Based on my experience in the biotechnology industry, beginning in the mid-1970s as an investor, company founder and now turnaround chief executive officer, it seems clear that there are two distinct businesses that populate the biotechnology sector—product development companies and service providers for such companies. The product development companies often are based on very esoteric technologies that seem to have the potential to create blockbuster products. Their strategy (or more precisely the venture capitalists that initially finance them) seems to be to raise lots of OPM (other people’s money) and bail out before anyone really understands what the company is all about. Traditionally, these companies raise their big money with IPOs when the markets are in their “silly” periods and, despite dire predictions of impending doom because of evaporating cash, unprofessional management, missed projections and dubious scientific theories, they prove devilishly hard to kill. Just when the end seems inevitable, the market enters another silly period and the un-fundable becomes fundable again.

On the other hand, service companies such as CMOs appear to be rather straightforward (although they often are not), so that analysts believe they can understand the business. This means that management has to be concerned with decidedly non-esoteric issues such as revenues, EBITDA, cash flow and long-term growth rates. Of course, there are hybrids—service companies with unfathomable technology—but these will have to lie beyond the scope of this article. Thus, at least in theory, for service companies to prosper they have to identify a business model that can drive exponential earnings growth.

For the manufacturing outsourcing industry, history is not particularly encouraging. The early CMOs failed for several reasons, and to understand the impediments to achieving a successful business model we need to revisit these failures.


Vulnerability to In-House Capacity
Like the software industry, CMOs found that outsourcing dried up in slow times as customers retrenched and took diminished demand in house. It is not at all clear that this situation has changed today. The problem stems from how you view the experts’ contradictory projections. Some predict a serious shortage of biologics manufacturing capacity by 2006 and others project overcapacity. Which is true? It all depends on whether you believe there are 250 or 350 to 400 monoclonal antibodies in development and what you think the likelihood is that none, a few or many of the proteins identified by the human genome project and its offshoots will become products. If you believe that such proteins will have the same success that other proteins traditionally have enjoyed because of lack of safety problems, then the shortage theory may be correct. If on the other hand you think that safety questions associated with such proteins are a complete unknown—i.e., you worry that the situation with cytokines such as interleukin-12, which had a disastrous clinical history, will be more the rule than the exception—then the overcapacity scenario may be more likely.


Failure To Anticipate Product Failure
It was assumed that biotech products would be safe and effective because they were “natural.” Why people didn’t think about how really bad natural things can be (e.g., hemlock) was and still is a mystery. Nevertheless, the high failure rate of the early monoclonal antibodies because of safety problems and lack of efficacy led to a much lower demand for manufacturing than most people had believed. This translated into much less outsourcing than was projected, wholly apart from in-house capacity. However, as noted above, the success of the current crop of biologics may not be a given. While we appear to have solved at least some of the problems with MAbs, there is not yet enough experience to consider the case closed. And, again, based on history, I would be surprised if we didn’t encounter the same laundry list of problems with genomic proteins as we did with the early MAbs. The rule of thumb is that it takes four to five years and $150 to $200 million to construct and validate a cGMP production facility (based on a rough estimate of $2 million per thousand liters of capacity). Given the uncertainty of product approval and the high cost and long lead times for facility construction and validation and the “Field of Dreams” risk associated with “build it and they will come” is not only unquantifiable but extreme.


Building Production Before Approval
This requirement killed several companies that spent millions of dollars to construct and validate production facilities to produce Phase III/commercial materials only to see the product fail in clinical trials. You could argue that overly aggressive biotech companies can largely be blamed for the problem, but lead times are so long you could also argue that facility construction was necessary. In part this gave rise to the demand for CMOs, a rationale that continues today. Over the past ten years, the FDA has somewhat backed away from requiring that Phase III materials be produced under final manufacturing conditions and has shown flexibility in other areas, in certain cases allowing production at smaller scale to document ability to produce at larger scale. Recently, however, European regulators seem to have backtracked on flexibility despite the consequences to biotech companies (“if they fail, they fail”). They even claim that they are convincing the FDA to follow suit. If they succeed, repeating history may turn out to be the salvation for CMOs.


An Unready Market
What happened was that big pharma still considered in-house manufacturing to be the Holy Grail because of an institutional fear that their reputation would go down the drain if an outsourced product bearing their label were recalled. This now seems a bit strange in light of the number of big pharmas that recently have entered into consent decrees with the FDA because their older plants do not and probably cannot meet anyone’s definition of cGMP. Nevertheless, it is still not clear that big pharma has abandoned the Holy Grail. We in the outsourcing industry continue to see resistance to outsourcing production, although it now appears to be based more on internal power struggles than on operational considerations.


