CNN Article Shows Potential of STD Wizard
On April 15th, 2010 (a Thursday), CNN wrote and published an article related
to STDs (sexually transmitted diseases) on their website.
The article title was "The downside of 'friends with benefits'".
It included a link to the STD Wizard.
We experienced a huge increase in usage of the STD Wizard resulting
from the link in the article. The STD Wizard displays web pages as
it dialogs with the user to determine which screening tests to recommend.
The average number of pages served was 1,008 per day during the previous
week. The number served during the approximate 24 hour period after the
article was 211,540 pages! That's a huge increase. We have had some other
incoming links, but nothing like the impact of this CNN article.
The server notified us by email that there was increased activity.
At first, we were somewhat perplexed, and even worried some that maybe
the server was malfunctioning or even that we might be under attack.
It was not until the morning of the 16th that an apache log file
analysis made clear what had happened. It showed the incoming links
from the article.
The server reports a "load average", which is normally close to zero.
During much of the 15th, the load average was around six, which is quite
high. However, the server performed well during the entire time. We used
the STD Wizard once as a test during the heaviest period during the afternoon,
and it worked quickly and smoothly. The apache log showed no errors.
The CNN article gave us the opportunity to conduct a much-needed
real-world stress test, and it passed.
We keep track of the number of visitors who start and finish the STD
Wizard. There was an average of 52 per day who started in the previous
week. On April 15th, the day of the article, 13,270 started!
The previous week, 310 of the 365 who started also finished, or 84.9%. Of
the 13,270 who started Thursday the 15th, 10,578 finished, or 79.7%,
5.2% lower than the previous finish rate. Some users may have clicked on
the link just to see, which would explain the 5.2% lower finish rate.
Here is a listing of the number of visitors, pages, hits, and traffic during the
period. It spikes dramatically on the 15th, and stays high for days after.
Date | Visitors | Pages | Hits | Traffic
|
13 Apr | 135 | 1,599 | 3,985 | 0.07 GB
|
14 Apr | 134 | 1,694 | 4,025 | 0.08 GB
|
15 Apr | 18,573 | 279,587 | 640,090 | 12.27 GB
|
16 Apr | 6,273 | 103,771 | 250,926 | 4.30 GB
|
17 Apr | 2,435 | 39,247 | 96,368 | 1.65 GB
|
18 Apr | 1,557 | 22,446 | 56,188 | 1.00 GB
|
19 Apr | 561 | 7,062 | 16,195 | 0.29 GB
|
20 Apr | 327 | 3,772 | 8,732 | 0.17 GB
|
The number of visitors in April, 2010:
Here's another report with the number of STD Wizard starts and finishes (user gets recommendations). This
is based on a different log file analysis system that actually
determines if a user starts the wizard, not just viewed some
introductory pages:
Starts | Finishes | Date
|
63 | 53 | 2010-04-13
|
62 | 55 | 2010-04-14
|
13,270 | 10,578 | 2010-04-15
|
4,456 | 4,145 | 2010-04-16
|
1,671 | 1,564 | 2010-04-17
|
994 | 892 | 2010-04-18
|
288 | 256 | 2010-04-19
|
156 | 137 | 2010-04-20
|
We don't have demographics on who used the program, but we think many
users are younger people at risk of having an STD. The CNN article
encouraged some 15,000 people to get scientifically-based STD
screening recommendations, based on their risk factors.
|
|