Asymmetry, measured in milliseconds
Speed to insight, speed to do something with the insight...
It might have blown our minds when we heard that being on Wall Street itself was not the advantage - it was being closer to the data transmission. When Michael Lewis wrote a book on this, the idea that receiving information milliseconds earlier than 'the other guy' could be in any way useful was remarkable, until we learned how high-frequency trading works... One 'hack' to gain competitive advantage was a non-clever-but-elegant solution: move closer to the source, gain microseconds on the competition and execute a trade before the others do.
His new book Flash Boys is about the form of computerized transactions known as high-frequency trading, in which the fastest computers with the highest connection speeds get the information first, and make the trade before anyone else can. A millisecond — even a nanosecond — can make all the difference between how much money is made or lost on any transaction.
You'd be surprised to hear what investment banks do to get that nanosecond edge, and how they often use it in ways Lewis describes as predatory. The victims range from some investment houses to individual investors.
'he's sitting physically in lower Manhattan when he makes his trades. When he pushes the "buy" button, the signal from his computer travels up the fiber optics along the west-side highway of Manhattan and through the Lincoln Tunnel. On the other side of the Lincoln Tunnel is one of the 13 stock exchanges, called the BATS Exchange founded by high-frequency traders.
They're sitting there, and they get the signal that he wants to buy first. ... They can see what he wants to do. They discern his desire to buy Microsoft, and they have faster connections to the 12 other exchanges that are scattered across New Jersey, and they race him to the other exchanges, buy all the Microsoft in front of him, and sell it back to him at a higher price. ...
He was physically being raced by high-frequency traders who had faster and more direct fiber optic lines from the first exchange to all the other exchanges'
10 years later, systems to remove that steroidal source of advantage have still not been agreed - there are many perverse incentives that are keen to maintain the status quo.
What do you need to know? When do you need to know it? Because it would be rare to have the second answer be 'later' or 'when everyone else does', it forces the emphasis onto the first question, and being smarter.
At IDEA and Protodigm, we're very keen on our methodology of Intelligent Decision Economics. Keen on it because it introduces money into the consideration. As Michael Schrage wrote, 'the thing about Google isn't that it gives you great searches; it is that it does it 'instantly'... the same answers a day later wouldn't be something you'd want.'
Intelligent Decision Economics says: is it worth paying A to find out B, or would paying X to find out Y be a better investment? It says: what could we know in a day, a week, or a year? What could we know 'enough' vs in more depth? (This is a version of Colin Powell's 40-70 rule:
The rule states that you need between 40 and 70 per cent of the total information to make a decision.
With less than 40 per cent, you will likely make a poor choice, and with more than 70 per cent, you will end up taking too long, and the decision will be made for you!
(Let's pause on that for a moment: in pharma, we never have more than 40% of the total information available when we make a decision, even in a phase III-go decision. Wind that back to early phase, and you could take at least a zero from those percentages. Early phase cannot be based on the typical prior decision - 'our CD73 is going into disease x' - it has to be regarded as pre-decision, as exploratory.)
When resources are finite, and they certainly are in early phase, this becomes a critical planning tool. The estimated value at point B or point Y is important, as are input costs A and X. Equally, we need to acknowledge uncertainty - despite spending A, you might find the answer is not-B. This is the reason we call this 'decision economics' - what is the decision to be made, and therefore what would it be useful to know, knowing that the investment in finding out can be optimised. There's no 'perfect', but there is 'better'.
Advantages in pharma, in knowing before the competition, are not yet measured in milliseconds. But they are measured in billions of dollars. Remember, that asymmetry might not just be in 'does compound N hit target P?' but 'which other targets might it hit?', 'which other spaces exist in that market that could get us there faster?', 'which regulatory path could we pursue?'
Importantly, it is not just the presence of information, it is about the system that exists to exploit information, and a system that recognises the limits of the information. As a British minister said about the recent events in Afghanistan, "I've seen some lines about the failure of intelligence. History shows it's not about the failure of intelligence, but about the limits of intelligence.'
Milliseconds only count if you have a way to trade in the microseconds before your competitor. In pharma, finding out information that your competitor doesn't have, on a molecule within a class, is only valuable if the system you have internally is set up to pivot in the presence of that information, or was set up to look for it in a way that your competitors aren't. Either way, the learning process is critical.
Remember, stock trades are simple: they're simply directional. Knowing what to do when you see something can be delegated to a computer. Pharma development is vastly more complex. But, the situation remains the same: knowing first, and doing something about it, remains a competitive advantage. Just be ambitious when it comes to finding out what to know, and how to find out.