Failure To Satisfy Customers
Poor quality, poor interaction with customers, scientist CEOs who had no clue about what it takes to make a product and put totally unreasonable demands on CMOs certainly contributed to the last crash. Unfortunately, this situation has not changed all that much. While the CMOs are generally pretty well managed, the customers often are not and the less than ideal interaction between customer and manufacturer remains an industry-wide problem. With all apologies to the editor, “communication is key” and customers are often not very good at this. Combine this with the customer’s expectation of getting something for nothing and the tendency to pay slowly, and you have the makings of a bad business.

So what have we learned? It is not that scientist CEOs are bad people, but they tend to have little understanding of how to run a business. As mentioned above, the outsourcing business is by its nature a real business concerned with real issues, such as cash flow and the like, while product development biotech companies are more like beauty contests—their value is in the eye of the beholder. Thus, the two don’t really fit together very well and conflicts are inevitable. I also mentioned the cycles in the “silly” financing periods; in a down cycle, money can be impossible to obtain on any terms. If you are in the middle of manufacturing clinical materials or a product for a customer that has run out of money, it is far from trivial.

It’s not just the financing markets that can cause a sudden absence of money. Things like unresolved patent conflicts, unanticipated competition and dropping of support for a product by a partner can have an even more dramatic effect, as was the case of Pharming1. Big pharma customers are not the answer either, since they also can “run out of money” in a different sense – through a change in priorities within the company. While they will insure that you are not “out of pocket,” it is rare to be compensated for lost opportunity cost. Other problems that impact CMOs negatively are the overly optimistic assessments of the status of a product that are endemic to biotech companies and poor preliminary work on a product that arrives on the CMO’s doorstep. This often takes the form of a customer’s assurance that it has a “fully developed purification process” that turns out to be un-scalable. I don’t know how CMOs get compensated for doing the necessary purification development work because they run into egos the size of planets back at the customer’s shop. Suffice it to say that interaction with customers has to improve but the jury is still out on whether it will.

Given this environment, how is it possible to pursue a successful business model? We have implemented several ideas that seem to be working.


Get Paid as Much as Possible in Advance
Customers hate this and will resist it no matter how much reason and logic support the CMO. As mentioned above, there are many ways for a customer – large or small – to run out of money in mid-stream, with the CMO will inevitably end up on the short end of the stick. The CMO can try to protect itself with “take or pay” arrangements or firm commitments for space, but if there is no money or priorities have changed, it is very difficult to collect. The CMO can sue of course, but it’s not such a great strategy to sue a customer, since it tends to make other customers nervous. Wholly apart from the money question, it also is difficult to get a customer to pay for manufacturing a product that fails in the clinic, whatever the agreement says.

Problems associated with getting money up front stem from the fact that the CMO industry in its current form is relatively embryonic. Thus, it is possible for the “Mom and Pop” shops to claim that they are able to perform. The problem is, by the time it is obvious to the customer that they cannot, the money has been spent and the customer expects the CMO to reduce its prices so it can still meet its budget (and, thus, avoid admitting a mistake). Therefore, the CMO needs to do everything it can to meet this “garage” competition even though it is theoretically not competition.

Another competitive force is the product-oriented biotech or pharma company that has temporarily unused manufacturing capacity and offers to make products for others. Because it is not in use, such companies generally offer very attractive prices that just cover overhead expenses in order to keep their people in place. Many customers have had poor experiences with this arrangement because, whatever the contract says, the biotech company will find a way to throw the customer out when it needs to manufacture for itself. While the customer may have legal remedies, in reality, these are pretty useless. I often say to companies that a biotech company with a lawsuit will end up with a single asset —the lawsuit itself.

Keep in mind that the toll manufacturer can easily calculate the relative risk/reward of honoring the contract with the customer or making its potential blockbuster drug in the space and will make the obvious choice. From the customer’s perspective, is the risk/reward relationship even remotely in balance when you think about saving some money on manufacturing versus being unable to have your clinical materials when the trials are about to start? Nevertheless, because of lack of experience, too many customers go for the bait.


