How to Choose the Right CDMO for Your Biologics Program
Selecting a Contract Development and Manufacturing Organization (CDMO) is one of the highest-leverage decisions you’ll make in a biologics program. The right partner can de-risk CMC, compress timelines, and smooth the road from preclinical to clinic (and beyond). The wrong one can create churn, scope creep, and avoidable regulatory headaches. This guide offers a practical framework—plus a scorecard you can copy—to help you choose with confidence.
Start with Your Program’s Reality (Stage, Modality, Risks)
Before assessing vendors, align internally on what you actually need. Clear inputs lead to better RFPs and more accurate quotes.
Self-assessment checklist:
Program stage: Preclinical • Phase I • Phase II • Phase III • Commercial
Modality: mAb • Bispecific • Recombinant protein • Vaccine • ADC/bioconjugate • Other
Expected volumes & dose: Required DS/DP batch sizes (e.g., tox batch, FIHT, PPQ) and vial/syringe format
Markets & regulators: US (FDA) • EU (EMA) • UK (MHRA) • Other (PMDA, NMPA, etc.)
Known risks: Aggregation • Glycosylation • HCP burden • Potency drift • Stability • HPAPI/containment
Critical timelines: Target IND/CTA date, site audit window, fill-finish slot, PPQ/validation horizon
Questions to self
What are our must-hit dates on the clinical path (IND/CTA, first-patient-in, PPQ)?
Which CQAs are most likely to drive rework (e.g., aggregates, HCP, potency)?
Do we need integrated DS/DP or can we split with clear interfaces?
Essential CDMO Criteria (What to Evaluate and Why)
Use the criteria below to structure your long list, shortlist, and due diligence. For each, align on proof you’ll request and how you’ll score it.
Technical Breadth & Fit
A strong fit looks like a CDMO that routinely runs your host and modality at your intended scales, with clear decision trees for when to deviate from platform processing. Mature groups can show credible scale-down models that predict yield and CPP/CQA behavior, and for ADCs they’ll demonstrate linker/payload handling, DAR control, and aggregation mitigation under appropriate containment. Trouble often starts when partners say “we can do anything” but can’t produce modality-specific proof, overpromise on perfusion without acknowledging media/filtration risks, or force late resin/single-use changes due to inventory gaps.
Evidence to request
Two anonymized process overviews for similar programs (yields, titres, CPP/CQA learnings)
Unit operations list + available columns/resins; conjugation chemistries (for ADCs)
Facility/equipment list with line drawings; allowable excipients/solvents/surfactants
Questions to self
Do our DS/DP scales and formulation needs map to their installed base?
Are our CPPs/CQAs compatible with their platform defaults?
If ADC: do they meet our OEL and conjugation chemistry requirements?
Quality & Regulatory Track Record
What “good” looks like is a site with recent FDA/EMA/MHRA inspections showing no critical findings, a culture that closes CAPAs on time, and a tidy document hierarchy with trained users. You’ll feel it in how openly they discuss deviations and trending. Common pitfalls include defensive audit behavior, reluctance to share QA metrics, and fragmented QMS practices that vary by suite or site—classic signals of data-integrity risk later.
Evidence to request
Summary of external inspections over 24 months + CAPA closure timelines
Redacted deviation/complaint metrics (rates, time to close)
Data-integrity policy and audit-trail management overview
Questions to self
Does their inspection cadence align with our target markets?
Are we comfortable with their CAPA velocity and trending?
Would our regulatory story be stronger with this site on our dossier?
Process & Analytical Strength
A capable partner demonstrates DoE-driven development, clear control strategies, and complete method lifecycles—from development through validation and transfer—for potency, purity, identity, residuals (including HCP/DNA/Protein A), and stability-indicating assays. Weaknesses emerge when critical assays are outsourced with long lead times, comparability is an afterthought, or HCP strategy is vague (e.g., no path from platform to process-specific coverage when needed). For deeper reading on HCP analytics and method lifecycle, see our page on Analytical Services for Biologics.
Evidence to request
Redacted control strategy + process characterization example (fishbone, DOE matrices, PARs)
HCP strategy (platform vs. process-specific), 2D-DIGE/LC-MS support, spike/recovery data
Sample OOS/OOT investigation reports
Questions to self
Do we need process-specific HCP now, or is a platform ELISA sufficient?
Are our specs feasible with their standard methods?
Can they meet our stability and release timelines with in-house QC?
