Why work samples beat resumes for overseas hires

Every resume from every finalists pool says the same things: 'detail-oriented', 'proactive', 'strong communicator'. These phrases cost nothing to type and nothing to verify. When a candidate is overseas and you're hiring remotely, you also can't rely on the informal signals that a walk through the office provides — the way someone navigates a tricky hallway conversation, the speed at which they read a room. You are working almost entirely from documents and calls.

Work samples change that equation. They ask the candidate to do a version of the job, not describe doing the job. The output is gradable, comparable, and visible before you schedule a single interview. The practical implication: a five-person e-commerce brand can screen thirty applicants in less time than it takes to run ten Zoom calls, and do it on evidence instead of instinct.

What a resume actually tells you

A resume confirms that someone held a title and stayed in a role long enough to list it. It says nothing about the quality of the work inside that tenure. A bookkeeper who listed 'QuickBooks, accounts payable, month-end close' for three years may have run a tight ledger or may have inherited a mess they never fully cleaned up — the resume looks identical either way.

Experience listed in years is also a poor proxy for skill level. Someone who spent four years doing rote data entry has four years of experience and very little breadth. Someone who spent eighteen months inside a fast-moving startup finance function, closing books monthly under a CFO who actually reviewed them, may have built more applicable judgment. The resume surface doesn't distinguish these two people.

What a work sample actually tells you

A good work sample is a compressed version of a real task from the role. For a bookkeeper, that might be a messy transaction log with three deliberate errors: can the candidate find them, categorize correctly, and flag the anomaly clearly? For an executive assistant, it might be a cluttered inbox summary exercise: can they surface the three things that need action today and draft the replies in a register that matches your voice?

The output answers the question you actually care about: can this person do the work? Not 'have they done work like this' — the answer to that is always yes on a filtered applicant list — but 'does the evidence in front of me right now match what I need?' That's a different, harder, more useful question.

Grading matters as much as the task itself. A work sample graded by gut feel after you've seen the candidate's photo and name is not much better than a resume screen. Grade against a rubric, blind to identity where possible, before you open the cover letter. That sequence protects you from the halo effect that derails most screening processes.

Why this matters more for overseas candidates specifically

When you hire locally, you often fill in gaps with ambient information: a mutual contact, a quick reference call to someone you know, the way a candidate carried themselves in person. None of that infrastructure exists across a twelve-hour timezone gap with someone you've never met in a country you may never have visited.

The Philippine BPO industry employs roughly 1.9 million people, according to IBPAP, the industry association. That means there is a deep, experienced candidate pool, but also that the volume of applicants for any posted role will be high. You need a filter that scales and that produces comparable, gradable output across dozens of applications. A work sample does both. A resume review does neither at scale.

Language is also genuinely testable in a written task in a way it is not on a resume. A resume says 'excellent written English'. A work sample shows you the actual prose a candidate produces under mild time pressure, which is exactly the condition under which they'll be drafting your client emails or your internal updates.

How to design a task that isn't a trap

The goal is not to trick candidates. A work sample designed to be deliberately obscure tells you about your candidates' ability to handle obscure tasks, which is probably not what you need. The goal is a representative slice of the role's actual day-to-day.

Keep it short enough to be fair. If you're asking for unpaid work, it should take under ninety minutes to complete. Longer tasks either get skipped by the best candidates (who have options) or produce inflated output that doesn't reflect real working pace. Tell candidates exactly how long it should take.

Give context. A task that says 'write a product description' without a brand brief, a tone guide, or an example of what you like tests the candidate's ability to guess what you want, not their ability to write. A task with a one-paragraph brief and a sample you actually like tests the skill you're hiring for.

Grade before you meet. Build a rubric with four or five dimensions and score each dimension separately before you look at anything else about the candidate. If you're evaluating a customer support response, score tone, accuracy, brevity, and resolution-focus as separate items. You're looking for a total score, not a vibe.

The rubric question: what to weight and why

A typical work-sample rubric for an overseas hire might weight task quality most heavily, around a third of the total score, and then factor in scenario judgment, experience specificity in follow-up questions, written English quality, and salary-band fit. Those weights reflect what actually predicts on-the-job performance for remote roles where written output and independent judgment matter most.

Publishing your rubric to candidates before they complete the task is underrated. It signals that the process is fair and graded, not subjective. It also raises average output quality because candidates know exactly what you're looking for. Higher average quality is good for you, it raises the floor of who clears the bar.

One thing to avoid: weighting presentation over substance. A beautifully formatted spreadsheet with wrong numbers is worse than a plain one with correct numbers. Make sure your rubric rewards accuracy and judgment, not polish.

Where Rolemote fits into this

Rolemote screens overseas candidates using a published rubric: graded work sample at 35%, scenario judgment at 25%, experience specificity at 20%, written English at 15%, and salary-band fit at 5%. The weights and what they measure are public at the /how-we-screen page, nothing is proprietary or hidden. You can check the rubric before you start.

The whole search runs free, describe the role, get a free AI hiring brief, read your scored finalists. Payment happens only when you ask to meet finalists: a flat fee, not a percentage of salary. If nobody clears your bar, a re-run is free. Typical agencies in this space charge 25-35% of first-year salary, roughly $4,500-6,300 on common roles; Rolemote's flat fee is $1,995. There is also an optional Success Plan at $99 per month per active hire, which includes lifetime replacement and monthly AI check-ins that flag quit-risk early.

The reason to name the rubric publicly is that 'rigorous vetting' and 'top 1%' are claims you cannot check anywhere on any agency's site. A published rubric with published weights is a claim you can check, disagree with, or improve on. That's the point.

Before you post your next role

Suppose you're hiring a virtual assistant based in the Philippines at, say, $900 per month, well within the published range of $700-1,150 per month for that role. If that person quits in month two because the brief was vague and the role turned out to be nothing like what the resume suggested to either side, you've spent two months of salary, your own onboarding time, and whatever recruiting cost you paid to find them. None of that is recoverable.

A work sample built before you post, one that represents a real task, carries a real rubric, and is graded before names are visible, costs you a few hours once. It applies to every applicant. It is the single highest-leverage thing you can do before a single resume lands in your inbox.