surveyed shopper correctly reported the tax status of 7 of the 8 products asked about — they just did not add it until a tag did it for them.

Blake, Moshary, Sweeney and Tadelis (2021) ran the same comparison on StubHub, randomly assigning half of US users to see mandatory fees only at checkout and half to see fee-inclusive prices from the first inventory screen. The back-end-fee group spent 20.64% more over ten days and was 14.1% more likely to buy at all. The intensive margin is the interesting half: conditional on buying, those users paid 5.42% more per transaction and 5.73% more per seat, so at least 28% of the revenue gain came from trading up. StubHub's buyer fee runs about 15% of the ticket price, so every listing looked 15% cheaper while people were still choosing. The funnel shows the cost — back-end users were about 19% more likely to select a ticket but about 45% less likely to complete at checkout, where the fee finally appeared — and experience did not erase it: users with ten or more prior visits still showed a 15% revenue effect against 21% overall.

The same obscured total produces choices people regret. Santana, Dallas and Morwitz (2020) dripped optional surcharges across six studies. In the first, 54.5% of participants initially picked the lower-base-fare airline against 11.7% when fees were shown upfront, and 24.5% against 7.8% ended up paying more than the alternative would have cost. Satisfaction with the final choice was lower under dripping (M = 5.39 against 5.83), and the choice itself carried it: people who ended on the lower-base option rated it 4.60 against 5.92 for those on the higher-base one. The fix was not the total on its own but a comparison — showing the competing airline's price alongside the total at the confirm step cut the mistake rate from 15.2% to 6.9%, and from 39.4% to 16.0% among those who had taken the cheaper-looking fare.

The effect has a ceiling, and the ceiling is where the surcharge stops being ignorable. In Hossain and Morgan's (2006) matched-pair eBay auctions, a 1¢ opening bid with $3.99 shipping beat a $4 opening bid with free shipping by about 21% for CDs, but the gain vanished where the reserve was 53% of retail instead of under 27%. Component count has the same ceiling: the field review by Greenleaf, Johnson, Morwitz and Shalev (2016) reports participants landing within 5% of the real total 49% of the time with two price components against 23.3% with nine, an inverse-U in purchase intention where one surcharge beat two, and surcharges exceeding the base price being judged unfair, with fairness fully mediating the effect of surcharge size on purchase intention.

So grade this mixed, and keep the two halves apart. That hiding the total raises short-run revenue is about as well established as a pricing finding gets: two large field experiments in different categories, a mechanism confirmed by survey and click-stream, and consistent lab replication. That splitting a price helps you is not established at all — the pooled estimate is 0.08 and falling, and the sign flips with the total's presence and the surcharge's plausibility. The revenue half is also largely no longer yours to take: showing the all-in total up front is a legal requirement, not a preference, in the UK (section 230 of the Digital Markets, Competition and Consumers Act 2024), the US for live-event tickets and short-term lodging (16 CFR Part 464) and the EU (Directive 2005/29/EC and Directive 2011/83/EU). The lab work behind those rules is blunt about the damage: Huck, Schmid and Wallace (2013) found drip pricing wiped out 25% of consumer surplus while leaving the decision problem unchanged, and Rasch, Thöne and Wenzel (2020) found sellers competing only on base prices and setting the highest drip they could get away with. Treat an obscured total as a revenue number you are no longer allowed to earn.