Remote-first startup hiring playbook: entity, comp, async loop, and onboarding in 2026

Remote-first is no longer a differentiator, it is a default expectation for senior candidates in most functions. What still varies wildly between startups is how well they run the mechanics: entity setup, comp policy, async interview loops, and the first thirty days after signing. This playbook is the current best practice we teach founders on the Startup Roles platform.

Entity and employment: who to hire, how, and where

The single biggest early decision is whether to use an Employer of Record (EoR), set up your own local entity, or engage contractors. Get this wrong and you either overpay by 15 to 25% in fees, or you inherit tax, IP, or employment liability that will surface at diligence.

The default: EoR for the first ten hires

For your first ten remote hires across geographies, an Employer of Record is almost always the right answer. Deel, Remote.com, Oyster, and Multiplier are the four names that come up repeatedly. Expect 8 to 15% of gross comp in fees, plus statutory costs.

EoRs solve four things simultaneously: legal employer status in the country, payroll and tax withholding, statutory benefits, and IP assignment enforceable in that jurisdiction. You do not want to reinvent any of these before you have to.

When to open your own entity

The rough rule: three or more employees in the same country, expected to grow to five plus inside a year, at which point the local entity cost (£15k to £40k to set up, similar annual running cost) breaks even against EoR fees. UK founders hiring in the US usually open a Delaware C-corp and a payroll provider (Rippling or Gusto) before hiring a fourth US employee.

Contractors: proceed with care

Genuine independent contractors are fine. Misclassified employees who are functionally full time are the single largest employment liability we see in seed stage diligence. If you use contractors, be honest about it: contract length under six months, they use their own tools, they invoice on their own schedule. Anything else and you are running an EoR arrangement without the paperwork.

Comp policy: pick one model, publish it, defend it

Comp policy is the surface where remote-first startups leak the most trust with their team. The fix is to pick one of three models, publish it internally, and defend it at every offer.

1. Location-based bands

Adjusted to the local market using a cost of labour index (Numbeo, Payscale, or an in house version). Cheapest, but hardest to defend at the top of your band because two engineers doing identical work take home materially different amounts.

2. Global bands

One number regardless of geography. Easiest to hire against, hardest on burn. Companies like GitLab and Buffer moved away from this at scale. It is still viable at seed if you keep the team small.

3. Zoned bands (the pragmatic middle)

Three or four zones such as: Zone A (US, Switzerland, London), Zone B (UK excluding London, Berlin, Amsterdam), Zone C (Southern and Eastern Europe), Zone D (Rest of World). Each zone gets its own band; within a zone, comp is location agnostic. This is the model we see working at the majority of Series A remote-first companies in 2026.

Equity policy

Equity should be geography agnostic. Two engineers doing the same work take the same equity slice, regardless of where they live. Anything else creates a two class system that surfaces every time an option grant vests.

The async interview loop

A well run async loop looks like:

1. 30 minute intro call (live, video).
2. Take home work sample with a fixed two week window and a clear "spend no more than 4 hours" instruction.
3. One live paired session or working meeting (60 to 90 min).
4. Reference calls (async written or short live calls).
5. Founder or hiring manager final.

Total elapsed time under 14 calendar days. Loops that stretch past three weeks lose approximately 40% of the top of the funnel to competing offers or renewed inertia.

Making the take home worth doing

Senior candidates increasingly refuse take homes. The three things that reliably restore consent to do one:

• Cap the time explicitly. "Spend no more than 4 hours. If the task takes longer, stop and describe what you would do next."
• Pay for it. £150 to £250 per hour, invoiced normally.
• Use a real problem from your backlog. Not a synthetic puzzle. The candidate's output should be something you can plausibly ship.

Onboarding: the first thirty days decide the next twelve months

Remote hires who cannot answer "what am I supposed to do this week" reliably churn inside six months. Ship a written "First 30 days" document before their start date. The template we recommend:

First 30 days template

• Week 1. Environment setup, meet the 8 people you will work with most, first read only PR reviewed and merged.
• Week 2. First production PR merged. First customer conversation attended as a listener.
• Week 3. Own one shipped surface end to end. Write your first architecture or product note.
• Week 4. Present a "what I have learned and what I will change" document to the founding team.

The buddy and manager split

Every remote hire gets a buddy (a peer who answers "how do we do X here" questions without judgement) and a manager (the person who owns their scope, comp, and career). Never collapse the two roles into one person.

The remote-first tools stack

The stack we recommend for a remote-first team of 10 to 40:

• Communication: Slack or Discord, plus Loom for async video.
• Docs: Notion or Linear (pick one and stick to it).
• Meetings: Google Meet or Zoom, plus Tandem or Around for casual co-working.
• Recruiting: Ashby for scheduling and pipeline, Startup Roles for sourcing and outreach.
• Payroll and compliance: Deel or Remote.com for EoR, Rippling or Gusto for owned entities.
• Equity: Carta or Pulley.

Common failure modes in year one

• Time zone drift. Hires cluster in a single time zone despite policy. Fix by writing an explicit "overlap hours" requirement into every JD.
• Meeting inflation. Async companies quietly become sync companies as team size grows. Cap recurring meetings per person per week and audit quarterly.
• Onboarding decay. The First 30 Days doc goes out of date after the third hire. Assign an owner and refresh every quarter.
• Comp band drift. Individual offers stretch bands until the published policy is fiction. Pull comp data quarterly, publish the audit internally, adjust bands rather than break them.

Frequently asked questions

Should we go remote-first or hybrid?

Pick one and be honest. Hybrid with an unspoken expectation of three office days is the worst of both worlds for senior candidates.

How do we handle team offsites?

Two per year for teams under 30, one per year above that. Budget £2000 to £3500 per person per offsite fully loaded. Skip the first offsite and turnover in the following year rises noticeably in our data.

Do remote-first startups hire more slowly?

No. In our data they hire slightly faster because the addressable candidate pool is roughly 10x larger. What varies is offer acceptance rate, which is 5 to 10 percentage points lower for remote roles because candidates weigh multiple competing remote offers more actively.

What about US contractors on a UK entity?

Only for genuinely independent work under six months. Anything longer, run it through an EoR or open a Delaware entity.

For the sourcing side, read our passive sourcing playbook. For the JD side, see our 2026 JD framework.