By Bruce Booth, contributor

The Startup Genome project recently released a report on the DNA of internet start-ups, essentially trying to identify attributes that lead to success or failure. One of the things they found was that 74% of start-ups failed because of “premature scaling.” Sounds like an unfortunate medical condition, but it’s Internet companies spend too much money before they really know what they have – their product, customer, market, etc…

While reading it, and my tech partner Fred Destin’s post on it, I couldn’t help but think about the issue of premature scaling in life science start-ups. Spending too much, growing too fast – not an uncommon characteristic in biotech. It almost always leads to shareholder pain and a loss of invested capital.

Here are four types of premature scaling (or inappropriate scaling) I can think of in biotech, and we try to avoid them all:

1. Building a Big Science story too fast. This is the “Go big or go bust” strategy with a group of Nobel laureates: Raise enormous amounts of capital to fund a novel discovery or research platform without enough evidence of target validation in a disease setting, confidence in chemical (or biological) tractability, progress on a lead program, etc. This generates big teams, big footprints, big stories – and massive burns. If the substance and, in particular, the rapid progress on product development, doesn’t get in line quickly, a big gap in valuation emerges that can crush these investments.

The right way to build a Big Science story today involves scaling consistent with a science-led, capital efficient approach: Build a sound platform with 15-20 FTEs on modest equity raises, find partners to help offset the growth and validation of that platform and then grow into the Big Science story as R&D evolves. The wrong way to build these is through rapid scaling around a hype-led fundraising machine. More often than not, investors get burnt with these.

Synta is a good rapid-scaling example. They have raised and spent $350 million, had at one time a team of 150+ FTEs or more and built a big broad portfolio — but their investors have suffered considerably. Story is far from over, but at the 10-year point its looking tough for the early investors.

Sometimes this model works, at least for investors. If the company can achieve escape velocity with enough hype and buzz in the market, they can get public or acquired early. Sirtris is a good example of a high escape velocity ‘big science’ deal that made it pubic and was acquired; its fair to say that many spectators wonder if GSK is regretting its $720 million acquisition, but at the time the story had a ton of public relations momentum.

2. Building a big company when it’s really a project. Lots of venture money is wasted building “companies” when they are really just product development vehicles. I covered this theme under a prior blog around new liquidity theses. By stapling multiple programs together, building a big team especially on G&A and running multiple studies at once, investors often think they’ve diversified their risk. Most of the time they’ve just raised the capital intensity of their deal such that one product bump and the whole thing gets revalued enormously.

Big Pharma buys these plays for single programs typically and so if a company is lucky enough to have two winners, say a Phase 2 and pre-clinical program, they leave real value on the table. If you’ve got an interesting asset, then develop it. But there’s little reason to put the expensive trappings of a bigger company around it. Leverage a part-time group where possible; you probably don’t need a CFO or, God forbid, an HR person. Focus on lean product development. Stromedix and Zafgen are great examples in our portfolio.

3. Building too fast on back of a partnership. Biotechs often get seduced in premature scaling by the siren song of partnering: They do a big deal on their platform, and then expand their organization and footprint, and try to work on more projects – all increasing their net burn. In short, it’s often an illusion that the partnership actually brought non-dilutive runway extension to the company. Sadly, when the sugar daddy partner terminates the deal, the biotech is left way out of balance and has to RIF its staff. In the public markets, this has recently happened to Alnylam with Novartis and Targecept with AZ.

It’s much more painful for private companies with weaker cash positions. This strategy – of aggressively funding internal burn rather than buying runway – can work if the company is lucky enough to develop some interesting assets with the free cash from their partner. But there’s alot of luck involved (true of all of biotech, I guess). The alternative is to truly scale your organization to the partnership. Our portfolio company Vitae has managed this reasonably well. It last raised equity in 2004, and has used its pair of deals with Boeringher to extend its runway while selectively advancing internal projects. Plexxikon was similarly successful; hadn’t raised equity for years before it was bought for more than $800 million by Daiichi.

4. Building out before a big outcome. Drug approval, for instance. In Big Pharma, having product to sell into the channel the day after approval is the goal for most blockbusters; in the event of a CRL rejection, they can eat the costs. But in cash-strapped biotech, this tends not to work. Small companies that build out aggressively before an approval more often than not get crushed. Vicuron a few years back. Adolor. ARCA. All hired sales reps prior to a pending approval, failed to get approval, and had to RIF large numbers of their employees.

This is not only tragic for those sales folks, but it also exacts a huge tax on the capital intensity of these businesses (especially the small ones). Hubris and excessive optimism are the typical causes of this one, but sadly Boards still let this happen. Horizon Pharma (another one of ours) recently did it right – kept the company under 20 FTEs through approval of its new drug Duexa in April 2011. It’s now public and working on the sales & marketing organization.

The counterpoint to premature scaling is what we at Atlas coined P/B/S. No, not Phosphate Buffered Saline. It’s “Prove/Build/Scale.” Thinking through small bets to prove a hypothesis, slightly more to test the programs, when to spend to grow a company, etc. Most of our seed investing is done to prove concepts. Building requires partners (which is where we often exit). And scaling requires functioning capital markets so rarely happens today.

Premature scaling can kill companies and investments. Worth keeping that in mind for the next budget cycle.

Bruce Booth is a partner with Atlas Venture. He blogs at www.lifescivc.com.