The procurement team at a mid-sized insurer in Rotterdam is preparing its third framework agreement this year with a supplier of IT specialists. The first two were signed quickly, on the vendors’ templates. Six months later the cost of those shortcuts became clear. At the first vendor, a senior Java engineer left in the middle of a release and the successor arrived after five weeks, because the contract only said the vendor would “use reasonable endeavours to provide a replacement promptly”. At the second, the client asked to replace a developer who was struggling and was told the guarantee had covered the first two weeks, and the problem was raised in week three. Both disputes ended without any remedy, because neither contract provided one.
The procurement lead negotiating the third agreement is not asking whether the vendor offers a replacement guarantee. They all do. She is asking how to write it so that it works on the day she needs it. And because the insurer is a financial entity, she also needs the contract to sit comfortably in the company’s DORA register of ICT arrangements.
This article is written for her and for anyone buying IT staff augmentation. We show what an effective replacement guarantee consists of, which SLA metrics make sense for staffing and which are copied without thought from support contracts, how to enforce deadlines without going to war with the vendor, and what red flags to watch for in a vendor’s draft.
Why is the replacement guarantee the most important clause in a staffing contract?
Because in staff augmentation you are buying access to specific people, and people leave, fall ill and sometimes do not fit. Every other clause, from rate and notice to IP and confidentiality, assumes the specialist is there and working. The replacement guarantee governs the situation where they are not, which is exactly when you lose money.
The loss has three parts. First, the vacancy: days in which work does not move and the team waits or absorbs the missing person’s tasks. Second, re-onboarding: the successor spends the first weeks learning the system at reduced output while you pay the full rate. Third, knowledge that left with the predecessor and has to be partly rebuilt.
Purely to show how quickly this adds up: take a senior at €390 per day, the rate in the worked example on our pricing page. If the successor works at about half capacity for the first two weeks, you pay 10 × €390, or €3,900, for roughly €1,950 of real output. Add the team’s time spent onboarding and every vacant day on which that role’s work stands still. With two or three replacements a year on a contract of several people, it is a sum worth negotiating over.
A good replacement guarantee does not remove these losses, but it limits them and allocates them fairly. It shortens the vacancy with a hard deadline, moves part of the onboarding cost to the vendor and forces a knowledge handover. A bad one, “the vendor will use reasonable endeavours to provide a replacement promptly”, does none of that and means very little in a dispute.
Be honest too about what a guarantee does not do. It does not replace good selection at the start: a vendor that sends random people and then swaps them complies with the contract while you lose weeks to repeated onboarding. Always read the guarantee alongside the description of the vetting process and your right to interview every person before they start.
There is a second, less obvious reason this clause matters. It is the best test of a vendor before you sign. A vendor with a bench of vetted specialists and a sound matching process will accept a concrete deadline without fuss. A vendor that recruits every person from scratch for your request will try to blur it. How a vendor negotiates this clause tells you more than its sales deck.
How do a replacement guarantee, an SLA and a remedy differ?
These three are often blurred, and each does a different job. Separate them before you start drafting.
The replacement guarantee is the vendor’s commitment to act: replace the specialist with someone of comparable skills, within a set deadline, on set cost terms. It answers the question of what the vendor will do when a specialist leaves or does not fit.
The SLA, or service level agreement, is a set of measurable indicators of the engagement’s quality, with targets and a measurement method. It answers the question of how you will check that the vendor is performing, including when nothing goes wrong. The replacement deadline is one SLA metric, but the SLA can cover more, such as time to present candidates or account manager response time. Our glossary entry on the service level agreement gives the general definition.
The remedy, whether liquidated damages, a service credit or a rate reduction, is the financial consequence of missing a commitment. It answers the question of what happens when the vendor misses the deadline. Without it, the guarantee and the SLA are declarations that are hard to enforce.
An effective contract has all three layers: the commitment (what the vendor will do), the measure (how you check it did) and the consequence (what happens if it does not). A missing layer is the most common reason for the situation in the introduction: the vendor formally offers a guarantee and the client has nothing to enforce it with.
