A July 2026 working paper (Song, Liu, and Agarwal, arXiv:2607.03718) estimates that when app development teams adopt remote work, their iOS apps release more major versions and introduce more new features without losing originality, using job-posting-inferred adoption and a staggered difference-in-differences design. The authors explicitly leave open whether this positive causal effect survives direct measurement of remote-work policies from internal records, whether the proposed mechanisms (team-size expansion and collective skill capacity) can be formally established rather than suggested, whether the effect extends beyond successful top-ranked apps to failed or discontinued products, and whether it holds for digital products other than mobile apps. No published work measures product-level continuous digital innovation together with directly observed remote-work adoption in any product population, and idea-generation studies point in the opposite direction on novelty, leaving the sign, mechanisms, and generality of the effect genuinely open.

Background

Remote work (work from home, WFH) became a durable feature of knowledge-intensive labor markets after 2020, and firms now make divergent choices about it. A first economics literature measured its consequences for individual productivity and well-being: the Ctrip experiment (Bloom, Liang, Roberts, and Ying 2015) found performance gains for call-center workers; more recent studies of hybrid arrangements (Bloom, Han, and Liang 2024) find retention gains without performance damage; Emanuel and Harrington (2024) study selection and performance in remote call-center work. A second literature studies innovation and collaboration specifically and is more pessimistic: Brucks and Levav (2022) show in laboratory experiments that video-mediated communication curbs creative idea generation; Yang et al. (2021) show from corporate telemetry that firm-wide remote work made collaboration networks more siloed and asynchronous; Lin, Frey, and Wu (2023) show that remote scientific teams fuse fewer breakthrough ideas, and Frey and Presidente (2026) extend this analysis of remote collaboration and scientific breakthroughs. Gibbs, Mengel, and Siemroth (2024), using internal records from a large Indian IT-services firm, find that the quantity of employee-submitted improvement ideas did not change when employees worked from home but idea quality declined. These two literatures stand in tension: remote work seems benign or beneficial for routine individual output, but costly for the novel-idea component of innovation.

Digital product innovation in mobile-app markets is measurable at a granularity patents cannot reach. Release notes published with each app version distinguish major releases (substantial redesigns or new capabilities) from routine maintenance, and text-analysis of the notes identifies newly introduced features and whether each feature is original (first appearance in its app category) or imitative, an approach validated in the information-systems literature on platform innovation (Foerderer et al. 2018; Kaushik and Gokpinar 2023). 'Continuous innovation' denotes this stream of incremental releases and feature introductions rather than discrete invention events.

Into this gap, Song, Liu, and Agarwal (2026) pose the question directly for digital product innovation. They build an app-month panel of iOS apps in five categories (education, games, lifestyle, business, utilities), restricted to apps that appeared in the daily top-500 ranking of their category during 2018-2023, observed from May 2020 to June 2023 (1,782 apps with valid pre-treatment periods, development teams of at least three members). They infer the month a team adopts remote work from job postings (Revelio Labs data, 11 development-related roles, a rule-based classifier detecting permissive WFH statements), following the teleworkability-measurement tradition of Dingel and Neiman (2020). Identification uses the Callaway and Sant'Anna (2021) staggered difference-in-differences estimator, which compares outcomes of apps whose teams adopt remote work at different dates while allowing treatment effects to differ across adoption cohorts and over time (avoiding the biases of conventional two-way fixed-effects estimators under heterogeneous effects), with not-yet-adopting apps as controls; an instrumental-variable analysis uses lagged internet connectivity in the locations of indirect labor-market competitors (competitors' competitors) to address measurement error, reverse causality, and omitted variables. They find remote work significantly increases major releases and new features per app, does not reduce the share of original (non-copycat) features, raises downloads, and matters more for teams with prior modular-collaboration experience from open-source participation. Team-size expansion and collective skill capacity (the number of distinct skills on the team) rise with adoption and are associated with better outcomes, but the authors state these mechanism analyses are 'at best suggestive'; commuting-time reduction and app maturity do not explain the gains.

The authors' own limitations section states what remains unresolved: adoption is inferred from job postings rather than internal policy records; qualitative mediators (work-life balance, job satisfaction, motivation) are unobserved; the sample conditions on commercial success (top-ranked apps), so failed or discontinued products are excluded; mechanism evidence is suggestive rather than formal mediation tests; and the evidence covers only mobile apps.

Problem Statement

The open problem is to establish whether remote work drives or deters continuous digital product innovation — and through which causal pathways and in which product populations — now that a first well-identified estimate exists only for job-posting-inferred adoption among commercially successful mobile apps. Song, Liu, and Agarwal (2026) estimate positive effects of remote-work adoption on the log number of major releases and new features per app-month (Callaway–Sant'Anna staggered difference-in-differences, 1,782 top-ranked iOS apps, May 2020–June 2023, with a labor-competition-network instrument), and explicitly identify the following open directions for future research:

  1. Direct measurement. Re-estimate the effect of remote work on product-level continuous innovation using directly observed remote-work policy (internal policy records or equivalent direct measures) rather than inference from job postings, removing residual measurement error in adoption timing and intensity.
  2. Mechanisms. Formally test the proposed causal pathways — team-size expansion and collective skill capacity, currently supported only by suggestive associational evidence and attenuation when the mediators are added — using formal mediation analysis or an identification strategy that isolates the channels. The source notes that qualitative mediators (work-life balance, job satisfaction, motivation) are unobserved and that incorporating such measures could further illuminate these channels.
  3. Population. Determine whether the effect extends beyond successful products to unsuccessful or discontinued digital products, since the source sample conditions on top-500 ranking and its identification compares only later versus earlier adopters among survivors.
  4. Product scope. Determine whether the effect generalizes to digital products beyond mobile apps (for example desktop software, SaaS, games, or web products), where comparable release- and feature-level innovation measures exist.

