पाठशाला Pathshala · दल Dal, The team · Lesson 17 · Build

Remote, hybrid or office: the decision for Indian teams

The working model is a decision about the work, the cities and the seniority of the team, not about perks. What the evidence shows, the variables that differ in India, and the rituals each model needs to work.

Pathshala, The Founder Library · 11 October 2026 · 7 min read

Aerial black and white view of cars and trucks packed in a traffic jam on a Mumbai expressway. — Photograph: abshky · Pexels

Founders argue about remote work as if it were a question of belief. It is a question of fit: what the work needs, where the people live and how much of the team is still learning the job. Answer those three and the model mostly chooses itself. What is left is the part founders skip, which is writing down the rituals that make the chosen model work.

This lesson sets out what the best evidence shows, the variables that differ for Indian teams, a decision tree to walk your own team through, and the rituals that hybrid and remote each need.

Decide from the work, not the perks

The working model is usually discussed as a benefit: something the company gives or withholds. That framing produces bad decisions in both directions. A founder who believes in the office mandates five days for a team of experienced engineers spread across three cities and loses the two best. A founder who wants to be generous makes everyone remote and then wonders why the five graduates hired in June are still lost in December.

Three questions decide most of it. Does the work need people in one place? Is the team mostly early in their careers, learning by proximity, or mostly experienced? And does it live in one city or many? A fourth decides whether remote can work at all: does the team already write things down? Everything else, including the founders’ preferences, is a tiebreaker.

What the evidence shows

The strongest evidence on hybrid work is a randomised trial at Trip.com, published in Nature in June 2024 by Nicholas Bloom, Ruobing Han and James Liang. Of 1,612 employees in its airfare and IT divisions, half, chosen by whether their birthday fell on an odd or an even date, were given the option to work from home on Wednesdays and Fridays. Quits fell from 7.2 per cent to 4.8 per cent, a reduction of a third, and job satisfaction rose. Performance reviews and promotions over the following two years showed no difference between the groups, and the lines of code written by the 653 engineers showed none either.

The detail matters for a founder. The fall in quits was largest among women, at more than half, among non-managers and among people with long commutes. Managers’ own estimate of the effect on productivity moved from minus 2.6 per cent before the trial to plus 1.0 per cent after it. And the design was specific: fixed days at home for everyone in the group, three days together in the office. The authors note it is one company in one country, but it is the cleanest test there is, and it supports structured hybrid rather than either extreme.

The best-documented account of what goes wrong comes from GitLab, which runs fully remote and publishes its handbook on hybrid work. Its warning is the two-tier company: people who are in the office get more information, more informal access and, in time, more promotions, while remote staff become a satellite office. The office, it says, becomes a crutch for collaboration. Every model can fail; hybrid fails in this particular way unless someone designs against it.

The Indian variables

Cities and commutes. Long commutes in Bengaluru, Mumbai and Delhi NCR are a daily cost that the Trip.com result suggests matters most to retention. A company with an office in one corner of a city is, in effect, choosing which half of that city it can hire from. Anchor days reduce the cost without giving up the overlap.

A Mumbai local train running on tracks under overhead power lines. — In the Trip.com trial the fall in quits was largest among people with long commutes. Where a team lives decides what each office day costs it. Photograph: Abhishek Mishra · Pexels

Homes. Many early-career people in Indian cities live in shared flats or with family, without a quiet room, reliable power back-up or a good desk. Remote work for them is not a perk. Budget for a co-working seat or an equipment allowance where it is needed, or accept that the office is their best workplace.

Seniority. Indian startups often hire young: graduates and people two or three years in. They learn most from sitting near experienced people, which is the part of the office hardest to reproduce online. The more junior the team, the stronger the case for days together.

The law. The labour codes in force since November 2025 permit work from home in service sectors by mutual consent. Put the model, the anchor days and any equipment terms in the appointment letter or a written policy, so that consent is on paper. Checked 10 October 2026.

The decision, team by team

Walk each team through the tree. A company can reach different answers for different teams, and often should: the warehouse on site, the engineers hybrid, a senior designer in another city remote. The tree gives the default; write down the exceptions.

Take a company of thirty in Bengaluru. Eight people run a fulfilment hub on the city’s edge. Fourteen engineers and designers work from a small office in the centre, nine of them less than three years into their careers. Four salespeople live in Mumbai and Delhi NCR, close to their customers, and four people in finance and operations sit with the founders. The hub is on site; that is settled by the work. The product team is mostly junior with five experienced people who can be there three days a week, so it runs hybrid with Tuesday, Wednesday and Thursday as anchor days. The sales team is experienced and spread across cities, and it already writes call notes and deal reviews in the CRM, so it is remote, with a visit to the Bengaluru office each quarter. Finance and operations follow the founders’ anchor days. Four teams, three models, one written page that explains why.

The working model follows from the work, the cities and the seniority of the team. The rituals that make it work have to be written down, or the office quietly wins.

Hybrid done properly

Hybrid is the most common choice and the easiest to do badly. Four rituals make it work. Fixed anchor days for the whole team, as in the Trip.com design, so that the days people come in are the days the people they need are there; a hybrid policy where everyone picks their own days produces offices full of people on video calls with colleagues at home. One process for every meeting: GitLab’s rule is that every work meeting has an agenda in advance and that informal decisions are written down afterwards, so that nobody needs to have been in the room. Leaders not in the office every day: GitLab goes further and argues that the best place for leaders in a hybrid company is outside the office, where they meet the same friction as everyone working remotely. Watch promotions by location: once a year, compare who was promoted against how often they were in the office. If the two line up, the company has built the two tiers it meant to avoid.

Protect the anchor days from what erodes them. Customer meetings, interviews and travel go on the other days where possible, so that a day in the office is not spent on video calls. Put the team’s planning, its retrospectives and its one-on-ones on anchor days, and keep the home days for work that needs a closed door. And measure attendance on anchor days only, in aggregate, once a month: if fewer than three in four people turn up, the days are wrong, the commute is too long or the office is not worth the trip, and each of those has a different fix.

Remote done properly

Remote works for experienced teams that write. Decisions, specs and the reasoning behind them live in documents that anyone can read; meetings exist to discuss what has been written, not to transmit it. The [writing culture lesson](/library/writing-culture-decisions-in-documents) is the precondition, not an add-on. Two more rituals: a budget for the whole team to meet in person every quarter, for planning and for the trust that video does not build; and deliberate onboarding, because a new remote hire has no corridor to learn from. The [onboarding lesson](/library/onboarding-first-thirty-days) applies with more force, with the buddy doing daily video check-ins in the first two weeks.

The quarterly check

Every quarter, review the model against four numbers for each team: attrition, especially among people in their first year; output against plan; the time a decision takes from proposal to written answer; and, once a year, promotions by location. Ask new joiners at their ninety-day review whether they could find what they needed and who to ask.

Change the model when the answers to the three questions change, not when a founder has a frustrating week. A team that was mostly junior and is now mostly experienced may be ready for fewer anchor days. A team that has spread across cities needs the writing habit before it needs a policy. Write every change down with its reason, and give people a quarter’s notice of it.

The evidence comes from one large trial; your own attrition and output are the test that counts. Nothing here is legal advice.