Red Light Cameras

Chicago has been installing red light cameras.  I thought one caught me a few weeks ago, but never got a ticket, so the flash must have been for a car going the opposite direction.  The experience got me wondering about certain aspects of the systems.

If you search, you’ll find opinions that they are effective at reducing accidents, and others that they are not.  The only definitive RLC study I found, conducted by FHWA, has interesting conclusions.  Bottom line is red light cameras reduce both injuries and costs.  The FHWA study is much better than others, but it’s important to note it’s limitations.

The FHWA study discusses spillover effect, that is, the effect where RLCs alter the behavior of drivers at intersections without RLCs.  When a city places RLC’s at a few intersections, some drivers pay attention to whether a intersection has a camera, but others become more careful at all intersections. 

What the study can’t measure, because there is no data for it, is what the effect would be if all drivers acted like the second group, that is, they drove more carefully at all intersections.  This condition would occur if all intersections had cameras or a high enough percentage as that drivers assumed they were present without looking for them.

This limitation is important because all the benefits of this homogenous state, compared to the current mixed state, favor RLCs.  First, in the mixed state, drivers intentions and expectations vary more than in the homogenous state.  If a driver who knows a RLC is not present at an intersection is following one assumes all intersections are equipped, the following driver may assume the leading driver is less likely to stop than he is, increasing the risk of a rear-end collision.  If all drivers have similar goals, the risk is reduced. 

Second, if drivers look for RLCs as they approach an intersection, this is a distraction, and increases the risk of any type of accident, especially rear-end collisions.  If they assume the camera is present, plan to stop, and focus their attention on traffic and signals, the risks are lower.

The point is the proven, but modest benefits the FHWA study finds are underestimating the full value, and while we may each dislike the chance we’ll receive a ticket, ultimately they do save lives, injuries and costs.

Google, Brand and Beta

Reading Dare Obasanjo’s thoughts on how layoffs and project trimming at Google affect its culture led me to thinking about why the fatality rate for side projects at Google has risen so high.  It’s not just the economy, it’s more than that.  I think despite being the major innovator in web based advertising, Google may have failed to understand the meaning of brand internally.

What Google has been loathe to admit to itself is that a not insignificant part of its success in many areas is something other than the quality of their products.  Not that they don’t have quality, or that quality wasn’t important, but many of their products have succeeded by leveraging the brand built from the success and popularity Google web search and the following press obsession with the periods of rocketing stock prices.  This brand was critical for products like Gmail which asked users to change the way they work.  Gmail asked users to give up something in exchange for something else.  Products like that are a much harder sell then products that provide incremental improvements, and the ability to convince users to temporarily ignore their losses until they understand their gains is critical. 

That ability is called trust.  Google however has damaged their earned trust in somewhat unexpected ways.  Maintaining trust isn’t as simple as not breaking your promises, because people aren’t fully rational about trust.  Causing confusion is as reliable a mechanism to lose trust as lying, and Google has done a good job of confusing users with their extremely long list of beta products.  Sure, they are called beta’s and the implicit statement there is that it’s not complete, but since some of their beta products have had quality far above others, the meaning of beta changed.

What Google would benefit from here is to do a much better job of internally vetting the state of quality of their products and projects and communicating that to their users and potential users.  Currently I have a feeling that Google’s main metric of quality is the number of daily users.  That is a poor metric that is self-defeating in the brand arena.  Those products that have succeeded, and are already successful don’t need a stamp of approval, but those that do are left among the rather large and growing bucket of products that either haven’t yet made it, or never will.