Of these directions, direction 1 — a directly measured, credibly identified estimate of the sign and magnitude of the effect in a defined product population — concerns the effect under direct adoption measurement, while directions 2–4 further characterize the effect through its mechanisms and its generality.

At the core of the problem is the sign and magnitude of the effect of remote work on continuous digital product innovation under direct adoption measurement, in a defined product population and with an identification strategy credible against nonrandom staggered adoption and pandemic-era confounds; whether the proposed mechanisms hold up under formal testing, and how far the effect extends across product success and product type, remain open. A rigorously identified null or negative result under direct measurement, with adequate power, would reconcile or contradict the innovation-suppression findings of the laboratory, telemetry, and scientific-collaboration literature.

References

  1. Song, Fangchen; Liu, Yixuan; Agarwal, Ashish (2026). Remote Work: Driver or Deterrent of Digital Product Innovation. arXiv:2607.03718, DOI 10.48550/arXiv.2607.03718, https://arxiv.org/abs/2607.03718
  2. Song, Fangchen; Liu, Yixuan; Agarwal, Ashish (2026). Remote Work: Driver or Deterrent of Digital Product Innovation. SSRN working paper, DOI 10.2139/ssrn.7052699, https://doi.org/10.2139/ssrn.7052699
  3. Gibbs, Michael; Mengel, Friederike; Siemroth, Christoph (2024). Employee Innovation During Office Work, Work from Home and Hybrid Work. SSRN working paper, DOI 10.2139/ssrn.4909536, https://doi.org/10.2139/ssrn.4909536
  4. Lin, Yiling; Frey, Carl Benedikt; Wu, Lingfei (2023). Remote collaboration fuses fewer breakthrough ideas. Nature 623(7989):987-991, DOI 10.1038/s41586-023-06767-1, https://doi.org/10.1038/s41586-023-06767-1
  5. Brucks, Melanie S.; Levav, Jonathan (2022). Virtual communication curbs creative idea generation. Nature 605(7908):108-112, DOI 10.1038/s41586-022-04643-y, https://doi.org/10.1038/s41586-022-04643-y
  6. Yang, Longqi; Holtz, David; Jaffe, Sonia; Suri, Siddharth; Sinha, Shilpi; Weston, Jeffrey; Joyce, Connor; Shah, Neha; Sherman, Kevin; Hecht, Brent; Teevan, Jaime (2021). The effects of remote work on collaboration among information workers. Nature Human Behaviour 6:43-54, DOI 10.1038/s41562-021-01196-4, https://doi.org/10.1038/s41562-021-01196-4
  7. Bloom, Nicholas; Han, Ruobing; Liang, James (2024). Hybrid working from home improves retention without damaging performance. Nature 630(8018):920-925, DOI 10.1038/s41586-024-07500-2, https://doi.org/10.1038/s41586-024-07500-2
  8. Emanuel, Natalia; Harrington, Emma (2024). Working Remotely? Selection, Treatment, and the Market for Remote Work. American Economic Journal: Applied Economics 16(4):528-559, DOI 10.1257/app.20230376, https://doi.org/10.1257/app.20230376
  9. Bloom, Nicholas; Liang, James; Roberts, John; Ying, Zhichun (2015). Does Working from Home Work? Evidence from a Chinese Experiment. Quarterly Journal of Economics 130(1):165-218, DOI 10.1093/qje/qju032, https://doi.org/10.1093/qje/qju032
  10. Dingel, Jonathan I.; Neiman, Brent (2020). How many jobs can be done at home? Journal of Public Economics 189:104235, DOI 10.1016/j.jpubeco.2020.104235, https://doi.org/10.1016/j.jpubeco.2020.104235
  11. Callaway, Brantly; Sant'Anna, Pedro H. C. (2021). Difference-in-Differences with multiple time periods. Journal of Econometrics 225(2):200-230, DOI 10.1016/j.jeconom.2020.12.001, https://doi.org/10.1016/j.jeconom.2020.12.001
  12. Frey, Carl Benedikt; Presidente, Giorgio (2026). Death of distance? Remote collaboration and scientific breakthroughs. Industrial and Corporate Change 35:1006-1042, DOI 10.1093/icc/dtag012, https://doi.org/10.1093/icc/dtag012
  13. Foerderer, Jens; Kude, Thomas; Mithas, Sunil; Heinzl, Armin (2018). Does Platform Owner's Entry Crowd Out Innovation? Evidence from Google Photos. Information Systems Research 29(2):444-460, DOI 10.1287/isre.2018.0787, https://doi.org/10.1287/isre.2018.0787
  14. Kaushik, Nilam; Gokpinar, Bilal (2023). Sequential Innovation in Mobile App Development. Manufacturing and Service Operations Management 25(1):182-199, DOI 10.1287/msom.2022.1154, https://doi.org/10.1287/msom.2022.1154