Mary Meeker's slides on Internet trends

Check out these slides on Internet trends from Mary Meeker, partner at Kleiner Perkins and former analyst at Morgan Stanley:



Innovation Velocity at American Universities  

A few weeks ago, a friend of mine forwarded me an article that analyzed the correlation between the velocity of start-up creation at American universities and total research expenditures.  The author, Jerry Paytas of Fourth Economy, makes an argument that universities become more efficient at start-up creation as their research expenditures increase.  The article notes, 

“As the volume of research increases, institutions become more efficient. At $200 million to $400 million in R&D, institutions can expect only a modest increase in startup rates – getting one startup for every $92 million in research. The very best schools, those that produce more than 4 startups per year, are able to generate one startup for every $77 million in research. For the smaller institutions, implementing the best practices and doing everything you can to be efficient at producing startups might add one more startup every other year.”

It is true that smaller institutions do not create a whole lot of start-ups, but that does not mean they are not good at it.  In fact, smaller schools can be quite efficient at creating start-ups.  

If research productivity is really correlated with research expenditures, then the mega research centers (e.g. Caltech, Harvard, MIT, Stanford) that have research expenditures greater than $500 million should significantly outperform smaller institutions. 

Interestingly enough, when weighted by research expenditures, the University of Kentucky outperforms both Harvard and NYU with regards to the average amount of research dollars that are required to create a start-up.  Admittedly, Kentucky is a bit of an anomaly (later in this entry I will show aggregate data on mega research center performance), but it does highlight the fact that efficiencies do not always correlate with scale. 

What my data also shows is that it costs Kentucky more money to generate a license than it does Harvard.

Licenses are important drivers of innovation, yet often play second fiddle to sexy start-ups.  The reality is that both licenses and start-ups are proxies for the innovation efficiency of universities.  My reasoning is based upon the theory that if a license was not transformative or valuable (e.g. “innovative”), then no one would license it. 

Given the important role licenses play in promoting innovation, shouldn’t the measurement of university innovation productivity be a blending of licensing and start-up activities?

The chart above is compiled from a list of universities that have research expenditures that are greater than $200 million.  I chose $200 million as a cutoff point because schools with research budgets smaller than $200 million typically have research programs that are either incredibly specialized or lean heavily toward basic research.  Also, omitted from this list are state systems (e.g. State of Texas System), which aggregate research expenditures on a system-wide basis instead of at the university level.  In total, 59 universities made the list.

In line with the Fourth Economy data, the 59 universities in my data set required about $93 million of research expenditures to spawn a single start-up.  Interestingly, it took roughly 1/9 of that amount of investment to generate a license.  And, when licenses and start-ups were blended together, it took only $8.7 million of government investment to create a transformative piece of technology. 

Some might argue against this, but I think that $8.7 million is a reasonable return on investment for the government and is a testament to the efficiency of American universities.

Also in line with the Fourth Economy data, mega research centers do seem to be more efficienct drivers of innovation than smaller institutions.  That is not entirely surprising since schools like Caltech, MIT and Harvard typically attract premiere researchers.

What I did find surprising was the fact that universities with no medical centers seemed to outperform those with medical centers on a cost per innovation (license or start-up) basis.  Given the fact that non-medical funding sources are more scarce than medical, I did not expect non-medical schools to be more innovative than those with medical schools – and, to do it with less money!    

American universities are highly innovative and productive; however, the metrics commonly used to track innovation (e.g. start-ups per year) typically do not represent the full picture.  By leveling the playing field via weighing of research dollars, smaller schools can perform (i.e. be “innovative”) as well as large schools and non-medical universities can outshine those with medical schools.  


Partnerships - Value Creating or Value Destroying?

Whether or not to partner a therapeutic is a critical strategic question for any life science start-up.  Unfortunately, there is no simple answer or clear strategy for when or how to structure those partnerships.  In the July/August 2011 issue of In Vivo Magazine, Alex Lash wrote an article that pulled together some fantastic data on partnerships for therapeutic start-ups.

Alex created a list of 50 start-up companies that had FDA approvals from 2001 to the first half of 2011 (does not include Plexxikon) for their lead product (only counted first time approvals).  Not surprisingly, virtually all (Regeron’s Arcalyst is a notable exception) of those lead products were partnered in some manner (co-development, outright out-license, co-market, etc.).  From the list of 50 start-ups, Alex tried to see if any correlation could be drawn between the types of partnerships that were established and the ultimate success/value of those start-ups.
Alex notes, “Partnering strategy appears to be more important to a company’s long-term stock success.  The more US rights a company on our list has kept, the more chance it has to outperform its peers on the stock market.”

