Auto Ads by Adsense

Thursday, October 07, 2010

Review: Chasing Stars

I've always maintained that performance is extremely contextual: treating performance as an individual issue, not as a team effort is misleading, because we don't know how important the kind of support other people on the team goes into making a star valuable. Chasing Stars is the first book that examines performance in context, and therefore I consider it a very important book, and very much worth your time if you work in a highly intellectual profession.

The context in question is that of star stock analysts. Groysberg picks this field because stock analyst performance (as rated by customers) is widely publicized in an industry standard fashion. As a highly intellectual profession, such ranked analysts are easily compared with each other, and analysts consider themselves as having highly portable skills. It is also tempting for analyst departments in various corporations to raid each other for talented employees, since having a top analyst would presumably be valuable in attracting customers.

Groysberg analyzes a 15 year period of such poaching, and approaches the data in all the directions you would think of. He looks at ranking before and after a move. He looks at ranking if an entire team moves (known as a "liftout") versus an individual moving. He segments analysts by gender. He even takes into account investor sentiment (i.e., what happens to the acquiring company's stock price after the star analyst acquisition was announced). His thoroughness lends a lot of credibility to the conclusions he draws.

The first distinction he makes is the difference between corporations that spend a lot of time training and acculturating new hires and coaching existing employees and corporations that focus on generic job training. It turns out, for instance, that the corporations that do invest in training and the acculturation process do get something for their efforts: not only are they better at creating new star analysts, they also pay a reduced cost of retention: they pay about 5% less than market for the talent that they do hang on to. Furthermore, it is harder to hire star analysts from such departments, and even when they do move, they frequently discover that their skills are not as portable as they thought: performance (as measured by analyst rankings) deteriorates for a year or two after such a move. This is quite a result.

Secondly, Groysberg also distinguishes between corporations with a lot of resources to devote to the analysts in the form of IT support, staffing help, and other networking help and corporations that don't have such resources. It turns out that analyst performance improves (as expected) when moving to corporations with a richer set of resources, but not by as much as you would expect, because the overhead of building relationships to gain access to such resources takes time.

Entire team moves seem to be extremely favorable: performance hardly ever drops, and all the employees in the team already can work together. Groysberg does note that there's also a big difference between a team brought in to exploit an existing market that the corporation always works in, and to explore a new market opportunity. The latter induces much worse performance than the former, probably because of the interaction with the sales team to educate them as well as the learning curve involved in exploring a new market.

Finally, Groysberg covers women. As in engineering, women analysts were rare on Wall Street, and had a hard time breaking into the cozy old-boy's network. Because of this effect, women analysts learned to build relationships and access resources outside of the corporation they worked in, which allows them to have the most portable skill sets --- women who switched companies did the best amongst the switchers. It also turned out that women were more likely to consider whether the culture of the new company they were joining was conducive to high performance, rather than just jumping ship for higher pay.

The big question mark here is whether any of Groysberg's findings apply to software. I certainly think that the current big company trend of buying small startups in one fell swoop and integrating the engineering team is reflective of the understanding that keeping a winning team together is important. (Though just how important is still frequently under-recognized: some companies are bought only to have their team members split apart) The "nurture" versus "nature" debate as far as engineering culture is concerned appears to also have been lost amongst many firms: many companies devote relatively little time to training and acculturation, and pay the price with reduced performance of its engineers.

All in all, this is an excellent book, and very much worth your time and money. Highly recommended, and definitely one of the few books that will be short-listed for the book of the year.


KevinKevin said...

Stock analyst performance --> rated by customers

Programmer performance --> rated by... ? X% peers and Y% managers? Lines of code? Lines of code divide by average bugs in production? # of users? Latency of the product? Number of niners? What should X and Y be? Oh man, it's giving me headache just to think about this.

Piaw Na said...

I agree that engineering performance is far more ambiguous than analyst ranking. That's why engineers should be paid in stock.

KevinX said...

All engineers are paid in stocks in one way or another. The big question is, what is a good breakdown of stock vs salary? I see 25:75 and 15:85 a lot these days, but is that far from what you envision? The type of engineer you get from 100:0 (sweat equity) and 0:100 (corporate lifer) are very different.