Pay for Performance Gone Wild

You’ve probably read about Transocean’s recent safety-bonus faux pas. In what seems to be a case of positive incentives gone wrong, executives at the offshore drilling company were actually awarded bonuses for safety performance in 2010. This despite the fact that the company’s April 2010 British Petroleum Deepwater Horizon rig explosion killed 11 workers and poured 206 million gallons of oil into the Gulf of Mexico.

In Transocean’s April 1, 2011 SEC filing (see page 35 of the DEF 14A filing in the list here), the company announced that it “recorded the best year in safety performance in our Company’s history,” and therefore the company awarded the senior management team 67.4% of the safety performance payout measure. Clearly their safety metrics did not dive deeply enough.

Establishing performance metrics can be a smart business move because they motivate people to perform better. But what if your measurements and benchmarks aren’t relevant or valid – what if they don’t actually drive the performance you desire?

Mis-aligned incentive plans and measurements are not a new phenomenon. In the book Competing for the Future (1994, Hamel & Prahalad), the authors criticize a practice they call “denominator management”, which occurs when managers are incentivized to maximize a metric such as Return on Assets (ROA) or Return on Invested Capital (ROIC). The quick way a manager can increase these measures is to focus on driving down costs or investments; the more difficult way is to focus on growing revenue. The quick approach can drive companies into the ground as beneficial expenses are cut (e.g., marketing) or mission-critical investments are postponed. While the metric target may be achieved, the company’s future competitiveness may be compromised.

A recent Associated Press article describes how increasingly savvy investors see through this ruse and can tell when profits are being boosted the easy way without associated revenue growth.

Are your incentive plan measurements relevant and valid? Are they producing the intended results?