Choosing a CDMO: the 7 questions you should be asking before you sign
For most biotech sponsors, the choice of a Contract Development and Manufacturing Organisation (CDMO) is one of the most consequential decisions of the programme. Once a contract is signed, the sponsor has effectively handed over a significant share of the technical, quality and regulatory destiny of the asset. The decisions that the CDMO takes on process parameters, documentation maturity, analytical method transfer and deviation management will surface in the IND or IMPD dossier, in the FDA 483 observations if an inspection occurs, and in the comparability exercises that follow. For early-stage biotech sponsors, it is one of the decisions that weighs most heavily on the first years of development.
The stakes are not abstract. FDA inspection activity has picked up again since the pandemic years, and inspection findings at manufacturing sites have followed. Capacity, meanwhile, is uneven: sterile fill-finish remains constrained, and early-phase sponsors can still wait a long time for a slot at a biologics CDMO. Against this background, a superficial selection process no longer offers the margin for error that it once did.
And yet, in practice, CDMO selection is often driven by three factors that are easy to measure and insufficient on their own: price, slot availability, and the quality of the presentation given during the capability visit. The deeper questions, the ones that actually determine whether the partnership will work over the next two to three years, are asked less often, and sometimes not at all. As discussed in our introduction to CDMO partnerships, the evaluation needs to go further than a checklist of technical capabilities.
The following seven questions are the ones that, in our experience, tend to separate a productive CDMO partnership from a costly one. They are not exhaustive. They are the questions that, when asked properly and answered honestly, tend to surface the information that matters.
1. "Show us the last three programmes you ran that were similar to ours. What went well, and what didn't?"
A capability presentation shows what the CDMO can do in principle. A programme history shows what the CDMO has actually done.
The right version of this question is specific. Similar modality, similar scale, similar phase of development, ideally similar analytical complexity. The sponsor is not looking for a sanitised success story; they are looking for how the CDMO talks about its own work. Does the CDMO describe technical issues that arose, how they were investigated and how they were resolved? Does it describe deviations with dates, root cause analyses and CAPAs? Does it volunteer the things that did not go to plan?
CDMOs that answer this question well demonstrate a quality culture. CDMOs that deflect, generalise or provide only success narratives reveal something the capability deck does not show. This matters because, at some point in the programme, something will not go to plan. The question is not whether deviations will happen, but how they will be handled when they do.
2. "Walk us through your tech transfer process, step by step, including the engineering run and the analytical method qualification."
In our experience, technology transfer is a frequent source of delay in early clinical programmes, and analytical method transfer is often where it starts: if the receiving CDMO cannot reliably measure what the sending laboratory measured, the process cannot be controlled, and the data cannot be compared.
The question is not whether the CDMO has a tech transfer procedure. All CDMOs do. The question is what the procedure looks like in practice. A robust tech transfer for a biologic or an advanced therapy typically includes:
A structured intake phase during which the sending laboratory's full process description, including "failed" or non-optimal data, is transferred to the CDMO.
A gap analysis against the receiving site's existing platforms, equipment and raw materials.
A defined analytical method transfer workstream, completed before manufacturing begins, with qualification criteria agreed in advance.
At least one engineering run at near-GMP conditions before the first GMP campaign, designed to surface process sensitivities before irreversible commitments are made.
When the CDMO's description skips over any of these elements, particularly the engineering run and the analytical method qualification, the programme is at a higher risk of discovering problems at the wrong time. Another warning sign is a transfer plan that assumes discovery protocols or lab notebooks can be used as they are.
3. "What is your documentation maturity at our development phase?"
Documentation maturity is one of the least glamorous aspects of CDMO evaluation and one of the most consequential. It is what separates a GMP manufacturing activity that produces a usable investigational medicinal product (IMP) from one that produces a batch with unanswerable questions attached.
At early clinical phases, sponsors should expect a CDMO to demonstrate, at a minimum:
Master Batch Records and Executed Batch Records that are structured, traceable and audit-ready.
A quality management system with defined deviation, change control, CAPA and OOS procedures that are actively used rather than formally held.
Analytical method documentation suitable for phase-appropriate validation or qualification, with the ability to evolve as the programme progresses.
Material traceability from starting materials through to drug product release, including critical reagents, single-use systems and primary packaging components.
Documentation maturity varies significantly across the CDMO landscape, and it does not always correlate with the size or reputation of the organisation. Smaller CDMOs can be exceptionally well-organised for a given modality, while larger CDMOs can be uneven between business units. The relevant question is not whether the CDMO has documentation, but whether the documentation it produces will hold up in a regulatory inspection, in a due diligence by a partner or investor, and in the comparability exercise that almost every programme eventually runs.
4. "Can we see a recent inspection report, with the observations and the responses?"
This question is often considered sensitive, and some CDMOs will decline to share a full inspection report. That in itself is informative.
Mature CDMOs that take their compliance posture seriously are generally willing to discuss their inspection history in substantive terms. They will summarise recent regulatory inspections, describe the observations received, explain the remediation actions taken, and show how those actions were verified and closed. They treat inspection history as a demonstration of operational maturity, not a liability to be hidden.
Two patterns should raise concern. The first is a CDMO that presents itself as having no observations or only minor ones across multiple years; this is unusual in a facility that is being actively inspected, and may indicate limited inspection exposure or selective disclosure. The second is a CDMO that cannot describe how specific observations were resolved, which suggests that the remediation did not deeply penetrate the organisation. Both patterns are worth investigating further before committing to a multi-year relationship.
