Biopharma Benchmarking Unveils Performance Variance

Nov. 8, 2017
Performance gaps suggest that it’s time for biopharma manufacturing companies to focus on operational excellence

The complexity of biopharmaceutical manufacturing has made operational excellence a relatively low priority to date, with manufacturers focused primarily on delivering an adequate supply of quality product. As the industry grows and evolves, however, the focus on operational excellence is increasing, and manufacturers are beginning to look at their peers to understand best practices and their own performance potential. As they do, McKinsey’s proprietary Pharma Operations Benchmarking service (POBOS Biologics) reveals notable performance variations among biomanufacturing sites, reflecting the immaturity of these operations. These performance gaps suggest that biomanufacturing companies should take a good look at the way they run their operations and consider whether it is, indeed, time to step up.


Biopharmaceutical manufacturers have dealt for some time with their products’ complex and unstable production processes and relatively low yields. Securing product delivery at sufficient quality has historically been considered challenging enough, therefore, without taking the risk of pursuing production improvements or a transfer to better facilities. Not surprisingly, it is accepted in the industry that variation in output, yields, productivity and quality is simply inherent to biopharma manufacturing. Operations are run at different levels of effectiveness (for example, costs, labor productivity and capital productivity), with technical performance varying as well. As a result, management’s focus in biomanufacturing to date has — justifiably — been on supplying the market, rather than improving established operations.

Today, the landscape of the industry is changing. Biosimilars are becoming a reality, making it more difficult to command significant price premiums for biopharmaceuticals, particularly in areas in which innovation may become more difficult, such as in inflammation treatments. Yet the biopharma industry is still more profitable than traditional pharma and has grown steadily for a number of years. In fact, the share of cost of goods (COGS) sold attributable to biomanufacturing in Big Pharma is increasing steadily. Where biomanufacturing was once a minor diversion for pharma’s technical-operations organizations — generating a limited share of total costs — many Big Pharma players today have, or aspire to have, a substantial part of their operations in biopharmaceuticals. Simultaneously, biomanufacturing is becoming increasingly industrialized, moving steadily from the frontiers of science into a new manufacturing mainstream.

As the industry changes, executives in biomanufacturing debate the potential for true performance improvement in their operations. Their expectations range from quality improvements and multiproduct flexibility to faster cycle times or throughput and an enhanced cost position. As they pursue these enhancements, they look to understand the true potential of their manufacturing sites, addressing a broad set of performance dimensions — such as process robustness, capacity utilization and lead time — that are as important as, or more important than, productivity itself.

As a result, there is already a strong sense that the industry is moving in the right direction, with some players beginning to take steps to achieve both technical and operational excellence. These players are following a path similar to the one taken several decades ago by a number of chemical active pharmaceutical ingredient (API) manufacturers, moving one step at a time toward more effective operations. Some even find themselves ahead of the curve, having built, or begun to build, operational and technical expertise that puts them at the forefront of the biopharma industry. They are operating multiproduct facilities at a high level of utilization, have rapid batch and product changeovers, and are seeing excellent cost, quality and delivery results.

 It is generally understood that certain players perform better than others, but those who have tried to understand their performance vis-à-vis that of the industry have found little transparency, making it difficult to compare the results at different sites or discover the industry’s true level of competitiveness. Understandably, many manufacturers are asking themselves important questions:

• Which performance metrics should we consider?

• What does good performance look like?

• How big is our opportunity for improvement?

• Are there any trade-offs? For instance, does increased productivity hinder quality?

