Formulation and process · Process development
“Our Cpk is too tight, and we don't know which parameter drives the variation.”
Process optimization and process capability (Cpk, Ppk)
A process that only just meets specification costs batches and deviation reports. CovaSyn analyses your process and batch data, calculates Cpk and Ppk, shows which parameters drive variation and derives a design space under ICH Q8. You receive a report that your expert reviews and signs off.
Starting point
When you need this
- Process capability for a critical quality attribute is not sufficient, and the cause is unclear.
- The submission needs a design space with justified ranges (NOR, PAR) under ICH Q8.
- PPQ is coming up, and you want to know beforehand whether process and data hold up.
- Values shift slowly in routine production, and the control chart does not yet show a clear violation.
Result
What you get
Cpk and Ppk with interpretation
Short-term and long-term capability per quality attribute, with a check of the distribution assumption and whether location or spread is the problem.
Drivers of variation
Process parameters ranked by influence, with sensitivity and indications of causal relationships in the data.
Design space and robustness
Ranges in which the process meets specification with margin, robustness against fluctuations and proposals for established conditions under ICH Q12.
Drift and control charts
Detection of gradual shifts across batches, control charts and an assessment of PPQ readiness.
Three ways
Yourself, together, or done for you
Run it yourself
Draw the molecule in the sketcher or use the tools through the CovaSyn MCP directly in Claude, ChatGPT or Cursor.
Go to MCPLearn it together
In a workshop your team learns how to answer questions like this with AI, cleanly and traceably.
See workshopsDone for you
We run the evaluation with our tools and deliver the finished report. Your expert reviews and signs off.
See Done-for-You
CovaSyn
The tools behind it
Deterministic calculations instead of guessed answers: the same tools we use for Done-for-You work are available to you directly.
Covaopt
Process models, multi-objective optimisation, robustness, design space, established conditions, risk ranking, causal analysis, drift, Cpk and Ppk, and PPQ assessment.
Covadoe
Experimental designs to study open parameters in a targeted way, response surfaces and design space from DoE data.
Covastab
Control charts and batch variability for monitoring in routine operation.
Audience
Who it is for
For process development, CMC and manufacturing teams at CDMOs, generics manufacturers and specialty chemicals companies who need to make a process capable, secure it or justify it for submission.
What the computation does not replace
- A model is only as good as the data. Few batches or a narrow parameter range limit what can be derived with confidence.
- Cpk assumes a stable process in statistical control. If it is not, we show that instead of calculating a number that means nothing.
- A derived design space has to be confirmed experimentally before it goes into a submission.
Insights
Read on
Cpk vs Ppk: In Control Is Not the Same as Capable
Process capability for pharma: what Cpk and Ppk actually measure, why a stable process can still miss target, and a worked CovaOPT example with real SPC.
Read articleDesign Space, NOR and PAR: ICH Q8 Explained
ICH Q8 design space, NOR and PAR defined precisely, with real CovaDOE and CovaOPT tool output showing what a tighter spec costs you in operating window.
Read articleWhich CPP Actually Drives Your Variability
Critical process parameter sensitivity: how Morris, Sobol and SHAP differ, and a worked CovaOPT example ranking temperature against the rest.
Read articleQbD Process Development: Screen, RSM, ML in One Run
A QbD process development workflow end to end: DoE screen, response surface, ML model on one 29-run Pd coupling, with every number from real CovaDOE output.
Read article
Questions
Common questions
What is the difference between Cpk and Ppk?
Cpk uses within-subgroup variation and describes what the process can do in the short term. Ppk uses the overall variation across all data and describes what it actually delivered. If Ppk is well below Cpk, the process varies between batches or over time, and that is where CovaSyn starts the root cause analysis.
What to do when Cpk is too low?
First establish whether location or spread is the problem: a shifted mean can often be corrected through a set point, while too much spread needs the driving parameters. CovaSyn ranks the parameters by influence and shows where tighter control has the largest effect.
How do I define a design space under ICH Q8?
A model built from DoE or process data shows in which parameter ranges all critical quality attributes stay within specification, with a margin for model uncertainty. CovaSyn calculates design space and robustness and proposes NOR and PAR for your team to review.
How can I detect process drift early?
Control charts only flag violations once limits are crossed. A drift analysis across batch history picks up gradual shifts earlier. CovaSyn combines both and shows which parameter is linked to the shift.
Who can analyse our process data for us?
CovaSyn does this as Done-for-You: you describe the process and question in the intro call, we agree on data and scope, and you typically receive the report within 5 working days once we have your data. Handled confidentially from the start.
Can I calculate Cpk and design space myself?
Yes. The process tools are available through the CovaSyn MCP directly in your AI assistant, for example in Claude, ChatGPT or Cursor.
Formulation and process
Related problems we solve
Solve a solubility problem before you run out of material
Our API won't dissolve properly, and we have been trying solvents for weeks.
See solutionCrystallization process design: cooling or antisolvent
Our crystallization yield is too low, and we don't know whether to cool or use an antisolvent.
See solutionDesign of experiments (DoE): from design to analysis
We have too many factors and too little material, and nobody on the team is sure which design fits.
See solution
Next step
Tell us about your process problem.
30 minutes on Microsoft Teams with the CovaSyn team. If it is not a fit, we will say so.