While the sample set contains 10 years of data, there are only 50 data points (start-ups with FDA approvals that met the criterion) from which conclusions can be drawn.  That being said, some loose conclusions can be drawn from Alex’s work:  

Holding onto US rights is important.  The US market is the most mature and lucrative therapeutic market in the world.  Therefore, holding onto US marketing rights is essential for start-ups to achieve maximum value.  That being said, it is not so easy for start-ups to hold onto all of a product’s value, especially in challenging development areas like cancer or cardiovascular that require extensive amounts of capital.  Giving up some value in order to get a product over the finish line is the reality for most start-ups.
Of the 31 active and independent start-ups with FDA approved products, 14 have never partnered their product in the US (e.g. ALXN, AMLN, CBST, INSM, REGN, UTHR).  Besides great management and products, those 14 companies also focused on niche markets and/or reformulated products with known safety/efficacy profiles - thus, mitigating some of the development risk and costs.  
Partnerships do not necessarily tie a start-up to one company.  Some people believe that partnering can limit exit opportunities, especially if a partner has a Right of First Refusal or Right of First Negotiation.  That may be true to some degree, but the chart above shows that 2/3 of acquired companies were acquired by an non-partner.  As drug development costs have skyrocketed, pharma companies have increasingly begun to look at partnering as a way to share risk - which in turn has created an excess of partnerships.  With so many partnerships in play, as well as the rate at which pharma companies seem to be switching corporate strategy, the negative perception risk of a strategic not acquiring a partner might not be as significant as previously thought.
Only 20% of the start-ups are profitable.  Of the 50 companies with FDA approved products, only 11 are profitable.  Of those 11, 9 partnered their lead product and only 1 sold all of their US rights.  This data would suggest that partnering ex-US rights (if structured and timed in an appropriate fashion) correlates strongly with the ultimate success of a start-up post FDA approval.  That being said, many of these companies/products were partnered because they tackled large chronic markets with high development costs, which in turn correlates with high company profits post approval.
Few start-ups successfully commercialize a 2nd product.  Only 10 companies on Alex’s list have brought a second product to market.  He notes that, “nearly all originated outside the company” and cited United Therapeutics (in-licensed tadalafil from Eli Lilly) and Millennium (Velcade, Leukosite acquisition) as examples of in-licensing successes.  This data highlights both how hard it is to develop therapeutics and why biotech analysts so heavily discount the value of start-up’s pipelines.  Also, because 2nd products are so heavily discounted by analysts, many start-ups and investors encourage early stage companies to partner those products early to support the development of the lead product.  This can be a source of discord between investors and portfolio companies, as most companies - and even some investors - like to hold onto 2nd products as downside protection in the even that the lead product has development issues.

Positioning a Start-up for a Partnership
Partnering is a complex strategic decision for which there is usually no right answer.  The best a start-up can hope for is to produce exceptional data that will entice a strategic’s interest under favorable terms.   

A lot of start-ups refer to partnering in investor decks and usually do so in terms of comps (e.g. partnering at “X” date at “Y” price just like company “Z” did) instead of data.  While comps might inform partnering talks, no two partnerships are alike.  And, unless one has inside information about why a strategic formed a partnership with a start-up, it is hard to draw hardened lessons that can be broadly applied.  

Start-ups that have great data will always generate partnering interest at favorable terms.  That is why many start-up CEOs spend considerable amounts of time with strategics getting to know exactly what data would most excite those strategics at the various stages of drug development.  Strategics are usually quite forward about what data they are most interested in, which in the long run can save start-ups significant time and money from going down the wrong path.    

The Future
Alex’s data indicates that the best way for start-ups to create shareholder value - at least at the moment - is to develop niche/specialty/reformulated products and hold onto US marketing rights.  Those takeaways seem to have caught the attention of investors and companies.  Three companies with upcoming approval decisions (Corcept Pharmaceuticals, Horizon Pharma, Onyx Pharmaceuticals) all retained full ownership of their niche/specialty products.  Horizon and Onyx are slight exceptions to the rule as both are submitting their second products for approval - but, should provide interesting data points points to view as analysts reset company valuations based upon the approvals of those products.