Which events should trigger a replacement?
The guarantee should list the events that trigger it rather than say “where necessary”. Five events deserve separate treatment because each follows a different logic.
Poor fit in the initial period. The specialist passed interviews, but in practice cannot handle the work, does not fit the team, or their real skills differ from those declared. This is the classic scope of an initial guarantee: replacement at the vendor’s cost, without a long procedure, within a short period after the start.
The specialist leaves, on their own or the vendor’s initiative. They resign, take another offer, or the vendor moves them to another project. This sits on the vendor’s side however long the engagement has run, so it should be covered for the entire term.
Long absence. Illness, parental leave, personal emergencies. The contract should say after how many days of absence the client can require cover and whether it pays for the absence. With billing for days worked the second point usually solves itself, but the first needs a clause.
Removal at the client’s request for cause. A breach of confidentiality or security rules, gross negligence. The client must be able to require immediate removal and replacement regardless of the initial guarantee period.
Declining performance during the engagement. The hardest trigger, because it is the most subjective. A specialist who performed well for a year starts to deliver less. Here a procedure works better than automatic replacement: written notice, a conversation with the vendor’s account manager, an improvement plan with a deadline, and only then a right to replace.
Note that only the first event is typically time-limited. A frequent drafting error describes the whole guarantee as “a guarantee during the first weeks”, so a specialist leaving in month six is formally not covered at all.
How do you define a poor fit so the guarantee actually works?
An initial guarantee is useless if each side reads “not the right fit” differently. The vendor will defend its person, you will have an impression that is hard to prove, and the guarantee period will expire in an email exchange.
The answer is not a detailed definition of poor fit in the contract but a procedure that settles upfront that the assessment belongs to the client. Three elements are enough.
Assessment on the client’s side. During the initial guarantee period the client may require a replacement without detailed reasons, only a short explanation. The vendor does not judge whether the request is justified; it acts on it. That is fair, because in this period the risk of poor selection should sit with the party that selected.
Form and timing of notice. Written or email notice to a named person, and the date of notice, not the date of replacement, decides whether the guarantee period was met. A problem raised on the last day of the period is then covered even if the replacement happens later.
A role profile as the reference point. Attach a role profile to the contract or order: technologies, experience level, scope of work. It is not there to prove poor fit, but it makes the conversation easier after the initial period, when a replacement does need a reason.
Agree a review within the guarantee period. If the period is two weeks, set a short assessment with the team lead after the first week, so the decision is made inside the period, not after it.
Remember there is a person on the other side. A specialist replaced after two weeks is not necessarily weak; often they landed in a team whose way of working does not suit them. A good procedure lets the vendor move them to another project without damage to their reputation and lets the client part ways without a long argument about fault. That is one more reason a short, clear initial guarantee is better for everyone than a long period with a duty to prove poor fit.
Which deadlines belong in the contract?
Deadlines are the core of the guarantee and they must be numbers. “Promptly”, “within a reasonable time” and “with due diligence” are not deadlines but an invitation to a dispute. Four matter most.
The initial guarantee period. How long after the start the client may require a replacement without reasons. A short period, such as two weeks, forces quick assessment but is enough to see whether the specialist copes with the first tasks. A longer one gives the client more time but is harder for a vendor to accept without a higher rate.
Time to present replacement candidates. How many business days from notice the vendor has to present candidates. The vendor fully controls this, so it can be short.
Time for the successor to start. How many business days from notice until the successor’s first working day. This is the key metric because it measures the actual vacancy. State that it runs to the first day of work, not to the delivery of a CV.
Overlap for planned departures. If the specialist leaves with notice, the vendor knows in advance. The contract should require the vendor to tell the client immediately and ensure the successor starts before the predecessor leaves, with a few days of joint work for handover.
For the start deadline, also agree who is responsible for delays on the client’s side. If the client cannot find time to interview a candidate for a week or does not grant access, the vendor’s deadline should extend accordingly. That is fair and avoids an argument about who caused the delay.
