Sin of Incentives

I admire Joel Spolsky's writings, whether it be at www.joelonsoftware.com or Inc magazine. No wonder I loved the latest Inc article he has written on "Sin of Commissions". He is essentially talking about why putting sales commissions could result in really bad end results. I couldn't agree more. I strongly recommend reading the article in full.

I have also had related experiences, mostly bad. I have worked in three companies prior to moving to Yahoo. Two of these were pre-IPO startups and the strongest incentives they could give were stock options, which is what they did. The third company was an established non-public-listed company and hence stock options did not make sense.

To incentivize its engineers, they had a bonus scheme. These bonuses would be paid out to the whole team if the team met its own goals and performed well. These were generally some sort of revenue/profit goals. Now, here in the lies the problem. Most technology products take significant time to get to revenue making stage. Websites, for example, take months before you can make profits off them. However, all these months are not spent idling around. You need to build the product, iterate on the features, work on crashes, stabilize the problems and only then are you close to making money.

Now, it is not necessary that the engineers in the team remain constant. Some are moved to different teams because of resource planning, or due to promotions, or due to change in profiles, and some just leave the company. So, the guys who get the bonuses (when revenues/profits start coming in) are not the same guys who helped build the product from start.

Moreover, if the product is already stable, engineers are mostly spending their time keeping things stable. All work comes through strict Change Request processes and are triaged before accepted for work. Often, these maintenance cycles are relaxed points and engineers can easily manage it working 9-5.

On the other hand, early stages need people working much harder to build the product (from scratch mostly), meeting tough early deadlines, meet moving targets, and handle all sorts of crashes happening at night times, because the product isn't stable enough or the monitoring tools haven't matured.

So, this means guys who are working in best conditions are paid out the bonuses, while guys working their asses off get none of it. This could be a big mess.

The normal reaction is - don't pay out team bonuses, but only stock options, so you get paid only when the company succeeds. This works in pre-IPO stages because the strike price is normally trivial and you get paid if the company goes big. However, for an already listed company, your strike price is established based on when you join and hence it is highly randomized.

Essentially, if you join when the market is at its peak, you get stock options which you can never exercise, while others who joined a few months later can just loiter around the office and make thousands off their stock options, which were priced too low.

Off late, many companies try out Restricted Stock Units (RSU) which have nothing to do with the point of time you join, but you payouts depend on when you sell these units, which is a far more controllable piece for most employees. I have been issued RSUs at Yahoo and have a 3 year lock in, which hasn't completed, and hence can't comment on how well it works.

Any comments?