Sometimes I wish I were R. Daneel Olivaw, able to adjust anyone’s emotions for the greater good of mankind (Of course following the Zeroth Law). However until then, I have to rely on its poorer ancestor - behavioral economics. I have been a fan of the field ever since I heard Dan Ariely’s talk ‘predictably irrational’.
Since the last time we spoke of incentives, we have tried a few experiments- one successful, one marginal and have recently introduced some new ideas. As one who loves listening to his own voice, let me talk about the success.
Professor John List at the University of Chicago has conducted some pioneering field experiments using behavioral economics. We borrowed from his field experiment, “The Behavioralist Visits the Factory: Increasing Productivity Using Simple Framing Manipulations” in which he raised the productivity of workers in a Chinese factory by simply ‘framing’ the incentives differently.
In his study, List has used the risk averse nature of humans, i.e., we try to avoid losing much more than trying to gain something. Instead of earning an incentive as the month progresses, he provided the same incentives at the beginning of the month. He then deducted these incentives if required performance standards were not met during the month. In the former case, workers would just not earn something they could have if they didn’t hit required standards. In the latter, they now lost something they had already earned. Here, their innate risk aversion came into play and they worked harder to protect what they had already earned and hence there was a small but significant increase in productivity with no cost to the company.
Using the same principle we started giving all incentives in advance at the beginning of the month. Their instructions were simple – they would receive all their incentives in advance. However, every time they missed a supply notification, points would be deducted from his account. At the end of the month, we would provide them a paper receipt that would show them how many points they lost out of the total they had received in advance that month.
I’ll be honest. I couldn’t contain my excitement the day we introduced this new initiative. Even before we started receiving any results, I was in self congratulatory mode. I thought how amazing I was, to apply state of the art research that would surely accomplish miracles.
So, what would you expect? Like all predictable twists, after the end of the first month – it achieved nothing. What’s more, the valvemen didn’t even notice. I’d ask my colleague, Aseem, did you explain the “Points Lost” receipt to them and he’d say yes. What was the response – they said. “Meh! Ok.”
If the initiative didn’t improve anything, may be I would have understood, but to not even create a stir amid our test group, that was inexcusable. I started to hunt for reasons. Why? Why don’t they care? It made me question the entire program itself. What was the point if they just didn’t care at all!
The next month was marginally better in that the valvemen only asked if they could spend all these newly minted advanced points, well, in advance! A tiny flicker of hope – I allowed it. But now, they used all the points at the beginning of the month so they had nothing to lose during the month. So we introduced a requirement of a minimal account balance to be able to use points in advance.
Finally, one day, by chance, I told Aseem to create the “points lost” receipt in Kannada and told him that we should keep that standard practice from now on.
The reaction was immediate. No sooner that the first of these Kannada receipts were distributed, valvemen started inquiring why they had so many points deducted. When we would show them each instance of a missed notification in the performance reports they would resort to haggling and bargaining. Eventually, they started to comprehend their cost of each missed notification. Talk about importance of using local language!
Even so, this framing of incentives has not been the raging success I had hoped it to be. However, it has certainly played a part in improving valvemen performance. One of the major gains from the initiative has been establishing the action and consequence loop for performance. If you miss notifications you will lose points. This simple rule has been established in their minds and no one now tries to question or bargain for points.
The program commenced in February and we started receiving reactions from the valvemen in March. At the beginning of March we had 28 consistent valvemen providing correct notifications to 115 areas o the city. This number increased to 35 and 158 respectively at the beginning of May (See chart).
During this period the only other new incentive introduced was a “Valveman of the Quarter” contest which rewarded the best 2 valvemen from each tank with a one time cash award. Significantly, the winners of the contest were performing well since January. While I have not done any rigorous testing yet (I intend to write a paper with David Argente, who suggested using List’s research ) the program seems to have been successful in raising the threshold of performance across all tanks.
Finally, it has primed the valvemen for my next ace ‘in the hole’. We have just introduced what we like to call the “Smiley System”. But I am not going to reveal what it is just yet. For that you will have to “watch this space”. (Hint: SMS, emoticons and behavioral econ).