Define how days are counted. “Five business days” means something different in an ordinary week and between Christmas and New Year. With a vendor in Poland and a client in the Netherlands, state whose public holiday calendar applies. The detail looks trivial until a specialist resigns on 20 December.
Who pays to onboard the successor?
This question is most often left out of contracts and is among the most expensive. For the first days, often weeks, the successor learns the system, gets access, reads documentation and asks the team questions. They work at reduced output and the team loses time onboarding them. If the contract is silent, the client pays the full rate throughout.
There are three reasonable solutions, and they can be combined.
Free replacement during the initial guarantee. If the replacement follows a poor fit in the first weeks, the vendor does not bill the specialist who did not work out, or does not bill the successor’s first days. The risk of poor selection then sits wholly with the vendor.
A free overlap on departure. When the specialist leaves on their own or the vendor’s initiative, a few days of joint work by predecessor and successor are at the vendor’s cost. The client does not pay twice for the same role.
A reduced rate during onboarding. For a replacement mid-engagement, the successor’s rate in the first week or two is reduced by an agreed percentage. A compromise that vendors find easier to accept than a full waiver.
Also pin down whom the vendor sends. The contract should speak of a person with “comparable skills” and say what that means: at least the same experience level and the same key technologies, confirmed by an interview with the client. A successor must not become a way to lower the team’s level at the same rate, or to raise the rate for the same role. The client should keep the right to reject a replacement candidate, and the replacement deadline should not restart with every rejection justified by a mismatch with the profile.
Agree these rules at signing, when nobody has a specific replacement in mind. Negotiating onboarding costs at the moment a specialist has just left and a release is waiting puts the client in its weakest position.
Which SLA metrics make sense for staff augmentation?
In staff augmentation the vendor is not responsible for how your system runs, only for supplying and retaining the right people. So SLA metrics should measure the staffing process, not technical outcomes. A common procurement error is importing metrics from managed services or support contracts, such as system availability or time to fix a critical bug. A staffing vendor does not control those results, because you direct the specialist’s work.
Six metrics make sense in most augmentation contracts:
- Time to present candidates from a request or replacement notice to vetted profiles.
- Time to start from candidate approval to the first day of work.
- Time to replace from notice to the successor’s first day, measured separately for poor-fit and departure cases.
- Specialist retention, the share of specialists who did not leave on their own or the vendor’s initiative in the period.
- Account manager response time to a reported issue, for example by the end of the next business day.
- Invoice accuracy, the match between invoices and approved timesheets.
Each metric needs three things: a target, a measurement method and period (for example a quarterly review), and a consequence for a breach. A metric without a consequence is information, not a commitment. Not every metric needs a financial consequence; for some, a duty to present a remediation plan is enough, but the replacement deadline is worth backing financially.
Keep a sense of proportion. Six well-measured metrics beat twenty that nobody tracks. An elaborate SLA that the vendor does not report and the client does not check gives a false sense of security.
Agree who produces the report and in what form. Simplest is a short monthly or quarterly summary from the vendor: roles, start and leave dates, replacement notices with dates and fulfilment, any breaches. The client confirms it in a review with the vendor’s account manager. That rhythm turns the SLA from a document in a drawer into a relationship management tool, and gives you hard data for the next tender.
What does a checklist of replacement and SLA clauses look like?
The table collects the clauses from this article in a form you can lay next to a vendor’s draft. The “reference value” column shows what a clause might look like; adapt the figures to the role and market. The last column lists wording that should raise a red flag.
