Or: Where Pay for Performance Goes Bad
A friend of mine sent me an article recently that had me pulling out my hair – and I don't have much hair left to pull out. The subject was pay for performance, and I responded with a piece in the Electronic Recruiting Exchange, seeking to counter the madness with some common sense.

The result sparked an instructive debate on the board (with footnotes!). Because performance-based pay at the executive level is once again in the news, I thought I’d share the original piece here, along with some of the ensuing comments.

To begin: It seems a 3,400-person bank in Ohio had a big problem with turnover among its check-proofers (a low-paying job that combines the benefits of tedium with the rewards of low pay and mission-critical accuracy). According to bank managers, proofers were "vital to ensure the correct amount of money is moved from one account to another... The bank's reputation and the customer's satisfaction depend on it."

Someone's innovative "solution" was to institute a pay-for-performance plan. In fact, the spokesperson said the pay plan worked so well that they intended to implement it throughout other departments. As the Church Lady on Saturday Night Live used to say, "Well, isn't that special?"

Is Compensation a Science? Yes, and No.
I totally agree that good performers should receive good pay. But most research shows that pay only has a temporary effect on performance. Also, tinkering with pay programs can be like getting caught between a bus tour and a casino buffet – hazardous to your health.

Does such tinkering even work? One reader pointed to the famous “Skinner box” psychology experiments. Citing the logic of operant conditioning, or response-stimulus manipulation, she argued that by carefully regulating compensation, organizations can indeed engineer performance in a fairly effective way. Yes, to a point. But what works for lab rats turns out to be more complicated when you’re dealing with people, as studies have abundantly shown. Human cubicle-dwellers get far more emotional about pay scales than their caged-rat counterparts.

Economics is also regularly applied to compensation issues, with similarly mixed success. Most pay programs more or less follow the laws of supply and demand: More demand for a specific job combined with fewer qualified people yields higher pay. But – and I mean BUT – economic laws assume every comparable employee has equal skills and is equally productive. Anyone who has ever worked in the real world knows "equal skills" are pure fantasy. Employees naturally tend to separate into top-scale people, middle-of-the-roaders, and bottom feeders.

Traditional practices, traditional failures
In most cases, this bell curve follows roughly a 20/60/20 ratio. But did you ever wonder WHY?

Traditional practices are amazingly ineffective:

• People (even professionals) do not systematically compare detailed pre-hire interview data with detailed post-hire performance, so they believe that interviews work.
• Human decision-making flaws encourage hiring managers to predict job performance based on social impressions and impact, instead of empirical data.
• The same people "forget" that only about 50% of employees who pass hiring screens become good performers. No news about new-hire job performance becomes good news.
• And worse: because no hiring manager wants to publicize a 50% success rate, they often rewrite history by changing expectations to match reality.

Best practices can help:

• Wake up and smell the coffee. Effective placement practices are not a learn-as-you-earn activity. Study as much about best hiring practices as possible. The Uniform Guidelines on Employee Selection Procedures are a good place to start.
• Thoroughly understand job competencies by doing a rigorous job analysis.
• Choose professionally-developed tests that accurately evaluate specific competencies.
• Make sure every test score is causally associated with some specific aspects of job performance.
• Measure each competency at least twice using two different methods to reduce error.
• Follow up and monitor the system, tweaking it as necessary.

Did you notice that "pay for performance" was not included as part of a good selection process?

Are commissions the exception?
“Wait, what about commissions?” I hear you ask. Several readers joined a debated about commission-based employees in general, and recruiters in particular. Sure, these populations seem to argue strongly that in some cases, even 100% pay-for-performance works. However, the truth is, these folks have actively sought sales-type jobs where they can “put it all on the line” with the chance of a very big upside. They have quite different motivations and drivers from people who seek salaried jobs – and their own process of self-selection, by definition, is nearly 100% accurate.

When Pay for Performance Goes Bad
Some horror stories:

• One company paid incentives for orders. Sales went up. Returns went up.The company implemented a restrictive return policy. Clients became very unhappy.
• Another company paid their programmers for catching software bugs. Payroll increased. Software bugs also increased.
• Auto mechanics were paid based on the size and number of their repair bills. Customers complained about high prices, unfixed problems, and unnecessary repairs.

Several psychological processes tend to drive this kind of behavior:

• People tend to slack-off when they feel underpaid. That is, they adjust their productivity to compensate for low pay. One reader pointed out the danger of giving incentives only to high performers – the “awards create losers” syndrome.
• This works both ways: when pay is tied to group performance and individual performance is ignored, group productivity tends to fall. High performers tend to resent slackers.
• What is measured (and rewarded) tends to get more attention, while everything else takes second priority. A reader noted the importance of understanding what’s truly important for your business, before you choose how to pay for it.
• Individual pay for performance generally leads to self-serving behavior
• Innate attitudes and beliefs, or a history of poor treatment by management, can lead to unproductive work behaviors such as workplace violence, low integrity and theft.

Bottom line: Pay is a very complicated process that often leads to unexpected consequences.

Recommendations
I have two recommendations that are so simple they almost seem complicated: 1) use best practices to hire right-skilled employees and managers, and 2) get out of the way and let them do their job.

Best practices ensure that marginally qualified people are screened out pre-hire, instead of later. They also maximize diversity by basing decisions more on skill measurement than impression management.

Getting "out of the way" means training, managing, and evaluating people based only on the job they were hired for, and then rewarding them fairly. It also means searching for and removing silly obstacles that inhibit work.

A bank-wide pay-for-performance system? It won't be a pretty sight. Chickens produce better eggs when they are healthy and fed a balanced, measured diet. More feed only leads to fat fowl, not more eggs.