For sponsors operating across multiple regions, it is also worth asking about the CDMO's experience with health authorities relevant to the programme, including the FDA, EU national inspectorates, the MHRA and, where applicable, the PMDA and NMPA.
5. "Who will actually work on our programme, and how stable is that team?"
CDMO proposals are typically presented by senior leaders. The people who will actually execute the programme are usually not in the room.
This gap matters. The success of a tech transfer and a first GMP campaign depends heavily on the experience, continuity and authority of the operational team assigned to the programme. Sponsors who ask to meet the project manager, the process development lead, the analytical lead and the QA contact before signing typically get a much more accurate picture of the CDMO's capacity to deliver.
The follow-up question is about stability. CDMO staff turnover has been a persistent challenge across the industry, and a programme that loses its key contacts mid-way through a manufacturing campaign often loses weeks of time recovering the context that departed with them. Asking the CDMO how it protects the continuity of its project teams, and what happens when a key person leaves, provides useful information about how the organisation actually operates.
This is also the right moment to ask how many programmes the assigned team is running in parallel. A CDMO team stretched across multiple high-priority transfers rarely delivers the level of attention that a first-in-human programme requires.
6. "What happens if we need to change something mid-programme?"
Every clinical programme changes. Process parameters evolve as development data accumulates. Raw materials are qualified and sometimes replaced. Analytical methods mature. Scale changes. Container-closure systems are refined. A CDMO's ability to absorb these changes without losing time is directly proportional to the quality of its change control system and the flexibility of its contractual and quality framework.
The relevant questions are practical:
How does the CDMO handle a mid-campaign change request? What is the typical cycle time from request to implementation?
How are change controls documented, and how do they flow into the regulatory CMC documentation the sponsor will eventually file?
What is the CDMO's experience with comparability exercises following a process change, and how are those comparability data packages structured?
How does the contract handle scope changes, additional batches, extended storage, or modifications to the analytical test plan?
For biologics and advanced therapies, this line of questioning is particularly important. Minor process changes can have measurable consequences for product structure or function, and a CDMO that treats change control as an administrative step rather than a technical one will not produce the data needed to support comparability arguments later in development.
7. "How will our programme be positioned against your other clients when capacity is tight?"
CDMO capacity is uneven across modalities. Sterile fill-finish remains constrained, and availability in other segments, such as plasmid and viral vector manufacturing, can change quickly. Where capacity is tight, slot allocation is no longer purely contractual; it is also political.
A sponsor with a Phase 1 programme competing for capacity against a CDMO's larger Phase 3 or commercial clients needs to understand, before signing, where they sit in the priority hierarchy. This does not mean that smaller sponsors should avoid larger CDMOs. It means that the conversation about priority should happen explicitly, not implicitly, and that contractual protections around slot reservation, minimum batch commitments and rescheduling notice should be negotiated with full knowledge of the CDMO's portfolio dynamics.
Two related questions are worth asking at the same time. The first is about the CDMO's own operational resilience: its financial position, its parent company structure, its recent site changes, and its capital investment plans. CDMO restructurings and site closures have become more frequent, and sponsors who assumed long-term continuity have sometimes found themselves scrambling for alternative capacity. The second is about contingency: if a critical step at this CDMO became unavailable, what is the realistic fall-back, and how quickly could it be activated?
What sits behind these questions
These questions are not a scoring template. They are designed to change the nature of the conversation with a potential CDMO. When a sponsor asks the first three of these questions seriously, most CDMOs respond differently: less sales pitch, more substantive technical dialogue. The conversation becomes a mutual evaluation rather than a one-way presentation.
The underlying principle is that CDMO selection for biotech programmes is not a purchasing decision. It is an extension of the sponsor's quality system and a structural commitment to a particular technical trajectory. The right partner is not always the most experienced or the most expensive; it is the one whose capabilities, documentation maturity, operational culture and development trajectory align with the sponsor's programme.
Several patterns tend to distinguish successful selections from unsuccessful ones. The evaluation is structured around the specific requirements of the programme, not around generic capabilities. Multiple CDMOs are compared, but only after an explicit set of criteria has been agreed by the sponsor's technical and regulatory functions. The sponsor's team includes, or has access to, someone with direct prior experience of multiple CDMO relationships, ideally across different modalities and regions. The commercial negotiation is sequenced after the technical evaluation, not before. The contract and quality agreement are built to reflect the realities of a development programme that will change, rather than as a transactional document for a fixed scope. Our article on risk assessment in drug development provides a useful companion view on how to structure this kind of evaluation.
For sponsors going through this process for the first time, the absence of this structured approach is a common source of avoidable risk. Bringing in someone who has already run several CDMO selections is one way to close that gap.
The underlying shift
For a time, many sponsors could select a CDMO on capability and price and absorb the consequences later. Inspection activity has since picked up, several CDMOs have restructured or closed sites, and capacity has become uneven across modalities. That margin has narrowed.
Sponsors who continue to approach CDMO selection as a primarily commercial decision are operating on an outdated model. The CDMOs that sit at the centre of successful programmes today are those that function as genuine extensions of the sponsor's technical and quality organisation, with the documentation, the culture and the operational stability to support that role. Identifying such partners requires a different kind of evaluation than a capability visit and a price comparison.
For biotech sponsors entering or deepening a CDMO relationship, the seven questions above are a starting point. The answers, taken seriously, typically reveal more than any data room or capability deck.
If your programme is evaluating or managing a CDMO relationship, 3Biotech helps sponsors structure the selection process, conduct technical and quality due diligence, and set up the governance framework that protects development momentum.
Learn more about how 3Biotech supports biotech companies through our CMC strategy and development services.