| Clause | What it should define | Reference value | Red flag in a vendor’s draft |
|---|---|---|---|
| Replacement triggers | Poor fit, departure, long absence, removal for cause | A list of events in the contract | ”Where necessary”, “at the vendor’s discretion” |
| Initial guarantee | Period after start in which replacement needs no reasons | E.g. first 2 weeks, date of notice decides | Period measured to the replacement date, not the notice |
| Time to present candidates | Business days from notice to profiles | A few business days | No deadline at all |
| Successor start | Business days from notice to first day of work | E.g. 5 business days on departure | ”Promptly”, “reasonable endeavours” |
| Successor onboarding cost | Who pays for overlap and first days | Overlap and initial-period replacement at vendor’s cost | Full rate from the successor’s first day |
| Planned departure | Duty to inform and to overlap | Successor starts before predecessor leaves | No duty to inform the client |
| Missed deadline | Credit, no billing, termination right | Right to terminate the role without notice after the deadline | No consequence, or a token amount |
| SLA metrics | Staffing process, measurement, review | 5–6 metrics, quarterly review | Technical metrics the vendor cannot influence |
| Knowledge transfer | Ongoing and on departure | Repositories and documentation on the client’s side | Knowledge held in the vendor’s tools |
| Notice and exit | Period, scaling down, data | Notice in weeks, no termination fees | Long notice and exit charges |
The table does not replace legal advice, but it lets you check quickly whether a draft addresses these points at all. In practice the first and last columns carry most of the value: if a clause is missing or worded like the red flag, you know where to start negotiating.
Service credits or liquidated damages: how do you enforce deadlines?
When a vendor misses the replacement deadline, the client has several tools. Choose deliberately, because a large penalty is not always the best protection.
Liquidated damages. A pre-agreed sum payable when the deadline is missed. The advantage is that you do not have to prove how much the vacancy cost you. The limit is that courts will not enforce a sum that is really a punishment. Under English law the UK Supreme Court in Cavendish Square Holding v Makdessi (2015) asked whether a clause imposes a detriment on the contract-breaker out of all proportion to any legitimate interest of the innocent party. Other European legal systems allow courts to reduce excessive contractual penalties. A figure unrelated to any real loss is therefore worth less than it looks.
Service credits or no billing for the vacancy. With billing for days worked, you already do not pay for days without a specialist. You can go further and agree a credit against the vendor’s invoices for each day the replacement deadline is exceeded. It is proportionate and rarely disputed.
A right to terminate without notice. If the replacement deadline is exceeded by a set number of days, the client may end the role with immediate effect and source the specialist elsewhere. Often the most effective tool, because it gives the client a real way out.
A remediation plan for repeated breaches. If the vendor regularly misses SLA targets, the contract can require a remediation plan and, if it is not delivered, a right to terminate the whole framework agreement.
In practice a combination works best: no billing for the vacancy as a standing mechanism, a moderate service credit for missing the deadline, and a termination right after a longer overrun. The aim is not to earn money from penalties but to make meeting the deadline cheaper for the vendor than missing it.
How do you protect knowledge continuity when a specialist is replaced?
A fast replacement achieves little if the successor starts from zero. Continuity depends on clauses that apply throughout the engagement, not only at the moment of replacement.
Repositories, documentation and tickets on the client’s side. Code, technical documentation, decision records and task history should live in the client’s tools, not in the specialist’s private notes or the vendor’s systems. This is the simplest and most effective continuity clause.
A duty to document as you go. The contract can require the vendor’s specialists to document key parts of their work to the client’s standard. Not heavy documentation; architecture decision records and run instructions are enough.
Handover on planned departure. A list of items to hand over, sessions with the successor and confirmation from the team lead that handover happened.
Access and security. A procedure for revoking the leaver’s access and granting the successor’s, with deadlines on both sides. Late access can consume half the replacement deadline.
The best continuity safeguard, though, is not a clause but how the work is organised. If only one person holds the key knowledge of a module, every replacement will be expensive, whether that person is a contractor or an employee. Code review, task rotation and pairing in critical areas turn a departure into an inconvenience rather than a crisis.
These clauses overlap with the wider duties of both parties covered in our article on the rights and obligations of the parties in a body leasing contract. Read the two together: that one sets out the foundations of the contract, this one deals with its most sensitive mechanism.
What should the contract say about notice, scaling and exit?
