Discussion:
Weird download behavior--bots?
(too old to reply)
Joseph Rosevear
2024-08-16 08:08:56 UTC
Permalink
I have a website at https://rosevearsoftware.com. I use it to distribute
the free software that I write. At least I'm trying to use it for that
purpose. I recently noticed odd download behavior. I'm wondering
whether this is due to bots? And if so, then what can I do about it?

I have a product called SAM-1.3. It is a menuing tool that works in a
special way. (You can read about it on my website, but that's not what I
wanted to talk about.) Between June 3 and June 24 my counter registered
about 67 downloads or more than 3 per day. Then suddenly they stopped
and I had four days of no downloads followed by a gradual rate of 13
downloads in 35 days--a little more than one every days. That's 1/9 th
the original rate. The graph is striking. It wiggles around, but not
that much. Underneath the wiggles you can see two straight segments and
a "knee" in the curve.

There is another product on the website that has a download counter that
generated data with the same basic shape--again with a distinct knee.

It could be a result of my not having uploaded any new versions recently,
but still that seems to me not enough reason for this behavior. The
world is a big place. It seems to me that a saturation effect, caused by
interested potential users finding and downloading my software, would not
show a knee like this. I'm thinking it is *bot* behavior.

My counters--homemade php code--already check the IP address against a
list. I did this to avoid counting my own downloads that might happen
from one of several places. So I'm thinking I could get a list of "bad
IP addresses" and check those also? That would make my list much longer.

Hmmm.

Any suggestions?

I feel a little foolish. I was happy to get so many downloads.
Obviously, I need to harness Ai to keep the bots out. Ha!

-Joe
Henrik Carlqvist
2024-08-16 18:04:18 UTC
Permalink
Post by Joseph Rosevear
Any suggestions?
Maybe you should not only compare adjacent weeks in the graphs, maybe you
should also compare this years weeks with previous years weeks. During
this time of the year, some parts of the world has their summer vacation.
Maybe that could affect the number of downloads?

regards Henrik
Joseph Rosevear
2024-08-19 09:17:20 UTC
Permalink
Post by Henrik Carlqvist
Post by Joseph Rosevear
Any suggestions?
Maybe you should not only compare adjacent weeks in the graphs, maybe
you should also compare this years weeks with previous years weeks.
During this time of the year, some parts of the world has their summer
vacation.
Maybe that could affect the number of downloads?
regards Henrik
Thanks!

That's a great idea, because I really *don't* know what is happening. It
is best not to assume an explanation.

Ah, but it will take diligence to either automate my data gathering or to
persist in gathering the data manually--which is how it is currently done.

-Joe

Loading...