The productivity “sweet spot” in hybrid work (2024–2025): what the evidence actually says

Its a contentious topic and one that seems to have leaders at an impasse. Do executives follow the likes of Amazon CEO, Andy Jassy in requiring corporate employees to return to the office 5 days a week. Others who have implemented strict RTO policies include AT&T, UPS, BlackRock, JPMorgan with Goldman Sachs, Tesla, Disney, GM and Apple pushing employees to be in for 4-5 days per week. On the flip-side of the argument you have the fully remote or "Work from anywhere" model championed by Atlassian (Team Anywhere policy), Autodesk, Twilio, CloudLinux, Spotify and Docusign etc.

If you are a Chief People Officer or other senior executive exploring the vast array of hybrid, fully remote or in-office options, this blog post is hopefully going to be a useful, evidence backed, guide to help you make an informed decision that may have far reaching productivity and financial implications for your business. At the very least, it should help direct your thinking towards an option that suits your business, customers and employees.

Spoiler alert.

Across the highest-quality studies we have (including a large randomized controlled trial and multiple quasi-experimental papers), the weight of evidence points to hybrid schedules with ~2–3 in-office days per week as the best overall balance for productivity, retention, collaboration, and training—provided those office days are used intentionally for activities that benefit from colocation (e.g., onboarding, mentoring, design sprints, problem-solving). Fully remote can work well for focused, individual work but risks weaker innovation and slower skill development for juniors; forcing full-time office presence rarely improves output and often harms retention.

1) What we mean by “productivity” (and why it varies)

“Productivity” isn’t one number. Studies measure it via output per hour (calls handled, code submitted, papers/patents), performance ratings and promotions, retention (lower attrition prevents productivity loss), and macro indicators. The right answer depends on task mix (focus vs. collaboration), seniority (mentorship needs), and tooling/management quality. Government and academic reviews stress exactly this nuance.

2) The best causal evidence: the Trip.Com randomized controlled trial (published in Nature)

• Design: 1,612 graduate professionals randomized to a 3–2 hybrid (home Wed/Fri) vs. five days in-office. Followed for 6 months + 18-month outcomes.

• Results: Attrition down ~33%, no hit to performance ratings, promotions, or measured output (including lines of code). Managers who initially predicted productivity losses revised upward after the experiment. Trip.com rolled out hybrid firm-wide.

  
  

Implication: When structured, hybrid preserves productivity while improving retention—a major, compounding productivity driver.

3) Why 2–3 office days tends to be the “sweet spot”

Several data sources converge on roughly two or three in-office days:

• The RCT’s successful schedule concentrated in-person time on two anchor days (with a third on-site day), maintaining collaboration without sacrificing focused work.

• A/B-tested deployments described by Bloom and co-authors for knowledge-work companies report comparable performance with two–three office days and lower quit rates, shifting sceptical managers.

• Global survey/behavioural data show hybrid has stabilized (not fading), and typical hybrid patterns cluster around part-week on-site; college-educated workers average around one day WFH globally but much more in remote-capable roles in English-speaking countries—so hybrid norms commonly land at 2–3 office days.

  
  

Why this works:

• In-person days are disproportionately valuable for mentoring, onboarding, complex coordination, and fast conflict resolution.

• At-home days maximize deep work by reducing commute and interruptions. Hybrid captures both benefits if office days are intentionally loaded with colocation-benefiting work (not solo tasks you could do at home).

  
  

4) The collaboration & innovation caveat (and how hybrid fixes it)

• Creativity/ideation: Lab and field experiments show videoconferencing can depress idea generation compared with in-person brainstorming. Use your in-office days for ideation sprints.

• Breakthroughs/knowledge flow: Large-scale analyses of papers/patents and within-firm proximity find face-to-face exposure boosts knowledge transfer and long-run human-capital development, though it may trim short-run output. Hybrid lets you schedule this exposure without giving up the focus gains of remote days.

  
  

5) What national and macro data are telling us (2024–2025)

• Worldwide: Post-pandemic, WFH settled into a new steady state; hybrid persists despite high-profile RTO mandates. This persistence—and the RCT evidence above—suggests hybrid is not a productivity drag when managed well.

• Australia: The 2025 Productivity Commission review finds the WFH shift isn’t behind Australia’s recent productivity dip; hybrid’s impact on productivity is neutral to positive and benefits participation, especially for carers and people with disabilities. Policy focus should be on investment and management quality, not banning hybrid.