The Summer of Companion Diagnostics

This has been an exciting summer for the biotech community as two highly anticipated approvals for targeted therapeutics, along with their respective companion diagnostics, were given the green light by the FDA.  The approvals of Zelboraf (Roche/Plexxikon) and Xalkori (Pfizer) was a great highlight for the pharma industry this summer, and could result in the accelerated development of other personalized medicines.  

Zelboraf and Xalkori are both orally administered therapies that address a specific patient subpopulation that is identified via a companion diagnostic.  While these two drugs address modestly sized markets (see chart below), their ability to effectively treat a specific patient population enables them to garner premium pricing.  Some industry insiders are now referring to similar drugs that address niche markets and command pricing power as “nichebusters”.  

There are pros and cons to the nichebuster model as it relates to companion diagnostics, and I thought it might be helpful to lay out some of the facts regarding the recent approvals of Zelboraf and Xalkori.  Though quite a bit older (approved in 1998) than the other two therapies, Herceptin provides a good comp:  

These three oncology products have markedly different approaches to pricing - but why?

  • Therapeutic Cost: Herceptin helped set the price range for targeted cancer therapeutics at $50,000-$100,000.  Because doctors are able to identify likely Herceptin responders with a companion diagnostic and only treat those patients, the drug is able to command a premium price.  The more powerful the diagnostic is at predicting likely response, the better chance the drug developer will be able to command premium pricing.
  • Diagnostic Costs: While the Zelboraf companion diagnostic is a pretty straight forward PCR test, Herceptin is a little more complex - involving immunohistochemistry (IHC) and sometimes fluorescence in situ hybridization (FISH).  Despite the variance in technical difficulty and time associated with the lab test, the companion diagnostics for Herceptin and Zelboraf are both about $100.  In the face of those precedents, Abbott intends to charge $1,500 for its own FISH-based test to identify Xalkori responders.  At $100 and $150, doctors can be fairly price insensitive when prescribing a diagnostic test for Herceptin or Zelboraf.  But, doctors may not be so price insensitive when it comes to a $1,500 test.  Even if doctors do prescribe the Abbott test, it is unclear if CMS will reimburse the product at $1,500, let alone $1,000. One last thought to chew on: what happens to pricey companion diagnostics when the cost to sequence a person’s genome breaks the $1,000 price threshold or even the $500 threshold?  
  • Dx Hit Rate & Market:  At what utility level (what I refer to as “Hit Rate”) and price point does it become a no-brainer to use a companion diagnostic?  In the case of Herceptin and Zelboraf, both have companion diagnostics that are relatively cheap ($100-150) and those diagnostics identify mutations that are relevant in a large percentage (25-50%, depending on the test) of the patients who will be tested.  Xalkori, on the other hand, is quite different as it has a high price tag and is not relevant for 24 out of 25 patients that are tested.  To look at it another way, it takes $37,500 (25 tests x $1,500 cost) to identify a patient that is likely to respond to Xalkori - making the real cost of Xalkori more like $117,500 ($80,000 Tx + $37,500 Dx).  Despite the fact that Xalkori is quite efficacious for those patients harboring ALK mutations, prescribers may balk at the all-in price tag of $117,500.  What I cannot speak to is how the all in cost of Xalkori relates to the long term economics of treating NSCLC.  There is a chance that the overall economic benefit of identifying and treating ALK positive patients might outweigh the all in cost of $117,500.
  • Bundling:  Roche is one of the few pharmas that has in-house expertise in both therapeutics and diagnostics.  That dual capability enables Roche to subsidize the cost of its diagnostic and transfer the cost of that subsidy into the price of the therapeutic.  By bundling low cost diagnostics with high cost therapeutics, Roche is able to drive adoption of its high margin therapeutic by basically giving away its diagnostic.  It is unlikely that Xalkori will have a bundled Tx/Dx because both components are made by different companies.  If Pfizer were to subsidize the cost of the diagnostic by paying Abbott a portion of the diagnostic’s cost, then it would have to charge quite a bit more (new cost = $80,000 + 25*subsidy) for an already expensive drug.  Charging more for the therapeutic could have a negative impact on adoption. 
  • Reimbursement:  The price range for simple analyte-based test like PCR has been firmly set by CMS at $50-400. With regards to complex tests (also referred to as “esoteric tests”), the agency has not formally set pricing guidelines.  That being said, CMS has indicated that it is likely to become more price sensitive for diagnostics that are priced above $1,000.  In the case of Xalkori, it is still unclear whether its FISH-based test will be considered an esoteric test and warrant a high price tag.  
  • Regulatory:  As therapeutics and diagnostics are regulated by two different groups at the FDA, coordination between those two groups is a must for any company trying to develop a companion diagnostic.  For many early stage therapeutics companies, it is cost prohibitive to develop in house expertise in both therapeutics and diagnostics.  Therefore, most start-ups shift diagnostic development to third party groups.  Providing the required oversight to ensure that third party development of the diagnostic is of high quality can be a challenge for a start-up that has a core competency in therapeutics.  


