If Hiring Takes a Quarter, a Contractor Is a Calendar Decision
You do not hire a contractor the way you hire an employee. Paper, access and a paid pilot can run at the same time. The delays are almost always sequential process, not scarcity of engineers.
European senior backend hiring is slow even when the market is friendly. You write a role, wait for a shortlist, run four rounds, wait for a notice period, then wait again until the person can ship in your repository. Three to six months from “we should hire” to useful output is ordinary. A contractor is not faster because they type faster. They are faster because you can skip the parts of that calendar that exist to protect a multi-year employment relationship.
That is the whole argument for “hire a backend contractor quickly,” and it is easy to waste. Teams copy the employee pipeline onto B2B: extra interview rounds, an unpaid take-home, legal review that starts after the technical yes, access tickets that start after the signature. By the time the person can clone the repo, the reason you needed them has moved by a quarter.
I work B2B from Serbia with CET/CEST overlap, so I have seen both the fast path and the six-week stall. The fast path is not magic. It is parallelism: NDA, MSA, SOW, access, and a paid evaluation on real work, overlapping, with one named owner on your side who is allowed to unblock IT.
The calendar you are actually buying
A permanent hire buys durability. The process is long because reversing the decision is expensive. A contractor buys a bounded period of senior capacity. The process should match that product.
What you still need, and should not skip:
- Evidence they can work in your stack (PHP, Go, your database, your CI), not a whiteboard.
- Written scope: owned outcomes, days per week, overlap hours, definition of done.
- Legal and security paper proportional to the data they will see.
- A start date that is a date, not “after onboarding.”
What you should not copy from the employee pipeline:
- Four culture rounds for a ten-day pilot.
- An unpaid weekend project as a filter. Strong contractors decline; you filter for people who can work for free.
- Waiting for a headcount code when the freeze already told you there is no FTE. If the constraint is a hiring freeze, contracting is a different budget line. Treat it as one, or admit you cannot buy capacity at all.
- A laptop. They have a machine. Your delay is SSO and Git.
If the work is actually “own this domain for four years,” do not hire a contractor quickly. Hire the employee and tell the business the date. Speed is the wrong optimisation when permanence is the requirement.
Run NDA, SOW and access in parallel
The slow version is a chain:
- Informal chat
- Technical interview
- Internal yes
- Legal receives an MSA for the first time
- Security questionnaire
- Vendor form in the ERP
- IT ticket for GitHub
- First day, no staging
- First useful PR in week three
The fast version starts three tracks on the same day you decide the person is worth evaluating.
Track A: confidentiality and commercial paper
Send the NDA before a deep dive in the repository. Many teams already have a one-way NDA that legal will sign in 24 hours. Do not wait for the MSA to have a conversation about architecture at the whiteboard; wait for the NDA to have production code on a screen.
In parallel, send the MSA or framework agreement: liability, IP assignment, notice, governing law, insurance if you require it. IP assignment is not automatic the way employment often is. If it is missing, you will discover it when someone asks who owns the migration scripts.
The SOW or order form is the engagement: scope, rate, days per week, start date, data processing if needed. You can draft the SOW while legal reads the MSA. Multiple SOWs under one MSA is why the MSA is worth doing properly once.
If personal data is in scope, start the DPA and the list of systems at the same time. Privacy owners who first hear about a contractor after git clone are why start dates slip.
None of this is legal advice. It is a sequencing note: paper that sits in a queue after you already said yes is how a two-week start becomes six.
Track B: access
IT should see a request the same week as the NDA: identity in your SSO, Git host, CI, staging, observability. Write the groups, not “access to everything.” Read-only production logs are usually enough for a pilot; production deploy keys are a later conversation.
If your access process requires a company laptop, a named desk, and a photo for a badge, you have an employee process. Cut a contractor path: BYOD, SSO, VPN or zero-trust, time-boxed account. Security can still require disk encryption, a password manager, and MFA. Those are hours, not weeks, if someone owns the ticket.
Start the ten-day evaluation clock when the environment works, not when the PDF is signed. Otherwise you will measure onboarding latency and call it a failed contractor.
Track C: the work itself
Pick the pilot ticket while legal reads. Production-relevant, bounded, not checkout on day one. Agree the success criteria in writing. Agree who reviews the PR. Agree overlap hours in CET/CEST.
If you cannot name a ticket, you are not ready to hire anyone quickly. You have a product problem.
A paid pilot is the evaluation, not a delay
Interviews tell you whether someone can talk. A 10-day paid pilot tells you whether they can open a pull request in your repository, take review, and write down what they will not touch.
