How Better Place Could Have Done a Better Job... Using Blanks

Published: by

Unfortunately, Better Place is everyone's whipping boy today. After all, the company blew through upwards of $800MM (remember Webvan?) and is going through liquidation after it barely got started.

While most commentators will focus on the money lost, and especially the massive hubris of the founders and investors (and which wildly successful world-changing startup didn't start with lots of hubris? It takes chutzpah to believe you will change the world!), I find two points interesting.

Average Profit

Over its very short period of operation, Better Place sold 2,000 Renault Fluence Z cars. Ignoring true gross margin, if it spent $800MM, and sold 2,000 cars, then its average is a loss of $400,000 per car sold. In other words, it could have saved everyone money by buying its customers Lamborghinis (even at Israeli prices). Granted, it made a little money on each car (not as much as Renault did), but with that much money at stake, it had to have sold between one and two order of magnitude more cars, or something on the order of 100,000 cars, to begin to prove the model.

Which brings me to the more important point:

Prove Your Model First

It isn't just the $800MM waste that reminds me of Webvan, but the fact that it didn't prove its model. The guru of this science of startups is of course Steve Blank, with his "Four Steps to the Epiphany" followed by "Startup Owners Manual", and his student Eric Ries "Lean Startup". I cannot begin to do either Ries or Blank justice here, but I have no doubt that they would have advised Better Place to start small. Build a minimum viable product (MVP) to reduce your costs while you find the product/market fit, prove that the model actually works.

In the case of electric cars, that involves everything: the car itself (including features, style, battery life, etc.); the market (commuters? soccer moms? work-at-home parents?); charging/changing; financial model; warranty; insurance; etc. etc. Better Place management assumed it knew what everyone wanted, which is a very dangerous thing to do at any size, but especially when playing with $800MM.

Just to give one example: a changing station is very expensive. Then-CEO Agassi said it cost $500,000 to build each, although I have little doubt that once you add maintenance, taxes, cost of the land, and especially the huge R&D to design the stations, each one was many times that. Israel alone had 27 charging stations, while Netherlands had ~18, together ~45 stations. The company probably blew $100MM on R&D and another $100-125MM building these stations, plus labour at the stations, or more than one fourth of its total capital. What would an MVP be in this case? Labour. Cut a deal with existing gas stations to use their pits. Every time a car comes in, drive it over the pit, a person swaps the battery, away it goes. Sure it isn't as cool, and probably more like 2-3 minutes rather than the claimed 59.1 seconds of an automated change station, but it would cost maybe $30 per change including rental agreement and labour (excluding the battery, which costs either way). With only 2,000 cars on the road, if each changed once per week (unlikely, since most charge at home anyways), then there would be 52*2,000 = 104,000 changes per year, or $3.12MM in costs. Better Place would have $200MM extra left in it war chest to learn the lessons and fight another day.

Unfortunately, with SAP-experienced management, car company executives and big government dreamers, their ability to do it lean was culturally constrained. Somewhere out there, someone is planning on alternative cars. I do not know if they will complement or replace the internal combustion engine. But there is someone who will test their assumptions with MVP, fail in many small ways, learn the lessons, and eventually succeed where splashy Better Place failed.