Diagnostics offer pharma companies the ability to create new therapeutic sales channels and strengthen already existing ones.  Regulatory changes have decreased doctor access for sales reps, which then limits the ability of reps to drive the adoption of high margin therapeutics.  Diagnostics, and even devices, provide another avenue for sales reps to garner more face time with doctors.  By creating in house diagnostic divisions (ex: Novartis purchasing Genoptix) or partnering with existing diagnostic players (Roche, Abbott, LabCorp), pharma companies can leverage those diagnostic sales channels to better position and sell their therapeutics.  As pharma companies increasingly ramp up their diagnostic efforts, expect nichebusters to become a greater portion of pharma companies’ pipelines.


When States Become VCs

Companies in high growth sectors such as Cleantech, IT, and biotechnology are coveted by states around the country.  To facilitate the establishment of companies in those sectors, many states offer a variety of incentives, including grants and tax abatements.  More recently, states have focused on capital investment in venture capital funds as a way of attracting high-growth companies.  As VCs have historically invested in high-growth sectors, many states have chosen to either align themselves with existing venture funds or start their own state-affiliated funds.  

In 1976, Connecticut started the first state-promoted venture endeavor, the Connecticut Product Development Corporation (CPDC).  Through direct equity investments by the state, the CPDC enabled entrepreneurs with compelling ideas to obtain early stage capital.  This model proved to be quite successful and was eventually copied by a number of other states.  Today, there are approximately 150 state-affiliated venture (including Fund of Funds) funds.

Should States Act as VCs?

As the competition for jobs has ramped up, many states have felt the urge to provide a guiding hand to drive companies to their state.  State assistance is often seen as necessary since few states organically have the tools (abundant capital, cheap real estate, and abundant human capital) necessary to draw companies to their state.

While most states recognize the need to promote start-up creation, few are in a position to effectively deploy the capital and resources necessary to manage a cohesive start-up promotion program. 

Running a cohesive start-up programs requires:

  • Consensus building between the legislature, venture community, and the corporate community
  • Defined strategy - direct investing, tax incentives, grants, etc.
  • Requires the state bureaucracy to make informed business decisions
  • Requires the state to take on venture capital-like risk with tax payers’ money
  • Requires talented program managers at the state level to run an investment program within a bureaucracy 

These are real challenges that need to be fleshed out before a state decides to allocate significant capital into start-up creation.  

Investment Structures Used by States

Unlike state pension funds which invest from a defined pool of capital, state capital investment programs often borrow money to make their investments.  States are able to obtain low-interest loans from banks by using future tax receipts as collateral.  The loan is then used by  the state to invest in venture funds.  If the state’s investments in venture capital funds return all of the invested capital plus interest, then the state’s investment will be cost neutral.  However, should the venture funds not provide enough return to cover the loan plus interest (excluding other costs to keep things simple), then the state is required to make up the difference by tapping into its tax receipts.    

Below is an example of how state capital investments are structured:

The opportunity cost of capital for the state is quite substantial in this model due to the various management and service fees that have to be paid yearly to the venture funds, fund of funds, banks, lawyers, and auditors.  In aggregate, these fees have an 8-10% cost structure built into the state’s capital investment.  Therefore, the structure I outlined is only cost neutral if the investments provide an IRR greater than 10% - meaning that the state’s pooled venture capital investments must return at least double the capital invested just to break even.  

That being said, setting up investment structures like the one I charted out can be quite effective, even if the venture funds do not perform all that well.  For instance, by requiring venture funds to make investments in local companies, those funds will contribute to local job creation and a larger tax base.  Even if tax payer money is needed to pay off the LOC to the bank, that money will have contributed to job formation in the state.


I don’t think any one state has perfected an incentive model for start-up creation, but I do think Pennsylvania does a pretty good job of supporting start-up creation.  

Pennsylvania is fortunate to have numerous top tier research universities within its borders, and can leverage the talent at those universities to spur technology transfer and start-up creation.  Additionally, numerous corporate headquarters are located in the state and provide a large and stable talent base for start-up companies.