To uncover the true potential of a given biomanufacturing site, it is essential to ask the right questions, look at the right performance indicators, and make the right comparisons. Companies should begin by attempting to benchmark themselves against their industry peers, assessing the performance of each biomanufacturing site across the board, whether at the site, line or product level. Where available, a stringent benchmarking exercise will provide insights into important factors such as:

• Technical performance in relation to indicators such as yield, titer, success rates and improvement rates

• Operational performance characteristics such as utilization and cycle times

• Productivity factors such as costs, labor, capital and inventory

• Quality considerations such as the level of regulatory scrutiny, deviation rates and CAPA rates

• Structural factors such as capacity, technologies, automation levels, location and salary structure

• Complexity related to batch record entries, critical process parameters (CPPs), number of products and frequency of product transfers

• Organizational health indicators such as education levels, health and safety, turnover and labor allocation

 McKinsey’s global POBOS Biologics benchmarking has been used to assess these aspects across several biomanufacturing sites. This tool, which covers a big part of today’s global biomanufacturing network (including originators, emerging biosimilar players and CMOs) across various locations, provides a view into the reality of the biomanufacturing industry, perhaps for the first time.

One finding is the surprising variability in performance in the industry across all parameters (Exhibit 1). Even in the more standard fermentation of monoclonal antibodies, the cost per standardized batch  for some players is significantly greater than $1 million, whereas for others we have recorded significantly lower costs, even below $400,000 per standard batch. For the latter manufacturers, the COGS of the biopharmaceutical API (at less than $1 per dose) is so low as to be comparable to, or even negligible relative to the COGS required to fill and finish the drug product in a prefilled syringe (about $1.30 per unit).

Another important finding is that there is no real trade-off among the various performance dimensions. Players that do well in one category tend to do so across the board, from quality to cost and from lead time to success rate. In most cases, the gap between high and low performers depends on how well the operations are run, rather than on structural factors or complexity. In fact, there is no clear correlation between complexity — including such factors as the number of products, the number of product transfers and the number of regulatory agency registrations — and performance.

The impression from the field is that the competence and experience of each site drives most of the differences in performance. For example, several complex multiproduct sites — both top 20 pharma companies and CMOs — were doing more than tenfold better than a group of single-product sites, because the latter were relatively inflexible and conservative in their way of running operations.

However, there is also evidence that adding complexity does not help a site that is still relatively new and lacks the appropriate competencies. In one case, the transfer of an additional product to a site with below-standard competencies triggered a series of compliance problems, causing batch failures and significant delays in the manufacturing schedule.

Finally, it appears that high performers adopt new technologies to the greatest extent possible within the structural constraints of their manufacturing site, such as the addition of disposables in the upstream seeding processes. These high performers are not afraid to undertake the complications inherent to change controls or regulatory submissions when doing so will bring about performance improvements. Looking more closely, there may be even further interesting differences in the industry’s approach to day-to-day operations, including regulatory strategy, plant utilization practices and the approach to operational excellence.

Looking more closely, there may be even further interesting differences in the industry’s approach to day-to-day operations, including regulatory strategy, plant utilization practices, and the approach to operational excellence.   

Regulatory Strategy
In looking at the number of entries in a batch record, some players add complexity beyond the point of increasing control, whereas others have gaps in their regulatory strategy. In fact, we observe a variance of 3x among the various players. This difference in approach is confirmed by the fact that the complexity of the batch records strongly correlates with the number of CPPs in play, suggesting that players that adopt a stringent regulatory strategy in one area tend to do so across the board. (The observed variance for CPPs is even more marked, at 10x.)

Most interestingly, the approach to regulatory strategy also correlates closely with the site’s quality performance, albeit up to a threshold, indicating that specifications that are too simple may engender less-compliant operations. Above a certain threshold, however, tighter control no longer makes a positive contribution. 

Plant Utilization
The majority of the plants assessed to date appears to be vastly underutilized, with upstream time in operations normally ranging from 10 to 40 percent (on a 24-7 schedule). Both structural factors and managerial mindsets are behind this arguably limited performance.

Mono versus multi: Many sites have been built either as monoproduct sites or with lines dedicated to a single product. This creates a challenge for the manufacturer, because one product may not be enough to utilize a site’s full capacity, but two products may be too much. Given the high value of biopharmaceuticals, we find that COOs typically prefer to err on the side of excess capacity, allowing a site to be inefficient rather than risking a shortfall in the drug supply if market forecasts are inaccurate.

