Structuring Alliances for Innovation

Why does this matter?

Companies invest substantial resources into partnerships. While some of those partnerships pay off, the resources will be spent in vain for others. Here we look at two alliance structures--Collaborative and Sequential--and discuss when each one is appropriate. We also share a predictive model that provides a probabilistic estimate for how a particular partnership should be structured, based on partnership choices made by over 2000 biopharmaceutical firms. Spoiler: Of the 2000+ partnerships analyzed, those that choose the appropriate structure outperformed the others.

Alliance Types and Industry Examples

For ease of reference, we refer to two parties entering an alliance as the innovator and the partner, with the innovator owning intellectual property rights prior to forming the alliance.

Collaborative Alliance

A collaborative alliance is one in which the innovator and partner continue to exert joint, equally important efforts into achieving the remaining goals of the venture.

Sequential Alliance

A sequential alliance is one in which the partner largely takes over the reigns in terms of further efforts into achieving the remaining goals of the venture and compensates the innovator based on future outcomes.

Sanofi-BMS: A Collaborative Alliance

Plavix, a drug for heart disease was Initially discovered by Sanofi Aventis and then co-developed by Sanofi and Bristol Myers Squibb. Plavix had glabal annual sales exceeding $9 Billion.

Shionogi-AZ

Crestor, a drug for high cholesterol was initially discovered by Shionogi of Japan, and then licensed to Astra Zeneca. Crestor had global anual sales exceeding $7 Billion.


What does the data say?

Clearly, both types of partnerships have shown great promise as a path to phenomenal success. How then, does one choose among the two alliance structures? We find that, among other factors, addressing three of the information problems discussed in our previous post plays a key role in alliance structure choice. These are:


  1. Asymmetric Information

  2. The hold-up problem

  3. Risk-aversion


Tackling Asymmetric Information

One way to address asymmetric information is signaling. The party with private information can undertake an action that credibly conveys the information it has to eliminate the information asymmetry. Typically, signals are costly but the cost depends on the type of information held. A party that incurs a lower cost to send the signal can do so and segment the market.

Within the context of biopharmaceutical R&D, private information may be due to differing fields of expertise whereby the partner has a relative disadvantage evaluating the quality of the innovator’s research niche. Collaborative alliances demand more resources from the innovator than sequential alliances. The expected cost of these additional resources are higher for low success probability innovators who are less likely to recoup their further investment. A biotech innovator can signal the quality of its product candidate by continuing its involvement in the project through a collaborative alliance. Indeed, we find that, when biotechs partner with pharmaceuticals, the probability of forming a collaborative alliance increases.


Tackling Hold-Up

Investments at any stage of development are relationship specific; such as the recruitment of patients for clinical trials, the efforts of scientists, the training of employees for the marketing of a specific product. The relationship-specific nature of investments makes the partnership open to the hold-up problem. To address the problem, one needs to have parties with comparative advantage make investments and allocate residual control rights to them.

In the earlier stages of development, many steps lie ahead with both parties having a comparative advantage in different tasks. In later stages of development, comparative advantage shifts to the partner. Control rights are typically more equal in a collaborative alliance but shift to the partner in a sequential alliance. Indeed, we find that, the later in the development cycle an alliance is formed, the less likely forming a collaborative alliance becomes.


Tackling Risk-Preferences

Parties that are less able to diversify away their risk tend to have risk-averse characteristics. In our sample of over 2000 partnerships, the average partner firm is 7 times larger than the innovator, making it more likely that the innovator (who has fewer projects) is likely to have more of a concern about risk. The degree of concern, in turn, depends on the level of risk and/or the level of risk aversion.

Managers at biopharmaceutical firms have access to historical failure rates for drugs targeting different disease indications. For example, drugs targeting respiratory diseases are half as likely to succeed as those targeting gastrointestinal diseases.

The innovator’s exposure to further risk can be limited through a sequential alliance, where most of the downstream outlay of efforts and cash are borne by the partner. In contrast, a collaborative alliance requires a higher resource commitment from the innovator, thereby increasing their exposure to risk. Indeed, we find that alliances with a higher risk of failure are more likelt to adopt a sequential alliance structure and less likely to adopt a collaborative alliance structure.

We also find that this relationship is less pronounced for larger innovators who can better diversify their investments and reduce their risk-exposure.

Eventual Outcomes

The good news is that most alliances formed, about 75%, are in accordance with the above findings and address the different information problems that they face through their choice of alliance structure.

What is the consequence of not addressing information concerns?

For an answer, we can explore the fate of the alliances that did not adopt the expected alliance structure: Not aligning structure with project characteristics can lead to an increased probability of a termination. Though only 24% of alliances in our sample are terminated pre-maturely, the probability rises to 62% for those that our model predicts to be terminated.

The above findings are robust across sectors but the specific circumstances of a given firm or sector will warrant specific analysis. For insights curated to your unique circumstances, to make use of predictive models, for collaboration opportunities, and speaking arrangements, please reach out through the contact form.