What Happens When Products Fail?
In prior posts about the new product development process, I've focused on the steps when everything goes smoothly along the way. But what happens when things go wrong? Today, I'd like to talk about what happens when a batch fails or issues arise with the finished product. This could include product separating in the kettle, failing quality testing prior to release to market, or complaints from consumers.
Failure During Scale-Up
When working on a new product, the responsibility of the process engineer as part of the scale-up work is to understand the failure points of the formula or process. This is a normal part of the laboratory, pilot, and trial batching performed during the cosmetic product development process. It is important to understand the potential failure points so controls can be implemented to avoid issues during manufacturing.
However, there may be limited opportunity to test failure limits at the larger scales. It is quick and easy to make a laboratory scale batch and test out different mixing speeds, order of chemical additions, or processing temperatures and see what the impact on the final product may be. When making batches at the pilot or trial scale, there is an increased cost of operator time, larger quantities of raw materials, and kettle availability where failure may not be an option. Additionally, these batches may be used to produce samples for testing packaging compatibility, formula safety, or even clinical trials. These tests cannot be run on samples that do not meet the formula specifications or target aesthetics.
For new product development, the result of a failed batch at the pilot or trial scale is typically to investigate the cause of the failure, correct the issue, and repeat the batch. This can be difficult to do without impacting the project timeline depending on the availability of new raw materials or the understanding of the mechanism of the failure. Therefore, it's important to do a robust investigation of the failure and root cause analysis.
The worst case scenario for a project would be a failed batch where no root cause can be identified and the issue cannot be corrected. This could lead to the product launch being cancelled or delayed. This has a broad impact on the business because new launches are communicated to retailers and sales targets are projected well in advance. Any last minute portfolio changes or delays can hurt relationships with company business partners.
For cosmetic manufacturing sites, it is required by the FDA to have a quality system in place to investigate failures and implement corrective and preventative actions (CAPAs). Each site will have different processes and requirements, but the general procedure is the same:
1. Define the problem The problem should be defined as clearly and specifically as possible, including details about who, what, when, and where. For example: "Batch 123 of product ABC failed quality testing for pH. The batch was made on 10-Nov-2019 in Kettle 1 and tested by quality on 11-Nov-2019. The specified range is 4.0 to 5.0 and the pH for the batch was 3.8. The pH was retested on 12-Nov-2019 by a different analyst and the failure was confirmed."
2. Collect all relevant documentation Each batch will have a lot of documentation to be compiled and reviewed. There should be raw material pre-weigh information on the weighed quantities and lot numbers, a batch card which documents all steps of the recipe, quality testing results with the test method used, information on who completed all steps in the process, and more. For product that was filled into final packaging, there will also be information on the storage time prior to filling and all the settings on the filling line. The goal is to compile facts, not inferences.
3. Determine the most probable root cause There are a variety of tools that can be used to complete root cause analysis. In my experience, I have used the five whys and fishbone diagrams most often. The five whys is a simple method of asking why? over and over until the root cause is identified and is best for simple issues. A fishbone diagram is a good tool for brainstorming ideas to guide the investigation work for complex issues. Each "bone" is a category and traditionally they are referred to as the six "M's" - Machine, Method, Material, Measurement, Mother Nature (or environment), and Man (or people). You draw a diagram with six branches and document all potential failure modes by drawing additional branches off the six categories.
4. Identify and implement corrective and preventative actions Once the most probable root cause is identified, corrective and preventative actions are identified to address the issue and keep it from recurring in the future. This could be as simple as redefining a manufacturing process step to implement additional controls or as complicated as reformulating the product.
5. Make product disposition decision The final step is to determine what to do with the failed batch. This will be based on the failure and the corrective action. Using the original example of the out-of-specification pH result, if the root cause was determined to be inherent raw material variability and the corrective action was to widen the pH specification limits, the batch may be deemed acceptable. If the out-of-specification result was based on an addition of an incorrect raw material, the batch will be thrown away.
6. Document all findings in a report
Once all steps are complete, the data is summarized and documented in a report. This report will be approved by the quality team and stored with other batch documentation. If the FDA audits the site, investigation reports are typically included in their review.
Failure After Launch
So far, we've examined failures during the launch process. But failures can occur after a product has been successfully scaled up and launched too. In this case, the project team that launched the product has typically disbanded and moved on to other project work. There will likely be a process engineer who works closely with manufacturing who will investigate the failure in these cases.
The investigation steps are the same as outlined above. However, with a product that has already been launched, there is less leeway when it comes to revising the specification. Prior to launch, there has not yet been a finalized specification. After launch, the product has already been released to the market and used by consumers. Changes to the specification limits that could be noticed by consumers will not be supported unless there is a clear technical justification. It is more likely that a batch with a test result outside the specification limits will be thrown out.
In cases where the batch will be scrapped, it is still very important to complete the investigation and root cause analysis to keep the issue from arising again in the future. In these post-launch investigations, there will be additional documentation and business partners involved. The supply chain organization will need to understand the impact on inventory and the manufacturing team will need to understand how the failure could be prevented for the impacted batch or other products through additional personnel training, procedure updates, or engineering changes.
Conclusion Understanding failures during scale up and after the launch of a product is critical. Robust investigations help the technical team address failure modes and prevent future issues. Additionally, detailed technical justifications may allow batches to be released to market by proving that the batch is comparable to a control batch or that the original testing result was flawed. Performing investigations is an important part of manufacturing and is also an interesting way to apply the cosmetic chemist's knowledge of the formula and manufacturing process. The learning from investigations can also be applied to future product launches to prevent similar issues for new products.
Do you have any questions about root cause analysis or investigations? Let me know in the comments below!