The replacement guarantee protects continuity. Notice and exit clauses protect your ability to end the engagement when replacement does not help or needs change.
Notice per role. Staff augmentation should let you scale down with reasonable notice measured in weeks, not months. Short notice periods usually reflect the real flexibility of the model.
No exit charges. Check the draft for early termination fees, minimum volume commitments or charges for scaling down. With billing for days worked they have no justification.
A rate fixed for the engagement. A clause that the rate applies for the whole engagement and changes only by amendment. It prevents a replacement becoming a pretext to raise the rate for a role.
Regulatory requirements. If you are a financial entity, Article 30 of the EU Digital Operational Resilience Act (DORA, Regulation 2022/2554) requires contracts with ICT third-party service providers to include, among other things, service level descriptions, termination rights and minimum notice periods, and for critical or important functions also precise quantitative and qualitative performance targets and exit strategies with a mandatory transition period. Check with compliance whether IT staffing services sit in your register of ICT arrangements and shape the SLA and exit clauses accordingly.
If you are weighing how the billing model shifts risk between you and the vendor, read our article on whether a team subscription, time and materials or fixed price really shares risk.
What replacement terms does ARDURA Consulting offer?
Our replacement terms are published on our staff augmentation services and pricing pages, so you can compare them with other vendors’ proposals. We quote them word for word.
On the replacement guarantee: “ARDURA Consulting provides a 2-week replacement guarantee at no additional cost. If a specialist isn’t the right fit in the first 2 weeks, we replace them at no cost, and if a specialist leaves during the engagement, we guarantee a replacement within 5 business days.” In addition, all specialists pass our 5-stage vetting process, you interview and approve every candidate before engagement, dedicated account managers monitor performance continuously, and regular feedback loops ensure alignment. Our page on IT staff augmentation in Poland adds that within the first two weeks we replace the specialist with another vetted senior at no charge to you and absorb the onboarding overhead, so the replacement risk transfers to ARDURA Consulting.
On delivery time: we deliver vetted specialists within 2 weeks, and urgent placements in 48–72 hours from our pre-vetted pool.
On cost and exit: the daily rate covers the specialist’s salary, taxes, insurance, equipment and our management overhead; there are no recruitment fees, no setup costs and no termination penalties, and you pay only for the days worked, invoiced monthly. Rates are locked for the duration of the engagement. Scaling down requires 30 days’ notice, with no penalties.
Our specialist retention rate is 99%, and we hold a 5.0/5 rating on Clutch from 8 verified client reviews. Those numbers do not replace contract clauses, and that is the point: the terms a vendor publishes should appear in the contract you sign, with us or anyone else.
How should you negotiate these clauses, and where do you start?
Back to the procurement lead in Rotterdam. She does not need to invent a contract. She can lay this article’s checklist next to the vendor’s draft and ask five questions. Which events trigger a replacement? How many business days pass from notice to the successor’s first day? Who pays for overlap and the first days? What happens when the deadline is missed? Which metrics does the vendor report, and how often? Then she checks the result against the DORA requirements her compliance team applies.
Order matters. Start with deadlines and consequences, because they determine real protection. SLA metrics and review procedures matter, but they are easier to agree once the foundation is set. Keep proportion: a vendor that accepts concrete deadlines and reasonable remedies is worth more than one that signs any penalty because it does not intend to pay it.
Remember five things. First, a replacement guarantee has three layers: commitment, measure and consequence; lose one and it loses its value. Second, list the triggers, and make sure departure is covered for the whole engagement, not only at the start. Third, deadlines must be numbers, measured to the successor’s first working day, with the date of notice deciding the guarantee period. Fourth, a staffing SLA measures the staffing process, not system behaviour. Fifth, the most effective remedy is a combination of no billing for the vacancy, a moderate service credit and a right to terminate.
More on contracts and working with staffing vendors is in our staff augmentation hub. If you are preparing a contract and want to compare our terms with other offers, see ARDURA Consulting’s staff augmentation terms or book a call and we will walk through the replacement and SLA clauses with you.