In contrast, in facilities that are engineered from the beginning as multiproduct facilities, with the capacity and flexibility to handle a number of products, the variability of product-demand forecasting begins to balance out statistically, posing less of a challenge to product delivery as utilization rates increase.

Capacity management: Looking at site utilization, most sites have uptime  of 20 to 40 percent of available time, and net production time of 10 to 25 percent. Further, 20 to 30 percent of available time is spent on nonproduction activities and other losses, often leaving idle time  of as much as 40 to 50 percent. We believe there is room to optimize nonproductive time. Net production time is small compared with what the pharmaceutical industry is used to achieving in the manufacture of small-molecule APIs, i.e., 50 to 60 percent, because the nonproduction activities inherent to the equipment batch cycle are extensive and, in addition, there is a significant share of time that goes into maintenance activities and avoidable losses. Further, we have observed a few players that have already managed to operate their assets more effectively, reducing the amount of nonproductive time by using a mix of operational-excellence initiatives and adopting technical solutions such as disposable equipment.

The uncertainty, variability and performance issues that have characterized biomanufacturing operations in the past have underpinned the choice to build in high idle-time buffers to protect supply. Such a choice is surely savvy in most circumstances, given that most biopharmaceuticals have market values that do not justify any risk of a supply shortage. Nonetheless, the same players that have managed to gain better control of their nonproduction time and are running more effective operations do generally operate with higher utilization rates and a smaller idle-time buffer, without incurring any significant issue. A focus on performance excellence allows these sites to address many of the losses, failure rates, changeover times, breakdowns and lengthy preventive maintenance that are the main drivers of uncertainty.

Approach to Operational Excellence
Instituting operational excellence improves performance across the board; in fact, improving performance along one dimension brings improvement along other dimensions. For example, excellence in operations delivers improvements in quality as well as improving cost performance. We have observed that quality correlates strongly with costs, with an R2 of greater than 0.6. The rule of thumb is that the “major deviation per standard batch” key performance indicator (KPI) correlates with the “cost per standard batch” KPI, because each 0.1 increase in the incidence of major deviations per standard batch is linked to a corresponding increase in the standard batch costs of about $500,000.

Benchmarking can provide insightful transparency into what “good” looks like in a given industry and which dimensions should receive the most attention. In small-molecule, solid-dose manufacturing, the understanding is that a substantial share of the costs is variable (40 to 60 percent) and greatly linked to workforce optimization and productivity increases. In biomanufacturing, in contrast, the overall cost structure of a site is relatively inflexible, with relatively low variable costs. Hence, performance is strongly dependent on output volume and utilization levels. Although utilization is the most important factor, optimization is still possible on other dimensions.

Every path to success is different. As an example, one Asia-Pacific manufacturing site has been able to keep its costs low, its FTEs to a minimum, and its success rate high owing to a strong focus on process automation. In contrast, an EU site with a similar product focus has relied on high-quality, experienced personnel for its success to date. Although the site’s personnel-cost share per standard batch is somewhat higher than average, it has nonetheless managed to keep its overall cost point in line with benchmarks and achieve effective operations, delivering good performance on most other dimensions (e.g., success rate, quality level and productivity).

We have found that performance levels seem to be linked to the education levels of the workforce. Of course, the biomanufacturing industry in general tends to have a strong share of highly educated staff. Yet education levels vary widely. Across all sites, about nine-tenths of the workforce has some level of technical or life-science background — underscoring the importance of a scientific education to form the basis for effective operations. More interesting, at better-performing sites, more than one-fifth of the workforce has a master’s degree or above, and at least three-fifths has a bachelor’s degree. In contrast, the worst-performing sites tend to have less educated staff, with closer to one-tenth of the workforce having master’s degrees. One notable exception is a site at which we unearthed high performance, yet a workforce of which more than four-fifths lacked any higher education. Digging deeper, we discovered that this site’s employees had among the highest tenures we have observed in the industry, with significant know-how developed on the ground over many years. As a result, we see a clear link between performance and education levels, especially if the average tenure at the site is low.