Ten days is not a delay if you would have spent those ten days on another interview loop and a take-home. It is a substitution. You pay for the days. You get a merged change or a clear checkpoint. Both sides can stop.
Design the pilot so it can start the week paper is close enough, not after a ceremonial kickoff in month two:
- Real repo, not a sandbox fork that will be thrown away.
- A counterpart who answers the same day during overlap.
- Daily written notes: moved, blocked, unsafe.
- Pay the day rate. Unpaid trials select for people who can subsidise your process.
A pilot that cannot start because “we need to finish the hiring committee” is not a contractor process. You put the employee calendar back on top.
What actually blocks start dates
These are the blockers I see. Almost none of them are “no senior backend engineers in Europe.”
Sequential legal. Engineering said yes on Thursday. Legal opened the MSA on Monday two weeks later. Fix: legal sees the MSA when engineering sees the resume.
Vendor onboarding designed for agencies. Insurance certificates, three years of audited accounts, a DUNS number, a purchase-order cycle that runs monthly. A sole trader B2B contractor will fail a pack built for a 200-person consultancy. Either maintain a contractor path or admit you can only buy agencies.
SSO and Git as a post-signature surprise. The account cannot be requested until there is a vendor ID, the vendor ID cannot be created until there is a signed SOW, the SOW cannot be signed until security has seen the Git groups. Break the cycle: a time-boxed guest account for the pilot, then the full vendor record.
Unclear SOW. “Help with the backend” cannot be papered. Legal will bounce it. Engineering will not know what done means. Write a paragraph: the worker, the upgrade slice, the incident class, the days, the overlap.
Waiting for an unpaid exercise. Two weeks of silence while a senior person with other clients declines to do a toy app. You lost calendar and signal.
The hidden employee steps. Background check products that require a local employment contract. Equipment procurement. A start date aligned to a cohort onboarding in another country. None of that is required for a remote B2B engineer with their own machine.
No counterpart. Access exists, the board exists, and every question waits for a tech lead who is in a hiring loop for someone else. The contractor is idle. You will say they were slow.
Scope that cannot be touched until a freeze meeting. If finance has not decided whether contractors are allowed, that decision is the critical path. Do not run interviews as theatre.
Security questionnaires after architecture review. Send the questionnaire with the NDA. Most answers are boring: no subprocessors for the coding work, machine in the contractor’s control, MFA, no copying production dumps onto a laptop. If your questionnaire assumes a SaaS product with tenants, you are using the wrong form.
Hoping the start date is “ASAP.” ASAP is not a date. Put a Monday on the SOW. If access slips, move the Monday. Do not start the paid clock in a vacuum.
What you can decide in one week
A realistic week, if someone on your side has authority:
- Day 1: NDA out, MSA out, access ticket opened, 60-minute conversation about a real system they built, pilot ticket named.
- Day 2–3: technical follow-up on your PHP or Go constraints; security questionnaire returned; SOW drafted.
- Day 4–5: legal comments on MSA (or a known fallback: their paper for the pilot only, yours for the continuation); staging login tested; start date set.
That is not a complete MSA for a two-year programme. It is enough to start ten paid days. Continuation can sit under a fuller MSA. Teams that insist the full vendor pack is finished before a single clone are choosing a hiring-quarter calendar. They should be honest about that.
Timezone helps. CET/CEST overlap with a contractor in the EU neighbourhood means the access test and the first pairing do not wait on a transatlantic day. It does not replace paper.
When hiring quickly is the wrong goal
Speed is a bad reason to put a senior in a codebase you cannot supervise. If nobody can review, nobody can answer domain questions, and nobody will own the result, a fast start produces orphaned code faster.
Speed is also a bad reason to skip IP assignment, a DPA, or production-data rules. Parallel is not the same as absent.
And speed is a bad reason to staff a permanent role with a contractor because the job ad would take time. That role will still exist in six months, and you will have paid a premium for the privilege of not opening it.
How I start, when the calendar is the constraint
My commercial path is built for this: B2B, monthly EUR invoice, CET/CEST overlap, €300/day, up to five days per week, starting with a 10-day paid pilot on real repository work. I sign the client’s NDA and MSA rather than inventing a delay of my own. Details sit on the senior backend contractor page.
If you want to hire a backend contractor quickly, the work on your side is not finding a person who can start “immediately” in the abstract. It is putting NDA, SOW, access and a ticket on the same week. The engineers exist. The sequential process is what takes months.