????兩年前，我投身風險投資行業，重點關注清潔技術類初創公司。這就像是通過觀看八年級學生（譯注——相當于中國的初二學生）打籃球來發掘未來的NBA球員：他們還需要很多年的時間來成長，一旦他們真的成長起來，情況就會大不一樣。選擇的關鍵之處在于出色的模式識別——早在成功到來之前就找到相應的征兆。

????清潔技術風險投資領域里的同行大多來自于其他領域，通常是半導體或通信行業，他們在模式識別方面的偏好也帶著這些行業的烙印。我不禁想問：“要是大家的偏好都錯了，該如何是好？”要是我們在選拔苗子的時后看走了眼，該怎么辦？

????我曾從事于技術分析工作，因此我希望找到一種基于事實的方法來找到這個問題的答案。于是，我選擇了一批成功的、獲得風險投資的初創公司，試圖找出他們的共同點。我所選擇的初創公司：

????? 已經取得了明顯的成功。我將“成功”定義為在大型股票交易所上市。我知道，這個標準遠遠稱不上完美，但至少是清晰明確的。

????? 在2000年之后上市。這是為了把那些在1998-99年互聯網泡沫時期搭便車上市的能源公司排除在外。

????? 得到了風險投資機構的支持。請注意，絕大多數的清潔技術上市公司實際上都沒有得到風險投資。

????到目前為止，我選擇了18家公司作為樣本：

????I entered the venture capital business two years ago, focusing on cleantech start-ups. It's kind of like picking a future NBA draft by watching eighth graders play basketball: There are many years worth of development to go, and the subjects will look a lot different once they're all grown up. The key is excellent pattern recognition – identifying the antecedents of success long before it happens.

????Most of my fellow cleantech VCs came to this field from some other domain – often semiconductors or telecom – and brought their pattern recognition biases along with them. The question I found myself asking was "What if everyone's biases are wrong?" What if we're picking the wrong eighth graders?

????Coming from a career in the tech analyst business, I wanted to find a fact-based way to answer the question, so I assembled a set of successful VC-backed start-ups and looked for what they had in common. I picked companies that:

????? Had achieved unambiguous success, which I defined as an IPO on a major exchange. That's far from a perfect criterion, I know, but at least it's clear-cut.

????? Went public after 2000, to rule out the boomlet of '98-'99 energy IPOs (when the Internet bubble's rising tide lifted all boats).

????? Were backed by institutional VC. Note that the majority of public cleantech companies have, in fact, not been venture-funded.

????As of today, this sample consists of 18 companies:

????上面這份名單能告訴我們什么？平均來看，這些清潔技術初創公司的好苗子：

????? 通過發行股票籌集了1.22億美元。其中不包括補貼和債務。融資規模相差很大，但如果把Q-Cells公司和REC公司作為特例排除在外，它們的融資規模都不低于3,000萬美元。如果某份新的商業計劃提出了比這個數字更低的生命周期資金需求，必須有無懈可擊的理由來加以證明。

????? 上市時估值達到9億美元。估值差別也很大。假如投資者在這些公司上市時持有80%的股份，這些股份將帶來5.9倍的投資回報率。

????? 從成立到上市需要8.3年。清潔技術公司從成立到上市大概需要10年時間。但請注意，那些在2010年后上市的公司所需時間較短，平均為6.7年。這類初創公司在成立時或成立后不久就獲得了風險投資；相比之下，前者常常在自力更生了數年之后方才得到風險投資。

????? 在1/50的成功率中幸存下來。這18家公司在1995年至2007年間首次獲得風險投資。風險投資行業追蹤機構風險資源（VentureSource）的資料顯示，清潔技術行業在此期間共進行了986次種子融資或首輪融資。毫無疑問，今后將有更多的初創公司獲得成功，尤其是最近幾年獲得了風險投資的初創公司。但到目前為止，得到風險投資支持的初創公司成功率僅為18/986，也就是1.8%。

????如果將某些財務標準考慮在內，我們可以了解到更多情況。平均來看，這些獲得成功的清潔技術公司：

????What can we learn from this? The average NBA draft pick of cleantech:

????? Raised $122 million in equity financing, which excludes grants and debt. The distribution is pretty broad, but I'd note that if you remove Q-Cells and REC as special cases, the floor is ~ $30MM. If a new business plan poses a lower lifetime cash requirement, there should be an airtight argument to justify that claim.

????? Went public at a ~ $900 million valuation, also with a broad distribution. If you assumed that investors owned, say, 80% of these companies at the IPO, the lot would have delivered a 5.9x return on capital invested.

????? Took 8.3 years from founding to IPO. If you're signing up to build a cleantech winner, reserve a decade. Note, however, that this elapsed time has shortened for companies that have gone public since 2010, where it averages 6.7 years. Start-ups in this later group received venture financing either right at their founding or soon after it, in contrast to earlier companies which often bootstrapped themselves for years before taking venture capital.

????? Survived a 1-in-50 success rate. These 18 companies received their first venture investment between 1995 and 2007. A quick VentureSource search indicates 986 seed/Series A rounds done in cleantech during that period. While there are doubtless further successes to come, particularly from start-ups funded in the later years, the yield so far is 18 / 986 or a 1.8% of start-ups funded.

????If we include some financial metrics, we learn a bit more. The average cleantech winner also: