Extend, Convert, or Release: How to Decide When a Contract Ends

Last updated: September 11, 2026

By Tom Kenaley, President and Senior Partner, KORE1

Extend a contractor’s contract when the work continues but conversion is blocked, convert when the work is permanent and you’d rehire the person tomorrow, and release them once either the work or the fit has run out. Two facts settle nearly every contractor contract extension question. Both are knowable a month before the end date, which is about three weeks earlier than most teams go looking for them.

The end date on a contract is usually the least accurate number in the whole agreement. Somebody wrote it on day one while guessing how long a project would run, and by month four the project, the budget, and the person sitting in the seat have all moved. Nobody updates it. Then the staffing firm emails to ask whether the assignment still ends on the 30th, and a decision that deserved a month gets made between two other meetings.

That email has three possible replies. Extend, convert, or release. Each one is correct somewhere. The problem is the default, which is “extend another 90 days,” because it’s the only reply that doesn’t require anybody to schedule a conversation. Meetings are hard to book.

At KORE1 our contract staffing desk watches this decision land on managers in IT, accounting, and engineering every week. The ones who handle it well aren’t smarter about contractors. They look at the calendar sooner.

What a Contractor Contract Extension Changes, and What It Leaves Alone

A contractor contract extension is a written amendment that moves a contract worker’s end date, and sometimes the bill rate or scope along with it, while every other term of the original agreement stays in force. The contractor remains on the staffing firm’s W-2 payroll. Their employment status doesn’t change at all. That’s the appeal. It’s also the risk.

On paper an extension is small. Usually it’s a one-page change order against the statement of work or the purchase order, a new end date, and maybe a new rate if somebody asked for one. Procurement may want fresh budget approval for the added hours. On our side the assignment record just gets a later date. Nobody re-interviews anyone.

Conversion is bigger. The contract ends, the person leaves the agency’s payroll and joins yours with a salary, benefits, and an offer letter, and the agency is usually owed a conversion fee that depends on how many hours the contractor has already billed. Release is the smallest of the three on paper and the one I see handled worst in practice. More on that further down.

Two Facts Sort Most of These

I’ve seen managers build weighted scorecards for this. Seven criteria, colored cells, a formula at the bottom. I built one myself, years ago. Mine, like most of them, ended up restating two facts in a more complicated way.

Is the work still there after the end date? And if you were starting the search from scratch tomorrow, would you pick this same person off the shortlist?

Three of those four boxes are easy. People still get them wrong, mostly by waiting, but the answer itself is never in doubt.

The top-left box is where the judgment lives. The work is permanent and the person is good, yet something stands between you and a conversion. Maybe it’s a headcount freeze that doesn’t touch contract spend, which happens often enough that we wrote a whole page on keeping critical seats staffed through a freeze. Maybe the contractor doesn’t want a salary. Some don’t. I’ll get to them. Or maybe the work is real but it’s twenty hours a week of it rather than forty. Any of those can justify an extension. None of them justifies an open-ended one.

There’s a trap in the bottom-left box too, a quieter one. The project finished. Everyone liked the contractor. And suddenly there’s a “documentation phase” nobody requested, or a loosely defined support role that exists mainly so a good person doesn’t leave. I understand the impulse. I’ve also watched it turn a clean six-month engagement into an eighteen-month one with no owner and no end, which happens to be the exact shape that makes employment lawyers nervous.

When an Extension Is the Right Call

An extension is right when the end date is tied to something real that moved, or to something real that hasn’t happened yet.

The cleanest example I can think of came off our accounting desk last winter. A manufacturer in Anaheim had a contract senior accountant covering month-end close while it cut over from an old on-premises ERP to NetSuite. The original end date was January 31. The cutover slipped to March. Meanwhile the controller had already signed a permanent hire who couldn’t start until April 13, so the extension ran to April 24, with two weeks of overlap built in so the new person learned the close from someone who had already run it on the new system. Nobody argued about it. One amendment, with a date anybody in the building could explain. It ended exactly when it said it would.

Compare that to the extension nobody decides.

A medical device company in Irvine kept a QA automation engineer on 30-day extensions for five straight months. Nobody was being careless on purpose. Her department’s contractor budget got approved one month at a time, so that’s how the paperwork went out. She had written most of their Playwright regression suite and knew which tests were flaky and why. Nobody else did. Around the third week of the sixth extension, she accepted a permanent offer somewhere else, and I don’t blame her, because every month she’d spent roughly ten days not knowing whether she would be working in thirty. When the suite broke in November, her replacement spent close to three weeks just finding where the test fixtures lived.

Short, rolling extensions feel cautious. They’re the opposite. Each one hands your contractor a fresh reason to take a recruiter’s call, and the good ones are getting those calls constantly.

I should be upfront about something here. Extensions are the most profitable outcome on this page for a firm like ours, since the hours just keep billing, so telling you to stop rolling them 30 days at a time costs us money in the short run. I’d still rather you stopped. The contractor you lose that way is the one we end up replacing for you, usually in a hurry and usually at a bad moment.

Two more things come up at almost every extension.

Rate is the first. A contractor’s bargaining power peaks at the extension point, and experienced ones know it. Plenty of them ask for a raise, especially past the six-month mark. The bumps we see mostly land in the low single digits as a percentage of the pay rate. A skill whose market moved during the assignment can justify more. Budgeting for it ahead of time beats treating it as an ambush, because the alternative is pricing a replacement, and a replacement costs you weeks of ramp before it costs you anything on an invoice.

Length is the second. Three or four extensions in a row push an assignment toward a second year, and somewhere in there the question stops being about budget. Long tenure plus employee-style treatment is the pattern regulators and plaintiffs’ attorneys look for. Our contingent workforce management guide goes through where that risk sits and why tenure caps protect less than most companies assume. If you’re on the third extension and can’t say why, you probably already know which box you’re in.

Converting a Contractor Without Botching the Offer

Convert when the work has outlived the contract and the person is the one you’d choose again from a fresh shortlist.

If the role was set up as contract-to-hire from the start, most of this should already be on paper, including the evaluation checkpoints, the target date, and the fee schedule. Our employer’s guide to contract-to-hire walks through how those get set before day one. Plenty of conversions don’t begin that way, though. They begin as plain contract work that quietly turned permanent, and those are the ones that get fumbled at the offer stage.

Know where the contractor sits on hours before anything else. Most staffing agreements step the conversion fee down as billable hours pile up. Ours runs 20% of first-year base through 800 hours, 15% through 1,040, and nothing after that, which is about 26 weeks at forty hours a week. The dollar examples are in our contract-to-hire conversion rate benchmarks. An extension that carries someone past 1,040 hours can make the fee disappear entirely, which is worth knowing, though it shouldn’t set your timing if the person is already fielding other offers.

Then build the offer on the right math. This is where I see the most avoidable losses.

A logistics company in Ontario, California, had a data engineer on contract at a $95 hourly pay rate, which sits mid-range for the role in our 2026 contractor rate data. She’d been running their Snowflake and dbt pipelines for about eight months. The hiring manager wanted to convert her. He multiplied $95 by 2,080 hours, got $197,600, and assumed she’d expect something close to that. Finance had a ceiling of $165,000. He decided the gap was hopeless and started planning a release. He was wrong.

Nobody bills 2,080 hours, though. Take out ten holidays and three weeks of vacation, none of it paid on most contract assignments, and a realistic year is closer to 1,880 hours. At $95 that comes to $178,600. A salary keeps paying through those same holidays and vacation days. Then there’s insurance, mostly health coverage, and retirement contributions, which an employer largely pays for and which she’d been covering herself or going without. For private-industry workers in management, professional, and related occupations, the Bureau of Labor Statistics Employer Costs for Employee Compensation release put those two categories at $8.79 per hour worked in June 2026. Over 1,880 hours, that’s roughly another $16,500.

So the $165,000 offer was worth around $181,500 to her. Close to lateral. He made the offer at $168,000. She signed two days later.

Had he released her, he’d have been paying to replace someone he wanted to keep. Gallup’s widely cited estimate, from its 2019 research on turnover, puts the cost of replacing an employee at one-half to two times annual salary. She wasn’t his employee, technically. A data engineer who knows where every pipeline breaks is expensive to replace no matter what paperwork she’s on.

A few other things decide whether a conversion lands.

• The base should be checked against the market, and not only against the contract rate. Our salary benchmark assistant is a quick way to see whether $165,000 is fair for the role or just the number finance had left.
• Some contractors will say no, and it isn’t personal. People who contract on purpose often prefer the rate and the freedom to choose their next project over a salary and a 401(k). Ask early, before anyone drafts an offer.
• Does the new-hire process start over? Usually yes, with a benefits enrollment window and sometimes a background check your company runs separately from the one we ran. Say so up front.
• Title and level. People converted at the wrong level resent it for years.

Releasing a Contractor Cleanly

Release is the easy one on paper. There’s no severance to negotiate and no unemployment exposure on your side, because the contractor is on the staffing firm’s payroll rather than yours, and the firm is the one that formally ends the assignment. That ease is exactly why teams get sloppy with it.

Tell the contractor early. A good contractor who hears “this ends on the 30th” three weeks out will line up the next assignment, often through us, and will speak well of you to the next five people who ask. Hearing it on the Wednesday before a Friday end date, they tell a different story, and the contract market in any one metro is smaller than it looks from inside a single company.

Releases for fit are different. When the fit is wrong, I’d much rather hear it at week three than at the end date. Most agreements let you close an assignment on short notice, and a clear call early is kinder to everyone than a bad month.

The handoff matters more than people budget for. Most of what’s valuable about a departing contractor lives in their head, not in the repository. The best version I’ve seen came from a hospital system’s integration team in Orange, where the manager made the handoff a line item in the last two weeks of the contract and put her own name on it as the person receiving it. The contractor wrote up every scheduled job he ran. He listed the vendor portals he logged into, without handing over a single password. Then he spent an afternoon walking her through the three failures nobody else had seen, one of which was an HL7 interface that quietly stopped acknowledging messages whenever a certificate rotated. Four months later it did exactly that, and the fix took twenty minutes. Teams that skip this step tend to hear the contractor’s name again in a postmortem.

Then the accounts close, the same day.

NIST’s security control catalog for federal systems, SP 800-53 Revision 5, has two controls that fit this moment almost word for word. PS-4 calls for disabling system access within a defined time period when someone’s employment ends, and revoking their credentials. PS-7 goes further for outside personnel, requiring the external provider to notify you when a contractor holding your credentials or system privileges is transferred or terminated. Most private companies aren’t bound by it. Plenty of SOC 2 auditors test something close to it anyway.

One of our clients, a SaaS company in Newport Beach, found that out during SOC 2 Type II fieldwork this spring. The auditor pulled a sample of departed users, and one of them was a contractor released in February whose Okta account and GitHub organization membership were still active in April. The agency that placed him, not us as it happens, assumed the client would disable his accounts. The client assumed its termination workflow covered him, but that workflow ran off the HR system, and contractors weren’t in the HR system. It became an exception in the report. Nobody had done anything malicious. Nobody had owned the ticket, either.

Every account opened during contractor onboarding needs a matching close, scheduled for the last afternoon, with an owner’s name on the ticket.

Last comes the release where the person goes but the work doesn’t. That’s the top-right box, and it’s a backfill, which is a different search from the original one because the scope, the systems, and the stakeholders are all known now. SHRM’s 2026 recruiting benchmarking found that employers take a median of 39 calendar days to fill a nonexecutive job. Three weeks of notice doesn’t cover that. Contract cover can. Backfill staffing exists for exactly this gap, keeping the seat covered while a direct-hire search for the permanent person runs alongside it. We fill IT roles in 17 days on average, and contract cover tends to move faster than that.

The Last 45 Days, Week by Week

None of the above works if the decision starts the week of the end date. This is the calendar we try to get clients onto. It isn’t sophisticated.

