Friday, February 22, 2008

Stop studying the problem, and just try something!

Yesterday I tried to make the case that most business situations are complex and not predictable ahead of time. Time spent developing foolproof strategies, detailed plans, etc., is time wasted.

What to do? One answer is probes, or inexpensive experiments. I've already cited Boudewijn Bertsch's posts this week on the Cognitive Edge guest blog, and at the risk of going to the well too often, I'd like to discuss him again today.

He recently posted on the efforts by the Alexandra Hospital of Singapore to reduce waiting times for patients by offering a real-time webcam image of the emergency room waiting area on the web. People considering visiting the ER could view the image on the web and anticipate how long they might have to wait before seeing a doctor.

The webcam idea was a small, cheap solution. It might not have made any difference. But the hospital tried it, and found that waiting times improved. [One explanation: at times when the ER was already busy, patients with minor complaints either waited to come in, or found another hospital.]

If the webcam had not made an improvement--no problem, just take it down, and the hospital would have written off an investment totaling $400.

Boudewijn mentions that the Alexandra Hospital's continuous improvement program used elements of the Toyota Production System. There's a lot to the TPS, but one aspect that I find fascinating is that workers who propose ideas are asked to estimate the impact of the change. Then once the change is implemented the actual results are compared to the estimate and shared with the worker. Is it surprising that the estimates get better and better? [This Harvard Business Review article illuminated that feature very clearly.]

In a somewhat similar vein to the Singapore hospital, Google experiments with new products. They put them out there, without formal launches, people discover them, Google adjusts and tweaks, and the products develop a large following, or not. In the latter case, they retire the product, again with little fanfare.

Google has gotten some criticism about the frequency of their product failures, but what is being missed in this is that they are performing lots and lots of low-cost experiments. Their unconcern for media criticism allows them to put a lot of probes out there and "see what sticks," whereas many many competitors in the IT world stick with an approach of: develop for two years, study the market, then make a splashy launch.

Google gets a lot more at-bats than its competition. If they fail more, it doesn't matter, because they are cheap failures. The number of successful products, at least to my eyes, is much higher.

So if you've got a great idea, or even merely a good one, find a cheap way to try it out. If it doesn't measure up, kill it. Repeat often.

(Photo by svaziphil via stock.xchng)

, , , ,