  
  

6) Implementation playbook: turning 2–3 days into real gains

Anchor the right work on the right days

• On-site days (2–3/wk): onboarding, mentorship 1:1s, design reviews, cross-team decision meetings, difficult conversations, and relationship-building. (This is where proximity has outsized returns.)

• WFH days (2–3/wk): deep work (writing, analysis, coding), asynchronous collaboration, documentation, and individual upskilling.

Make the office day “worth it”

• Publish team-level on-site calendars 4–6 weeks ahead; co-locate allied teams on the same days to increase serendipity.

• Use meeting hygiene (clear agenda, decision owner, pre-reads) and prefer whiteboard-first sessions in-person; keep recurring status meetings async. (Inefficient meetings are a top productivity blocker in hybrid.)

Protect focus all week

• Adopt norms that defend uninterrupted blocks and discourage the “infinite workday” pattern (after-hours ping-pong and fragmented calendars). Track it and fix it.

Measure outcomes, not presence

• Team-level OKRs/business output, cycle times/throughput, defect rates, customer NPS, and retention; compare like-for-like before/after hybrid changes rather than using badge swipes as a proxy. (The RCT shows attrition is a meaningful performance lever.)

Mind the junior pipeline

• Require structured mentoring on in-office days (shadowing, code/design reviews) and instrument it (who meets whom, how often). This addresses the long-run skill-build benefits of proximity without erasing focus time.

  
  

7) Sector & role nuance (quick guide)

• Product/dev, research, sales engineering: Lean 3 office days during intense design/launch phases; drop to 2 otherwise.

• Content, analytics, backend engineering: Often thrive with 2 office days; keep collaboration rituals tight on those days.

• Early-career cohorts: Bias to more in-person mentoring touchpoints even if overall office days stay at two.

Bottom line

If you’re optimizing for sustained productivity and retention, the best-supported policy in 2025 is hybrid with 2–3 intentional in-office days per week. Use those days for the work that truly benefits from proximity, and protect deep work on the others. That’s where the evidence is strongest.

As a bit of light relief. I was speaking with a leader from a large Australian tech scale-up the other day who had the funniest take of the "If a tree falls in the forest" thought experiment. His version was "If an executive is in the office and there is no-one there to see him wearing his $5k suit, is he really important?" You can guess what side of the argument this exec has taken.

References (selected)

• Bloom, N., Han, R., & Liang, J. (2024). Hybrid working from home improves retention without damaging performance. Nature. (Trip.com RCT results summary & paper.) Nature

• Bloom, N., Han, R., & Liang, J. (2022/2023). How Hybrid Working From Home Works Out. NBER Working Paper 30292 (rev.). (Full RCT details.) NBER+1

• Bloom, N., Liang, J., & Han, R. (2024). One Company A/B Tested Hybrid Work. Here’s What They Found. Harvard Business Review. (Manager attitudes & outcomes under hybrid.) Harvard Business Review

• Aksoy, C.G., Barrero, J.M., Bloom, N., Davis, S.J., Dolls, M., & Zarate, P. (2025). The Global Persistence of Work from Home. (G-SWA, global stabilization of WFH.) wfhresearch.com

• Emanuel, N., Harrington, E., & Pallais, A. (2023–2024). The Power of Proximity to Coworkers (working paper; NY Fed blog summary). (Proximity boosts mentoring/skill-build; short-run output trade-off.) NBERLiberty Street Economics

• Brucks, M.S., & Levav, J. (2022). Virtual communication curbs creative idea generation. Nature. (Videoconference vs. in-person ideation.) Nature

• U. of Oxford & U. of Pittsburgh (2023). Remote collaborations deliver fewer scientific breakthroughs. Nature (news release). (Breakthrough innovation vs. remote). University of Oxford

• U.S. Bureau of Labor Statistics (2024). The rise in remote work since the pandemic and its impact on productivity (Briefing). (Why productivity impacts vary.) Bureau of Labor Statistics

• Stanford SIEPR (2024). Hybrid work is a “win-win-win” for companies and workers (research summary). SIEPR

• Australia: Productivity Commission (2025). Productivity before and after COVID-19 (overview). (WFH not the cause of productivity dip.) Productivity Commission

• CEDA (2024). More than a third of Australians still work from home (data & participation effects). CEDA