Capital Investment
It is often intuitively assumed that larger capital investments for a given amount of capacity will translate to better equipment and therefore higher manpower productivity and lower operating expenses. In biomanufacturing, however, that is not the case. Rather, we have observed limited to no correlation between the investment per installed fermentation capacity and either the manufacturing cost or the manpower productivity. In a few cases in which investments do seem to have delivered better infrastructure — for example, through increased automation — it has been difficult to verify performance improvement, usually because of underutilization. One exception is the previously mentioned site in Asia-Pacific, which has managed to realize value from its capital investment in automation by reaching top-quartile levels of utilization. In most other cases, the best-performing sites also have relatively low investment-per-installed-capacity profiles, while still emphasizing operational excellence. We therefore believe that in biopharma, how to invest is more important than how much to invest. This includes automation strategies that are deployed less for the sake of cutting costs and more to reduce human error, thereby drive quality outcomes. High-performing sites consume enough of a company’s capital expenditure to create well-engineered facilities but do not overspend — confirming that good engineering is not over-engineering.

Quality Assurance Staffing
We have found no standard or consistency in the industry that can help to determine the most appropriate QA-staffing level. In fact, there is no correlation between the number of deviations and the size of the QA organization, nor between the number of deviations and the number of CAPAs; nonetheless, we have made two interesting observations. First, we have found a moderate negative correlation between the size of the QA organization and the frequency of breakdowns and infections, suggesting that increased QA oversight could drive down the frequency of these issues. For better or worse, the higher downtime linked to increased infections and breakdowns does not really affect the cost point, most likely because this downtime is hidden in the idle-time buffer existing in most sites. Second, we have found some correlation between the number of QA personnel onsite and the level of CAPAs issued, hence indicating that CAPAs could be a proxy for QA workload and staffing requirements.

Scale & Labor
Among the many factors that potentially influence performance, we have found that the scale of operations has the greatest effect on costs, with an R2 of 0.7 correlating the costs per batch to the number of batches produced. Therefore, the more batches a site produces, the more competitive that site tends to be. After scale, labor productivity can have the biggest impact on unit costs. Labor costs in biomanufacturing are substantial, typically making up one-third to one-half of the total cost of a site. There is no primary department that generates the majority of these costs. The production workforce makes up anything between one-third and one-half of the total, while QA and quality control (QC) make up one-fourth to one-third and overhead and other production-support functions make up another one-fourth or so. As a result, labor productivity should be encouraged across the board.

Management should determine each site’s true performance potential relative to industry peers. Such a quantitative assessment may provide surprising revelations. For instance, the capacity a site can aspire to liberate can be substantial, whether through optimized changeovers (both product and campaign), improved management of unplanned downtime, better coordination of process steps or improved control of process variability. One company we observed was able to double its output from 50 to 100 batches in just one year by taking a leap of faith and challenging the current mode of operations: it increased the frequency of seeding and enhanced plant utilization, moving a sizable portion of its buffer time into manufacturing operation time.

Companies should begin by understanding the structural factors that define the maximum threshold of production performance in each of the relevant dimensions (output, lead time and quality). Structural limits are higher than they are assumed to be, and current assumptions should be challenged in a constructive way.

Once the true structural ceiling is determined, variables can be optimized one by one, allowing the company to set and then progressively realign targets over time on the basis of realistic performance-improvement expectations.

Finally, the belief that improving one aspect of performance will harm another is generally incorrect. On the contrary, poor quality generally leads to high costs, while the pursuit of excellence brings benefits across the board.

As the biopharmaceuticals industry matures and becoming progressively more mainstream, its managers are beginning to take a new look at their operations, opening themselves to questions about improving both their technical and their operating performance. Those ready to commit themselves to the task today have the opportunity to get ahead of the industry tide that we see coming over the next few years. As they do, they are likely to attain a new level of performance excellence, one that will give them a competitive edge and establish them as top performers in the biomanufacturing industry. 

About the Author

David Keeling | Ralf Otto