Why tell the contractor at 21 days? Because by then they’ve usually been taking calls for a week or two already. Contractors plan around end dates the way a renter plans around a lease. The American Staffing Association put staffing industry turnover at 376% for 2025, down from 416% the year before, according to its fourth-quarter 2025 employment and sales release. Much of that churn comes from short light industrial and clerical assignments, nothing like a senior data engineer’s. It still tells you what the contract labor market is built to do. Move.

The owner column matters more than the dates do. On the extension-by-default teams I described earlier, that column was blank. Nobody owned it. The staffing firm assumed the manager would decide, the manager assumed procurement would flag it, and procurement renewed whatever was about to expire, because renewing things that are about to expire is what procurement is for.

What Managers Ask Us Once the End Date Is Close

Is the Friday before the end date enough notice for a contractor?

Three weeks is about what a good contractor needs to line up the next assignment, so the Friday before rarely works, even when your staffing agreement only requires a few business days.

The contractual minimum is a floor, not a plan. Anyone good has been fielding recruiter calls since about the halfway point of the assignment, so they can move quickly once they know. Telling them at 21 days also gives us time to line up a replacement, if you need one, before the seat goes empty.

Can we keep extending the same contractor indefinitely?

Short answer: for a while, yes, but every extension after the second should come with a written reason and a date, or it’s a conversion you’re avoiding.

No federal rule caps how long a W-2 contractor can stay on one assignment. The legal exposure comes from long tenure combined with treating the person like an employee in every way that counts. The practical problem arrives sooner. Somewhere around month twelve, the contractor starts to wonder why they’re good enough to keep but not good enough to hire. They usually ask.

They want a raise to extend. Is that normal?

Normal, and more likely at an extension than at any other point in an assignment, because it’s the one moment when the contractor’s bargaining power and your switching cost are both at their highest.

Plan for it. A few dollars an hour, even with markup on top, rarely comes close to what several weeks of ramp time costs a team, and the replacement you would hire instead will probably ask for the same rate anyway.

What if they turn down the conversion offer?

$95 an hour can be worth more to someone than a $170,000 salary, depending on how much they value flexibility, so a no isn’t always about the number you offered.

Ask why before you counter. If it’s money, the fix is usually the base or a signing bonus. If they simply prefer contracting, you’re back in the top-left box, and the choices are an extension with a clear date or a planned release with a backfill lined up well ahead of it. Either one is fine. What ends badly is pretending the no didn’t happen and extending month to month while everyone avoids the subject, which I have seen go on for most of a year at one client before the contractor finally left for a competitor.

Is it cheaper to keep extending than to convert?

Usually not past the first year, because markup keeps billing every hour while a conversion fee is paid once, and under many agreements that fee shrinks as hours accumulate.

Contracting stays cheaper when the work really might end, since part of what you’re paying for is the ability to stop. Once the work is clearly permanent, you’re paying a premium for flexibility you’ll never use. Mike Carter ran the full break-even between contract-to-hire and direct hire with a worked example, if you want the curve.

Do we owe the staffing firm anything if we release early?

The hours worked and whatever notice your agreement specifies, and that’s usually it, since no conversion fee applies when a contract simply ends.

A few agreements carry minimum-hour commitments, mostly on specialized project work. Read the termination clause before the day you need it.

Put a Name Next to Every End Date

If you manage more than a couple of contractors, pull the list this week and sort it by end date. Anything inside the next 45 days gets an owner and one line answering the two facts. Nobody needs software for this. It takes about twenty minutes, and it’s the difference between making these calls yourself and having a calendar make them for you.

KORE1 has been placing people since 2005, on everything from a six-week contract to a permanent executive hire, with recruiters working searches in more than thirty metro areas, and our one-year retention on placements sits at 92%. Some of that comes from getting these end-of-contract calls right. If you have a contractor whose date is coming up and you can’t tell which box they belong in, talk it through with one of our recruiters. Sometimes the honest read is to let the contract end, and you’ll hear that from us too.