A contract research organisation sells expertise and capacity. This guide helps you decide which tasks to buy, how to compare vendors, what to put in the contract, and which parts, like data capture, you can keep in your own hands.
Free sandbox · No credit card · 21 CFR Part 11 aligned
Study
Phase 2, 3 sites
Team
4 people
Key points
Step 1
The sensible first question is not which CRO but which jobs. Running a trial involves a bundle of functions: protocol and regulatory strategy, ethics submissions, site identification and contracting, monitoring, data management, safety reporting, biostatistics, medical writing and project management. A full-service CRO takes all of them. A functional service provider takes one or two. Some teams do most themselves and buy a single specialist service.
The right split depends on your experience and constraints. If your team has run trials before, has clinicians who know the sites, and needs mainly a data system, you may not need a CRO at all; see how to run a clinical trial without a CRO. If you are a first-time sponsor with an investigational product, a multi-country plan and tight timelines, buying experience for the regulatory and safety functions can be worth the cost. Neither answer is better in general. The mistake is buying everything because the alternative looks unfamiliar, or buying nothing because a vendor quote looked large.
Write a one-page task table before you speak to a single vendor: for each function, who does it, what experience you have, and what a failure would cost. That table becomes your request for proposal, and it keeps the conversation about your needs. The software stack for running a clinical trial page covers the tooling side of the same decision.
Task by task
A starting point for the worksheet, not a rule. Your experience and the study's risk change the answer.
| Function | Often bought when | Often kept in-house when |
|---|---|---|
| Regulatory and ethics submissions | First-time sponsor, multi-country, novel product | Single-site academic study with institutional support |
| Site selection and contracts | You lack site relationships | Investigators are your own colleagues |
| On-site and central monitoring | Many sites or high-risk design | Few sites, risk-based plan, strong site team |
| Data capture (EDC) and cleaning | You want it fully managed | You have a data manager or trained lead and a self-serve EDC |
| Safety reporting | Complex product, expedited reporting duties | Low-risk study with clear reporting rules |
| Statistics | No in-house statistician | You have a statistician or university support |
| Project management | No one has run a trial | A senior person can own the plan |
General guidance only. Confirm regulatory duties with a qualified adviser.
Step 2
Every proposal will say the right things about quality and experience. To separate them, ask for specifics. Has the team run studies in your indication and your countries? Who exactly will be on your study, and will they be the people who presented the proposal? What is turnover like on projects of your length? Can you speak to a client with a comparable study? Ask for a reference who will give a candid answer, and ask what went wrong on a past study and how it was handled.
Look hard at the assumptions behind the price. A budget is built on a stack of assumptions: number of sites, number of participants, visits per participant, length of recruitment, number of monitoring visits, number of queries per page, number of protocol amendments. If the real study differs from the assumption, the price changes. Ask what is included in the base price, what triggers a change order, and how those changes are priced. Compare proposals line by line on the same assumptions, using your own clinical trial budget template.
Check the quality system and the inspection history in general terms: how do they handle deviations, training and corrective actions, and what is their approach to risk-based monitoring? The risk-based monitoring and good clinical practice entries describe what a modern approach looks like. Be wary of any provider unwilling to share their standard operating procedures for review, subject to confidentiality.
Questions to ask
You want written confirmation that you own your data and can export it, with the audit trail, at any time.
Ask for the process and a worked example. Amendments are common; the price should not be a surprise.
Named people, with their current workload.
Look for access to live enrolment and data-quality views, not only a monthly slide deck.
Look for a defined escalation path and what it costs. See the recruitment playbook.
Ask about timelines, formats and who signs off.
Delays on the sponsor side are a common cause of cost growth; understand your obligations.
Termination terms and transfer of data and documents if the relationship ends.
Build a study in the free sandbox and see what running data capture yourself actually involves. Every feature, no credit card, pay only when live.
Step 3
The contract decides who carries what risk. Start with scope: a clear list of services, deliverables and assumptions, attached as a schedule. Define the change process in plain terms: what constitutes a change, who approves it, how it is priced and how quickly. Ask for rate cards for common changes so that an amendment does not start a negotiation from scratch. Define data ownership and access: you should own the study data and documents and have the right to exports and audit trail information throughout, not only at the end.
Add performance expectations that matter to you: enrolment reporting, timelines for query resolution, quality metrics, and what happens if they are missed. Define transfer provisions in case you change provider mid-study; how to switch EDC vendors mid-study shows why that is a painful moment. Confirm responsibilities under GCP: sponsors can delegate tasks to a CRO, but the sponsor remains ultimately responsible for the trial. The ICH E6(R3) guide covers oversight expectations. This is general information, not legal advice; have the contract reviewed by a qualified professional.
Finally, keep your own file. Record why you chose the provider, the qualification evidence, and the oversight plan. Oversight does not end when the contract is signed.
The data capture decision
A middle path
Many sponsors end up with a hybrid: a CRO or consultants for regulatory, safety or monitoring, and the sponsor holding the data system. This keeps the data, the audit trail and the study build in one place that you control, while still buying experience where you lack it. It also makes a later change of provider less painful, because the system of record does not move.
It requires being clear about responsibilities, since a CRO working in your EDC needs the right access, and you need a process for approving forms and changes. Capture has role-based access, site coordinator portals and exports by site, and enterprise customers can get dedicated support and documentation to support sponsor validation. For the CRO side of the market, best EDC for CRO small studies and EDC for CRO-independent trials show how each party uses the system. If you are an early-stage founder, read running your first clinical trial alongside this guide.
Whichever model you pick, make the decision on evidence. Ask each CRO to show you the system it would use on your study, build a small version of your protocol in a sandbox yourself, and compare. The cost of looking is low, and the cost of choosing wrong mid-study is high.
List the functions you need to buy, request proposals against the same assumptions, compare named team and experience in your indication, examine what triggers change orders, check data ownership and EDC access, and take references. Decide task by task rather than all-or-nothing.
Not necessarily. A CRO supplies expertise and capacity. Teams with clinical and operational experience often run studies themselves, sometimes buying single services such as monitoring or safety. Match the choice to your experience and the study's risk.
Defined scope and assumptions, how changes are triggered and priced, data ownership and access, performance expectations, transfer provisions and responsibilities under GCP. Have a qualified professional review it; this is not legal advice.
It determines who owns the study build, how you access data and the audit trail, how amendments are handled and how hard it is to switch provider. Ask which EDC will be used and what access and exports you get.
Yes, many sponsors do. With a self-serve platform such as Capture, your team keeps the study build, data and audit trail while the CRO provides services like monitoring or regulatory support. Agree access and approval roles in the contract.
Under GCP, sponsors may transfer tasks to a CRO but remain ultimately responsible for the trial. Keep an oversight plan. Confirm specifics with your adviser.
They do different things, so they are not a like-for-like comparison. Capture is a self-serve data platform: building and testing in the sandbox is free, and you pay only when you go live. A CRO sells services on top of a system.
Keep exploring
How to run a clinical trial without a CRO
The self-run path in detail.
EDC for CRO-independent trials
Data capture without a vendor build.
Best EDC for CRO small studies
For CROs running small studies.
Clinical trial budget template
Compare proposals on equal assumptions.
Build vs buy EDC
Another make-or-buy decision.
Running your first clinical trial
The startup roadmap.
Build a study in the free sandbox before you sign a contract. No credit card, no time limit, pay only when you go live.