Over the four posts in this series, we have built a complete framework for evaluating a CDMO for advanced therapy development. We looked at the manufacturing landscape and what the FDA data tells us about CMC risk. We covered modality-specific expertise and the structural importance of integration across process development, analytics, and GMP. We went deep on process development philosophy, materials strategy, and facility readiness. And we closed with the partnership model and the questions that surface what a capability presentation rarely does.
This final post is about turning that framework into a decision. By the time you finish a CDMO evaluation, you will have absorbed a lot of information from a small number of organizations. The challenge is not gathering more data. It is organizing what you have learned in a way that makes the right partner visible.
What follows is the framework we recommend: a scorecard for synthesizing the evaluation, a sequencing model for the evaluation process itself, and a short list of decision pitfalls worth avoiding.
The seven criteria below cover the dimensions developed across this series. The weighting reflects our view of relative importance for advanced therapy programs at the IND-enabling stage. Adjust weighting for your specific program context: a viral vector program will weight different criteria than a stem cell program; a tech transfer engagement will weight different criteria than first-in-modality process development.
|
Criterion |
Weight |
What Strong Looks Like |
What Should Concern You |
|
Modality expertise |
High |
Specific named programs in your modality taken through IND. |
General capability claims with no concrete program references. |
|
Development integration |
High |
Same scientists from PD through GMP; analytics in parallel. |
Sequential handoffs; documentation-as-transfer. |
|
Process philosophy |
High |
CPPs defined in RUO; vessel transitions evaluated up front. |
CPPs deferred to GMP; scale-up treated as logistics. |
|
Materials strategy |
Medium-High |
Clinical-grade equivalents qualified in development. |
Materials revisited at tech transfer, not in development. |
|
Facility & QMS |
Medium-High |
Mature QMS, sponsor-audited, FDA-engaged. |
QMS still being built, no sponsor audit history. |
|
Partnership model |
High |
Direct scientist access; proactive communication. |
Account leads as primary interface; presentation team ≠ program team. |
|
Cultural fit |
Medium |
Candor about prior challenges; clear reasoning. |
Defensive about programs that did not go perfectly. |
Use the scorecard as a conversation tool, not a spreadsheet. The signals in the right two columns are the kind of evidence you should be able to point to from your evaluation conversations. If you cannot, you do not yet have enough data on that criterion.
|
How you sequence the evaluation matters as much as which criteria you assess. A common pattern, capability deck first, then technical conversation, then pricing, then references, tends to give CDMOs the format they are most prepared for. A more discriminating sequence puts scientific conversation earlier and gives the evaluation more signal per hour invested.
Start with a structured technical conversation. Bring two or three of the modality-specific questions from Part 2 of this series, and one or two of the process philosophy questions from Part 3. Ask to speak with the scientists who would be assigned to your program, not the BD lead. The depth and texture of this conversation is the single highest-leverage data point in the evaluation.
Once the scientific conversation has established that the CDMO is a credible candidate, move to capability. Walk through facility, equipment, QMS, regulatory history, and team structure. Treat this as confirming what was suggested by the scientific conversation rather than as a separate evaluation track. Capability without scientific depth is not enough; scientific depth without capability is not enough either.
Ask the partnership questions from Part 4 directly. How does the team handle disagreement? Who is in the room during a GMP investigation? How do they describe a program that did not go perfectly? This stage is about how the CDMO will operate, not what they can do.
Save references for last, and ask specific questions when you call them. Generic reference checks rarely produce useful information. Ask references about the specific dimensions where you still have uncertainty: how the CDMO handled a technical setback, whether the team you met during evaluation stayed on the program, how regulatory questions were navigated.
A few patterns recur in CDMO selection decisions that go badly. They are worth naming because they are easy to fall into:
Mistaking capability for fit. A CDMO can be highly capable and still not be the right partner for your program. Capability is a necessary condition. Modality fit, integration, and partnership quality are what determine whether the capability translates to program outcomes.
Optimizing for the lowest stage gate cost. Cost matters, but the cost of a program that runs eight months long because of a CMC question is far higher than the cost differential between two CDMOs. Evaluate cost in the context of expected timeline and regulatory risk, not in isolation.
Underweighting program team continuity. The team you meet during evaluation may not be the team that runs your program. This is one of the most consequential and most under-asked questions in CDMO selection. Surface it explicitly.
Skipping the difficult-program question. Every experienced CDMO has worked through programs that did not go as planned. The way they describe those experiences is among the most informative signals in an evaluation. Skipping the question, or accepting a polished non-answer, removes one of the few windows into how a partner behaves under pressure.
Confusing enthusiasm with depth. Enthusiastic generalities are easy. Specific, scientific, candid answers are harder to produce and more informative. Calibrate on the texture of the answers, not the energy of the conversation.
You have enough to decide when, for each of the seven scorecard criteria, you can point to a specific moment in the evaluation that informed your view. Not a general impression; a moment. A scientist's answer to a specific question. A facility tour observation. A reference call detail. A way the CDMO handled a follow-up request.
If you cannot point to those moments, the next evaluation step is not another presentation. It is a targeted conversation, with the right people, about the specific criteria where you do not yet have evidence.
The decision itself, when it comes, tends to feel less like a tradeoff and more like recognition. The right partner for your program will have shown you who they are throughout the evaluation. Trust what you have seen.
The Artis BioSolutions CDMO Selection Guide pulls this entire series into a single resource. It includes the full framework, all of the evaluation questions, the scorecard above, and an expanded FAQ covering the questions program leaders most frequently bring to us. It is designed to be the resource you keep open during your next round of CDMO conversations.
Reach out to [contact@artisbio.com] to request a copy. If you are actively evaluating CDMOs, we are also happy to talk through what the framework looks like applied to your specific program and modality.
Artis BioSolutions partners with advanced therapy developers across process development, analytical development, GMP manufacturing, and IND-enabling CMC support. Our integrated team draws on deep drug development expertise across process development, analytical science, quality, and regulatory affairs, working as a true extension of your organization at every stage of the program.
If your program is in cell therapy, gene therapy, viral vector, or complex cell system development, we would welcome the conversation about whether the framework in this series, applied to our team, our facility, and our scientific depth, supports your program's needs.
Connect with the Artis BioSolutions team
inquiries@artisbiosolution.com | artisbiosolutions.com