Most of us adjusted backward our clocks one hour on Sunday. We had to do so to comply with the official time change, known as Daylight Saving Time. The original purpose of DST was to preserve the precious coal during WWI by adapting incandescent lighting usage to the duration of the day.
Nowadays DST is popularly regarded as a disrupting legacy practice of little benefit. Every six months we all have to adjust our behaviour, which of course is reflected in seasonal peaks of interest in DST:
We had the suspicion that many of the UK residents spent that additional time to sleep one hour more on a particularly lazy Sunday morning.
Back at the office, we were curious enough to check the impact of the time change on our metrics. We are always trying to improve how Nestoria users make the most of the property search engine. We measure that by looking at, among many other parameters, the number of visitors, their time on site and what they exactly do, particularly clicking on navigational features, property listings, pictures of houses, etc.
Impact of DST on the last Sunday in October 2009 on Nestoria.co.uk
We compared the activity on the Sunday 25th October, when clocks were adjusted one hour backward, with the previous Sunday, October 18th with the standard duration of 24 hours. The duration of the day was adjusted to identical periods of 24 hours and the number of visits normalized for the two days of comparison.
The distribution of the visitors in percentage of the total daily hardly varies beyond 7% on the total daily average, measured as the differences in the sum of percentages of visits of the visitors. Sundays are usually the days of the week with the longest average time on site, followed by Saturdays.
We found that Daylight Saving does show a significant shift in the hourly pattern of usage of the web site: up to 10% of users shifted the beginning and end of their visits during the beginning and the end of the day. Whilst the total variation is small in numbers, the shift of activity shows the expected behaviour of visits that start earlier in the day and start to decline also earlier.
No evidence that one additional hour of duration of the day generated proportionally more visits
The additional hour of the last Sunday in October increased by 4% (from 24 hours to 25 hours) the real duration of that day. The number of visits increased by 4%. It is unclear whether this variation in total number of visitors is due to weekly seasonality and/or increased duration of the day. It seems that additional availability of time is not spent in more activity online. That extra hour that day could dedicated to rest or leisure.
You may find a complete description of the analysis at Archive.org