Even with great research centers and corporate support, Pennsylvania is in a constant race to attract companies that can generate jobs for the Commonwealth.  Over the last 30 years, Pennsylvania has launched three major initiatives (I refer to them as the “Three Waves”) to promote job creation.

First Wave - Ben Franklin Technology Partners

Started in 1982, Ben Franklin Technology Partners (BFTP) operates four regional non-profit organizations to serve and promote companies across the state.  BFTP’s activities are subsidized by a yearly budget of $20 million to make investments in start-ups that range between $30,000 and $150,000.  The effect of BFTP on start-up creation in the state has been profound and BFTP has become the go-to source of start-up capital for most start-ups.  Since inception, BFTP has invested over $173 million that has been subsequently matched by over $408 million of private investment. 

In 2008, the U.S. Economic Development Administration, an arm of the U.S. Department of Commerce, named Ben Franklin “the most effective regional economic development organization in the nation.”

Second Wave – The Tobacco Settlement

As part of the Master Settlement Agreement of 1998, the major tobacco companies were required to contribute funds to ease states’ burdens in caring for residents with tobacco related illnesses.  With some of the proceeds generated from the tobacco settlement, Pennsylvania approved two one-time appropriations for strategic areas of investment - the Health Venture Investment Account and the establishment of three regional biotechnology research centers known as the Life Science Greenhouse program (LSG).

The Health Venture Investment Account invested $60 million in four life science venture capital funds that pledged to invest that capital in Pennsylvania companies.  Those funds also pledged to match every public $1 with another $3 from private investors, thus turning the state’s $60 million commitment into a $240 million war chest. 

The Life Science Greenhouse program was created to address a lack of valley of death (gap between academic work and early product refinement needed to attract VCs) capital for start-ups.  The three Greenhouses provide not only early stage capital (up to $1.25 million in companies over the life of an investment), but also mentoring and consulting services.

Third Wave – Venture Capital Investment Programs

In 2004, then-Governor Ed Rendell, put forth an ambitious $2 billion stimulus plan to help boost job creation in the state.  Knowing that Pennsylvania lagged its peers in start-up creation, Gov. Rendell carved out $310 million from the stimulus plan to directly fund two new venture programs: PA Venture Guarantee Program and the PA Venture Capital Investment Program.

Launched in 2007, the PA Venture Guarantee Program sought to leverage $250 million of capital to attract VC investors to the state’s start-ups and businesses.  The Guarantee Program was quite novel in that it incentivized VCs to take on more risk in high growth businesses by protecting against the potential loss of investors’ principal.  In return for investors committing at least $15 million of capital to a Pennsylvania start-up, the PA Venture Guarantee Program would cover the first loss of the aggregate amount of principal invested in the company up to 50% of the total principal invested (i.e. $7.5M covered on a $15M investment).  This program essentially encouraged VCs to double down on investments made in Pennsylvania companies. 

To participate in the program, the state created a prerequisite that the recipient (VC fund) would maintain an office in Pennsylvania staffed with at least one senior-level partner for the duration of the guarantee

Philadelphia and Pittsburgh attract the lion’s share of start-up capital, leaving vast regions of the state without funding sources.  Structured as a $60 million Fund of Funds, a second venture program - the PA Venture Capital Investment Program – was launched to promote investment outside of metropolitan areas.  Like the Health Venture Investment Account, the Venture Capital Investment Program has a 3:1 matching component, turning the state’s $60 million investment into $240 million of capital. 

The success of the PA Venture Capital Investment Program has been quite remarkable over the last five years and the program currently has 12 participating venture funds, 5 of which were 1st time funds.   Those 12 funds have collectively invested over $117 million in 45 Pennsylvania companies.   

Concluding Thoughts

Like entrepreneurs who are looking for their first seed funding, VCs often struggle to find anchor investors.  Given the current economic climate, states might want to consider playing a greater role in spurring venture fund creation and helping maintain the capitalization of existing funds. 

Five first time funds were funded over the last few years through the PA Venture Capital Investment Program.  Without the state stepping up to the plate as an anchor investor, it is hard to believe that those funds would have been raised during the economic downturn.  Yet, it is in such down economies that VC dollars are so impactful – providing needed capital to start-ups as exit windows remain closed and creating jobs through the funding of new companies.  

With the proper incentives and prerequisites in place, capital investment in venture capital funds by states is a proven driver of job formation via the growth of start-up companies.