Scale & Capacity
The best partners are transparent about capacity maps, earliest credible DS/DP start dates, and campaign planning that includes engineering or bridging runs. They’ll show multiple scale ladders (e.g., 50 → 200 → 1,000 L) with consistent unit ops to avoid surprises. Red flags include hand-wavy “we’ll find a slot,” double-booking risks, and hidden turnaround constraints like resin reuse or cleaning bottlenecks.
Evidence to request
Next three quarters of available windows for DS/DP
Historical OTIF for comparable programs
Changeover turnaround times; clean-hold limits
Questions to self
Is our critical path DS- or DP-driven?
Do their scale ladders match our clinical volumes?
What’s our tolerance for slot slippage?
Tech Transfer Excellence
Strong TT shows up as standardized templates with RACIs, explicit data-package requirements, acceptance criteria, and smart use of engineering runs to de-risk PPQ. Pitfalls are under-scoped TT (leading to a cascade of change requests), and missing raw data that forces time-consuming reconstruction.
Evidence to request
Redacted TT plan and gap-assessment example
Risk register snapshot from a recent TT
Change-control flow and decision rights
Questions to self
How complete is our data package today?
What engineering work are we willing to fund to de-risk PPQ?
Do we need bridging studies for comparability?
Project Management & Communication
High-functioning CDMOs run crisp weekly ops, monthly steerco, and maintain living RAID logs and KPI dashboards (schedule adherence, quality events, cycle times). Weak programs feel “deck-heavy” but light on issue closure, or rely on PMs without real escalation authority into production and QC.
Evidence to request
Redacted Gantt, KPI dashboard, and RAID log
Escalation policy and response SLAs
Sample meeting minutes and decision logs
Questions to self
How often do we need decision-grade updates?
Which KPIs matter most for our execs?
What’s our escalation path if milestones slip?
Supply Chain & Materials Management
Good supply chains forecast long-leads, pre-qualify alternatives for resins/filters/single-use sets, and keep incoming QC tight. Problems arise when BOMs are finalized late, critical resins are sole-sourced, or DP components (stoppers, syringes, PFS) become last-minute blockers.
Evidence to request
Long-lead material plan and inventory buffers
Approved vendor list and dual-source strategy
Recent shortage mitigations and outcomes
Questions to self
Which items on our BOM are highest risk?
Do we have alternatives pre-qualified?
What expiry and storage constraints will hit our schedule?
Downstream Path to Clinic & Commercial
Choose a partner that can carry you from characterization to PPQ and into CPV, with a clear philosophy for edge-of-failure studies, multivariate analysis, and lifecycle monitoring. Development-only shops struggle here; lack of a post-approval change plan or CPV framework often leads to avoidable regulatory churn later.
Evidence to request
Two anonymized PPQ examples and associated control strategies
CPV dashboards and APR/PQR templates
DP stability designs and prior regulatory feedback
Questions to self
What is our end-game scale and does this site get us there?
Are we ready to fund PC/PPQ studies when needed?
How will we manage CPV and signal detection post-launch?
Pricing & Total Cost of Ownership
The right lens is assumptions, not day rates. You want milestone-based pricing tied to deliverables, a transparent assumptions log, defined out-of-scope items with a fair CR process, and upfront inflation/indexation terms. Pitfalls include teaser quotes that explode via frequent CRs and under-estimated analytical or stability burdens.
Evidence to request
Costed proposal with inclusions/exclusions
Examples of typical change orders and drivers
Payment schedule aligned to deliverables
Questions to self
Which assumptions would we dispute later?
What budget variance can we tolerate?
Are we valuing speed/certainty over lowest headline price?
RFP/RFI Question Bank (10 to copy)
1. Share anonymized CPV dashboards and PPQ outcomes for a similar modality/scale.
2. Provide last 24 months of external inspections and CAPA closure summaries.
3. Walk us through a recent tech transfer plan and gap assessment—deliverables, RACI, and success criteria.
4. Detail your upstream and downstream platform options and when you diverge from them.
5. Show the method lifecycle for potency and HCP (development → validation → transfer).
6. Provide standard governance cadence, decision rights, and KPI set for sponsors.
7. Describe your long-lead material strategy and recent mitigation actions.
8. Break down pricing with assumptions, escalation clauses, and typical out-of-scope items.
9. For ADC/HPAPI: provide OEL bands, containment strategy, and cleaning-validation approach.
Questions to self
Which questions best discriminate our shortlist?
What evidence would “prove it” for each answer?
Who on our side will evaluate the responses?
