|
|
|
| ArminasX Saiman wrote on his blog about The Two Year Effect. People tend
to have a two-year attenstion span on projects such as hobbies, volunteer
work and participation in a virtual world.
Because we have a large sample of the Second Life Population, I have analyzed our data to see if this two-year effects exists.
At this moment we have more than 1,200,000 avatars in our database, for which we have birth day (the date when the SL account is created) and
the dates when they were detected by one of the visitor counters of our customers (1).
I have calculated the number of days between the birth day and the last detection by one of the visitor counters for each avatar. The first
table shows the frequency of the time avatars stay in SL. The first row shows the number of avatars that
stayed less than 100 days in Second Life, the second row the number of avatars that stayed between 101 and
200 days, etc.
| Days |
Number |
|
Days |
Number |
| 1-100 |
238012 |
|
1401-1500 |
1526 |
| 101-200 |
114849 |
|
1501-1600 |
1083 |
| 201-300 |
74101 |
|
1601-1700 |
676 |
| 301-400 |
60147 |
|
1701-1800 |
541 |
| 401-500 |
51020 |
|
1801-1900 |
431 |
| 501-600 |
40358 |
|
1901-2000 |
282 |
| 601-700 |
35323 |
|
2001-2100 |
176 |
| 701-800 |
33114 |
|
2101-2200 |
92 |
| 801-900 |
28099 |
|
2201-2300 |
59 |
| 901-1000 |
22843 |
|
2301-2400 |
24 |
| 1001-1100 |
16383 |
|
2401-2500 |
15 |
| 1101-1200 |
8204 |
|
2501-2600 |
4 |
| 1201-1300 |
3414 |
|
|
|
| 1301-1400 |
2058 |
|
|
|
A first visual expection of the results shows an exponential decline (See firgure 1). On average avatars stay
for 293 days on SL, but this number is not very informative if you look at the graph.
Chart.
On closer inspection, the graph has a poor match with an exponential curve (See figure 2). I have plotted a
trend line through the data that shows the deviation of the data to an exponential curve. When we zoom into the data
I notice that the first part of the graph shows an exponential decline (until 300-400 days), then a more
linear decline starts (Figure 3).
|
|
| Figure 2 | Figure 3 |
The data for the 400 - 1300 days have a good correlation with a linear decline (Figure 4).
Every day, 50 avatar that stayed in SL longer than a year are leaving (R2=0.98).
From day 1300 on, the decline
becomes less steep (Figure 5).
|
|
| Figure 4 | Figure 5 |
Avatars that are almost 4 years old, do leave very slowly (approximately one avatar each day).
These result indicate that we have several types of SL residents. The first group of avatars
stay in SL for a year or less, then a group of avatars that are in SL for 1-4 years and finally,
the avatars that are older than 4 years.
Why these 3 groups exist and what there behavior is in SL cannot be deduced from these results, but I wonder
if the first group are the typical socialites, while the second and third group are content creators?
|
|
|
Notes:
1. Our counters collect visitor data since january 2007. For this analysis I have only taken the visitor data from 2008 and 2009 (until 25-nov-2009).
I have excluded all avatars that were only detected once during the SL account creation. A total of 732,834 avatars were included in the analysis.
|
|
|