Understand the “Guts” Of the Business
It is dangerous to base a business model on the assumption that manufacturing biologics is essentially the same as manufacturing small molecules. Look at the almost complete lack of similarities between small molecules and biologics: lots of experience vs. very little; cost per gram to make—about $5.00 vs. $100 to more than $1,000; orders of magnitude greater quantities of biologics required than small molecules; generally much higher yields achieved for small molecules; and an actual capacity of biologics plants much lower than the theoretical capacity. All of these have to be taken into account when contracting or very unpleasant surprises can occur.


Be the Low-Cost Producer
You won’t get universal or even majority agreement with this concept but there is no more fundamental principal in business. If you are the low-cost producer, you can sleep at night knowing that no one can undercut your prices for long. Even experienced CEOs of CMOs will talk about quality and timeliness as factors overriding low cost. But on-time and on-quality are givens in any service business. With the uncertainty about demand and cycles that impact even the pharmaceutical business, especially the biotech segment, you will inevitably have to match or beat someone else’s price more often than not. We instituted strict cost controls and thoroughly indoctrinate our people with the concept of cost consciousness. We even let people go who didn’t or wouldn’t get it. As a result, we reduced our direct costs by 50% in a year through heightened employee commitment coupled with better materials and supplier management. This placed us in a much better position to withstand downturns. Cost controls also put you in a better position to price closer to the supposed non-competitors discussed above. Customers know there are differences among CMOs, and you can break through the barrier if you can come reasonably close to the other guy’s price. Furthermore, if you can keep your prices low, you are more likely to be able to get the customer to agree to milestone payments for successful performance. Since this is no-cost income, it puts the CMO on the road toward achieving exponential growth.


Proprietary Technology
I just don’t see how a CMO can achieve exponential growth in earnings if it cannot justify royalty payments because its platform technology is proprietary. Furthermore, an intellectual property estate has to be sufficiently broad to be able to give the customer reasonable assurances of freedom to operate. I don’t believe that for the reasons discussed above, a model based on markup on costs will work in the long run. Needless to say, this is another area in which there is less than universal agreement, al-though my camp seems to be getting larger all the time.

Royalty rates need not and probably should not be large. Why? Because a high royalty inevitably becomes a target for the customer—and, worse, its lawyers—to shoot at, especially in the age of concern about royalty stacking. Furthermore, any business model should be based as much as possible on statistical certainty—that is, making a large number of products for many customers. That way it becomes almost statistically certain that at least some will be approved and generate royalties. Even a small royalty on $500 million or $1 billion in sales is serious money.

Couple this with keeping costs low enough to be able to negotiate milestones and the CMO has a real chance at exponential growth. The difference between milestones and royalties is that, even if the product fails or is abandoned, milestones will be paid and can still result in exponential earnings growth. For this reason, a royalty-based model by itself is probably not viable because there are way too many reasons unrelated to failure in the clinic why a potential product never makes it to market. My best guess is that only about 20% fail because of poor clinical results. Other factors include: changing priorities, lack of money, unresolvable patent conflicts or royalty stacking issues.


Choice of Customers
Naturally if you could choose the right customers—those who would not change their priorities, run out of money, have products that will be approved, communicate effectively and honestly and will not challenge your royalty, life would be simple. But none of these can be quantified. On the other hand, you can listen to your “gut” when negotiating with customers, and if you have more business than you can perform, weed out those who do not appear to know what they are doing.

The question of partnering with customers by absorbing part of the product development cost is a difficult one. If you choose your customers wisely, it can lead to very big rewards even if you lack proprietary technology. But wholly aside from the question of what basis there is for deciding that a product managed by a particular customer is likely to succeed is the question of conflict of interest with other customers. I do not believe there is an easy answer to this question, especially in 20/20 hindsight.

I believe there are ways to achieve exponential growth in the biologics outsourcing industry. However, I also believe we are relatively early on the learning curve and are condemned to make many more mistakes before a workable model is found. In the meantime, I think the proper approach is to focus on reducing risk by controlling costs, keeping prices as low as possible, developing an IP estate for your platform technology, offering customers reasonable assurances of freedom to operate and incorporating milestones into contracts. With that under control, you can explore ways to achieve exponential growth, which by their nature involves unquantifiable risks, without “betting the farm.”


Notes
1. Pharming’s development project for Pompe’s disease was supported by Genzyme     until Genzyme suddenly acquired a competitor (Nova-zyme) and almost     immediately withdrew support for the Pharming project. Pharming’s market value     evaporated and within weeks it filed for bankruptcy. In-vestors are likely to end up     with nothing.