Extreme expectations

Original post:  Mar 31, 2014

We are often quick to adjust to the new normal. We gain a shiny new object that holds our fascination. In a few weeks, it starts to show some wear and tear. At one to two months, it is no longer the next new thing--if it's in regular rotation, then it is just another widget to us. After six months, it almost seems passé.

Smartphones are about six years old. What is amazing is how performance that previously would have been unthinkable now drives us to distraction because it is unthinkably slow!

This blog post in the NY Times discusses some of the increased expectations for the estimated one billion smartphones across the world. It references data from a company called Crittercism that analyzed three billion interactions with apps over the course of one month. There is a staggering amount of complexity in these relationships:

It’s useful to start by thinking about the overall network, with all its phones and carriers. That would include, say, an Android device being used in India, a BlackBerry on Wall Street, or an iPhone in Germany. Once you take into consideration all of the different phone models (2,582), carriers (691), and versions of operating systems (106), there are over 100 million possible permutations of how a signal might be processed.

Those apps are also changing dynamically. Many rely on cloud data. Some rely on multiple different sources that have to be combined and processed simultaneously. To top it off, it has to do all of these things between the time you press a virtual button to kick off the service and before your patience runs out.

Response time in the US averages 320 milliseconds. Canada actually has faster response times--about 260 milliseconds. India is at the other end of the scale at about 650.

For app companies, there is an even greater threat:

The real issue for companies, however, is not load times as much as failure rates. When that happens, customers are alienated and no business happens. A company with $1 billion in mobile revenue would lose $2.5 million a month with a 3 percent failure rate. Even losing one one-thousandth of the time is an $82,000 loss.

Not surprisingly, services that people pay for and services associated with making money fail less often. Google Analytics has an error rate of 0.1 percent, Crittercism said. Facebook, on the other hand, came in at 11th place, failing 2.5 percent of the time. Gaming apps, which are often free and usually very complex, in general crash 4.4 percent of the time.

The article concludes with a summary of the way we've come to demand the impossible:

On the web, applications now work (or show up, anyway) about 99.9 percent of the time. And that is one of the problems in the perception of mobile: People are used to that standard, and they don’t fully realize how much more complex this system is, compared with a fixed-line Internet.

So yes, you’re living in a technological miracle. And yes, you think it’s terrible. And no, that’s not unfair. After all, the people who gave it to you created these expectations along with their products.

Here is a link to the original source:  http://bits.blogs.nytimes.com/2014/03/27/smartphones-the-disappointing-miracle/?_php=true&_type=blogs&_r=0

Living up (or down) to expectations

Original post:  Sep 17, 2012

I heard a fascinating story this morning about expectations. An academic study focused on how teacher expectations might impact student performance.

The study told teachers in advance that there were certain students in their incoming class who had the potential for high achievement. The teachers did not know that the students had been selected at random and were no more (or less) likely to be high performers. At the end of the study period, the researchers found that the students pre-selected for high achievement actually outperformed their peers!

The researchers did a second study to try to find out why this had happened. It turns out that it was the accumulation of hundreds of smaller moments along the way. If a question was asked of a child, the response would depend on the expectations. If the child was considered low-performing, they would be given a short period of time to respond before the teacher would move on to another student. If the child was in the high-potential group, they would be given an extended period to answer (or possibly even prompted). The higher potential students were also given a bit more coaching along the way. Those little interactions ended up making measurable improvements over the course of a school year.

In the same way, this may be applied to our own interactions with our colleagues. There may be times when we are frustrated with other departments (or even with colleagues). If we can try to assume that our colleagues are working for shared interests, we may be more likely to give them the benefit of the doubt. We might actually help improve our own service in the short run. In the long run, we could help our colleagues improve as well!