pith. sign in

arxiv: 2607.00280 · v1 · pith:UYADBARYnew · submitted 2026-07-01 · 💻 cs.LG · cs.CY· econ.EM· stat.AP

Understanding Guest Preferences and Optimizing Two-sided Marketplaces: Airbnb as an Example

Yufei Wu , Daniel Schmierer This is my paper

Pith reviewed 2026-07-02 16:19 UTC · model grok-4.3

classification 💻 cs.LG cs.CYecon.EMstat.AP
keywords Airbnbcausal inferenceprice elasticityguest preferencestwo-sided marketplacepersonalizationeconomic modelingbooking behavior
0
0 comments X

The pith

Causal inference on Airbnb booking data estimates how guests respond to prices and other listing factors.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper combines economic modeling with causal inference techniques to recover how guests decide to book based on the prices hosts set, along with other listing attributes. It further examines how these responses differ across guests and across different types of listings. The resulting estimates are intended to guide the design of host-facing pricing tools and guest-facing personalization systems. If the estimates are reliable, they can help hosts set prices that increase bookings while keeping the marketplace balanced between supply and demand.

Core claim

By fitting causal models to observational booking records, the authors recover price elasticities and preference heterogeneity that can be directly used to optimize the pricing assistance Airbnb offers hosts and to improve how listings are shown to individual guests.

What carries the argument

Causal inference applied to booking data to estimate price elasticities and guest-level preference heterogeneity.

If this is right

  • Hosts supplied with price suggestions derived from the elasticity estimates will set more competitive rates and obtain higher booking volumes.
  • Personalized ranking and search results that incorporate guest-specific price sensitivity will produce better matches between guests and listings.
  • Market-wide use of these tools will reduce mismatch between supply and demand and improve affordability for guests while raising host earnings.
  • Heterogeneity estimates allow the platform to target pricing interventions at segments where guests are more or less price responsive.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same causal-modeling approach could be applied to other two-sided platforms that observe transaction-level price and choice data.
  • If the estimated elasticities prove stable over time, they could be used to simulate the effects of platform-wide fee or tax changes before they are implemented.
  • The work leaves open whether the recovered heterogeneity can be explained by observable guest or listing characteristics or requires additional latent variables.

Load-bearing premise

Observational booking records together with standard causal inference methods can recover the true causal effect of price on booking probability without large unmeasured confounding or selection bias.

What would settle it

A field experiment that randomly assigns different prices to otherwise identical listings and measures the resulting change in booking rates would show whether the observational elasticity estimates match the experimental ones.

Figures

Figures reproduced from arXiv: 2607.00280 by Daniel Schmierer, Yufei Wu.

Figure 2
Figure 2. Figure 2: Heat Maps of Channel Residual Impressions Across DMAs Over Time [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Hypothetical illustration of preference heterogeneity [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
read the original abstract

Airbnb is a community based on connection and belonging -- many hosts on Airbnb are everyday people who share their worlds to provide guests with the feeling of connection and being at home; Airbnb strives to connect people and places. Among our efforts to connect guests and hosts, we provide tools to enable hosts to set competitive prices, which helps improve affordability for guests while helping hosts get more bookings. We also personalize the guest experience to show them the listings that match their needs. To help inform these efforts, we combine economic modeling and causal inference techniques to understand how guests book stays based on the prices hosts set, among other factors, and how that preference varies across different guests and listings. Such understanding helps us identify opportunities for Airbnb to support the marketplace and better connect guests and hosts. For example, understanding how much guests respond to different prices helps optimize the tools that we provide to hosts, in order to enable hosts to choose and set competitive prices that further balance demand and supply. As another example, understanding heterogeneity in guest preferences helps us personalize the guest experience and better match them with the listings that meet their needs, based on how much they respond to different prices and other factors.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The paper claims that combining economic modeling with causal inference techniques applied to Airbnb booking data allows recovery of how guests respond to host-set prices (including heterogeneity across guests and listings), which in turn informs optimization of host pricing tools and personalization of guest experiences to better balance the two-sided marketplace.

Significance. If the recovered price elasticities and heterogeneity are causally identified, the results could directly support practical marketplace interventions such as improved pricing recommendations and listing matching. The work is notable for attempting to link micro-level preference estimation to platform-level optimization in a real two-sided setting.

major comments (2)
  1. [Abstract] Abstract: the central claim that 'causal inference techniques' recover guest price responses rests on an unspecified identification strategy. No instruments, policy discontinuities, supply-side model, or other source of exogenous variation is named, leaving the estimates exposed to simultaneity (hosts observe demand signals when setting prices) and selection (bookings reflect two-sided matching with unobserved quality and preferences). This is load-bearing for all downstream uses in pricing tools and personalization.
  2. [Abstract] Abstract: the weakest assumption—that observational booking data plus standard causal methods suffice to recover causal elasticities—is not defended or tested. In the Airbnb setting, host pricing is endogenous and many listing/guest attributes are unobserved; without explicit discussion of how these are addressed, the estimates cannot support the claimed optimization applications.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for these comments on the abstract. Both points correctly note that the abstract does not name or defend an identification strategy. We will revise the abstract to address this.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that 'causal inference techniques' recover guest price responses rests on an unspecified identification strategy. No instruments, policy discontinuities, supply-side model, or other source of exogenous variation is named, leaving the estimates exposed to simultaneity (hosts observe demand signals when setting prices) and selection (bookings reflect two-sided matching with unobserved quality and preferences). This is load-bearing for all downstream uses in pricing tools and personalization.

    Authors: We agree the abstract should indicate the identification approach. The manuscript combines economic modeling with causal inference on observational booking data; we will revise the abstract to name the key source of variation used to address simultaneity and selection (drawing from the methods section) so that the downstream claims are properly supported. revision: yes

  2. Referee: [Abstract] Abstract: the weakest assumption—that observational booking data plus standard causal methods suffice to recover causal elasticities—is not defended or tested. In the Airbnb setting, host pricing is endogenous and many listing/guest attributes are unobserved; without explicit discussion of how these are addressed, the estimates cannot support the claimed optimization applications.

    Authors: We agree the abstract does not defend the identifying assumptions. We will revise it to include a concise statement referencing how the paper addresses endogeneity and unobserved attributes (via the economic model and any robustness analyses), thereby clarifying support for the optimization uses. revision: yes

Circularity Check

0 steps flagged

No circularity detected; derivation relies on external causal assumptions

full rationale

The provided abstract and description outline a standard application of economic modeling plus causal inference to observational booking data for recovering price elasticities and heterogeneity. No equations, parameter-fitting steps presented as predictions, self-citations, or ansatzes are shown that would reduce any claimed result to the inputs by construction. The load-bearing step is the validity of causal identification assumptions, which are external and untestable within the paper itself rather than self-referential. This is the most common honest finding for industry papers describing applied causal work.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract contains no equations, parameters, or explicit assumptions; ledger is therefore empty.

pith-pipeline@v0.9.1-grok · 5743 in / 1014 out tokens · 31823 ms · 2026-07-02T16:19:31.054688+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

12 extracted references · 3 canonical work pages

  1. [1]

    Customized regression model for airbnb dynamic pricing

    Peng Ye, Julian Qian, Jieying Chen, Chen-hung Wu, Yitong Zhou, Spencer De Mars, Frank Yang, and Li Zhang. Customized regression model for airbnb dynamic pricing. InProceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining, pages 932–940, 2018

    2018

  2. [2]

    Search ranking and personalization at airbnb

    Mihajlo Grbovic. Search ranking and personalization at airbnb. InProceedings of the Eleventh ACM Conference on Recommender Systems, RecSys ’17, page 339–340, New York, NY, USA, 2017. Association for Computing Machinery. ISBN 9781450346528. doi: 10.1145/3109859.3109920. URL https://doi.org/10.1145/ 3109859.3109920

    work page doi:10.1145/3109859.3109920 2017

  3. [3]

    Real-time personalization using embeddings for search ranking at airbnb

    Mihajlo Grbovic and Haibin Cheng. Real-time personalization using embeddings for search ranking at airbnb. InProceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, KDD ’18, page 311–320, New York, NY, USA, 2018. Association for Computing Machinery. ISBN 9781450355520. doi: 10.1145/3219819.3219885. URL https://doi...

    work page doi:10.1145/3219819.3219885 2018

  4. [4]

    Prioritizing home attributes based on guest interest.the Airbnb Tech Blog, 2023

    Joy Jing and Jing Xia. Prioritizing home attributes based on guest interest.the Airbnb Tech Blog, 2023

    2023

  5. [5]

    Reducing in- terference bias in online marketplace pricing experiments.arXiv preprint arXiv:2004.12489, 2020

    David Holtz, Ruben Lobel, Inessa Liskovich, and Sinan Aral. Reducing in- terference bias in online marketplace pricing experiments.arXiv preprint arXiv:2004.12489, 2020

    work page arXiv 2004

  6. [6]

    Experi- mental design in two-sided platforms: An analysis of bias.Management Science, 68(10):7069–7089, 2022

    Ramesh Johari, Hannah Li, Inessa Liskovich, and Gabriel Y Weintraub. Experi- mental design in two-sided platforms: An analysis of bias.Management Science, 68(10):7069–7089, 2022

    2022

  7. [7]

    The price is right: Removing a/b test bias in a marketplace of expirable goods

    Thu Le and Alex Deng. The price is right: Removing a/b test bias in a marketplace of expirable goods. InProceedings of the 32nd ACM International Conference on Information and Knowledge Management, pages 4681–4687, 2023

    2023

  8. [8]

    Automobile prices in market equilibrium.Econometrica, 63(4):841–890, 1995

    Stephen Berry, James Levinsohn, and Ariel Pakes. Automobile prices in market equilibrium.Econometrica, 63(4):841–890, 1995

    1995

  9. [9]

    MIT press, 2010

    Jeffrey M Wooldridge.Econometric analysis of cross section and panel data. MIT press, 2010

    2010

  10. [10]

    Empirical models of demand and supply in differ- entiated products industries

    Amit Gandhi and Aviv Nevo. Empirical models of demand and supply in differ- entiated products industries. InHandbook of industrial organization, volume 4, pages 63–139. Elsevier, 2021

    2021

  11. [11]

    Random-coefficients logit demand estimation with zero-valued market shares.Marketing Science, 40(4): 637–660, 2021

    Jean-Pierre Dubé, Ali Hortaçsu, and Joonhwi Joo. Random-coefficients logit demand estimation with zero-valued market shares.Marketing Science, 40(4): 637–660, 2021

    2021

  12. [12]

    Estimating demand for differenti- ated products with zeroes in market share data.Quantitative Economics, 14(2): 381–418, 2023

    Amit Gandhi, Zhentong Lu, and Xiaoxia Shi. Estimating demand for differenti- ated products with zeroes in market share data.Quantitative Economics, 14(2): 381–418, 2023

    2023