Something Screwy Going on With TweetGlide
by Antone Roundy | 17 Comments | Product Reviews, Social Media/Networking, Viral
[Update: Check the comments -- there's a reasonable explanation after all.] Mike Filsaime's new Twitter/Facebook client app has generated a lot of buzz the last few days. For the most part, it looks like a great idea and a useful app (I've noted some of the problems with the beta version in my previous post and comments on it). But this morning, I'm seeing something...screwy.
My server logs show 934 hits in the last 24 hours where the "REFERER" header is set to "app:/TweetGlide.swf". Most of it (918 hits) is going to the URLs of the ads I have running. Great, right? I'm getting traffic from the ads.
I don't think so.
When I check my stats, I see that my ads have shown a total of 3 times. I don't even have anything vaguely approaching 934 advertising credits. Unless everybody's getting a TON of free ad impressions, that can't explain it.
The hits aren't coming from just a few IP addresses, so it's not the 3 people who saw the ad reloading the page a gazillion times each.
Perhaps most telling is that 700+ hits on one page have resulted in zero sales (WAY below my normal conversion rate), and 200+ on the other have resulted in at most 2 or 3 free signups (possibly none -- I haven't bothered checking whether any of the people who signed up in the last day came from TweetGlide).
What gives? I can only guess (what follows is pure speculation).
Is it possible that the app is secretly loading webpages in the background but not displaying them, even from the computers of users who don't see the ads?
Why would Mike do that? To get TweetGlide's name to show up prominently in peoples' referrer logs (some of which appear automatically on peoples' websites)? To make people think TweetGlide advertising brings in more traffic than it really does?
I could be totally wrong. All I know is I'm getting a suspiciously high number of hits that aren't leading to any conversions. Anybody care to explain?
Make that 942 in the last 24 hours.
October 30th, 2009 at 1:09 pm
Shenanigans Confirmed?
I checked my server logs and verified that when these hits from "app:/TweetGlide.swf" come in, ONLY the webpage is loaded. None of the accompanying CSS, images, scripts, etc. are loaded.
What this proves is that these hits are NOT from people clicking through my ads and viewing my webpages in their browsers.
Unless Mike comes out with a reasonable explanation for this, it looks to me like TweetGlide users are being used as sort of a botnet to advertise TweetGlide using server log spamming.
I'm turning off TweetGlide and recommending against using it until and unless I hear an acceptable explanation.
October 30th, 2009 at 2:34 pm
Hmm, just noticing that a large portion of this traffic came in one burst 24 hours ago. But it has continued more slowly since then.
October 30th, 2009 at 2:35 pm
Does seem strange
In the last 90 minutes I have been hit 15 times.
Of those 9 were calling the page with my additional ad parameter, but I have seen that retweeted as well.
So far today in the last 20 hours, I have had a total of 7 visits from my advert
It isn't just checking headers, it is downloading the whole page.
That isn't the worst of the bots that come to visit
October 30th, 2009 at 2:51 pm
TweetGlide
Hello!
I had a problem with Tweetglide earlier today. I downloaded it [URL removed] it worked for a while but then the ikon of Tweetglide disappeared from the screen and I couldn't get in again. But I got a black window with a message that said "Verification of credentials V0.1.1.38". That went on for about 20 minutes, and then I gave up. I uninstalled the program. And now I can't download it again, I just get the same message over and over again that says "Verification of credentials V0.1.1.38".
I checked the supoort message for Tweetglide and there it says:
Tweetglide
Publisher: UNKNOWN
Version: 0.1.1.38
something is wrong
October 30th, 2009 at 2:59 pm
Baldur,
That looks like beta software problems. I saw things similar to some of what you describe. It cleared up after rebooting my computer.
October 30th, 2009 at 4:04 pm
"That isn’t the worst of the bots that come to visit"
Right. It's not causing any real harm in terms of bandwidth usage. The more I think about it, the more it looks like a way to artificially inflate members' perceptions of the value of advertising in TweetGlide.
The fact that about 400 of the hits I counted came within about 20-30 minutes, followed by a smaller burst nearly 24 hours later, suggests that the feature may have been designed to trickle the hits in gradually, but perhaps when everyone upgraded to version 0.1.1.39 at about the same time (wasn't that about when that happened?), the traffic went out in a bigger burst than intended.
October 30th, 2009 at 4:08 pm
Antone--
All I can say is that I am glad you are on the case.
I've been reading about botnets lately, where some outrageous percentage of computers are being "used" to run spam, etc (or worse) without the owners knowledge.
It's funny, I haven't deleted Filsaime's endless supply of emails--from him or his legion of followers-- because I just want to keep tabs on what he's up to, but in THIS case....the "unsubscribe" button is just minutes from being pressed.
I will keep hope alive until I get some kind of confirmation, then sayonara.....
October 30th, 2009 at 9:19 pm
Why not approach the owner of Tweetglider himself? If no answer or wrong answer, then you know enough. I can hardly imagine that he did this on purpose. That would be stupid, isn't it?
October 30th, 2009 at 9:38 pm
I've submitted a question to the helpdesk. I'll comment here if I get a response.
October 31st, 2009 at 3:07 am
Hey Antone,
I am not sure what could be doing that. But I think it maybe coming from the page that shows the add. It is an RSS page (if I even know what I am talking about.)
http://tweet-glide.s3.amazonaws.com/tweet.rss
Now, if that page is somehow crawled or something by our site or something else that sends a ping to your site, I will find out for sure and post it here (or my Tech guys will.) I am on it now.
I can promise you this, it is not TG trying to show "Traffic" to its members. You want traffic that CONVERTS not traffic, The last thing we want to be known for is junk Traffic.
Before any conspiracy theory emerges LOL, please let me get back to you on that ASAP ;)
Thanks
Mike Filsime
October 31st, 2009 at 8:28 am
Here's the response I got from tech support:
"This is caused by the fact that when someone hovers the link in your ad the application loads the page title for the page the ad links to. We are working to find a solution for this."
When I first saw that, I wasn't too sure whether to believe it, because there was still the question of why I was getting over a thousand hits when my ads had only shown a few times.
But from Mike's response above, it looks like one ad credit puts your ad in the ads feed once, which potentially gets it viewed by a bunch of people.
I imagine the reason for doing it that way would be so that the ads can be loaded from Amazon S3 rather than putting a potentially huge load on Mike's servers. That makes complete sense.
Perhaps the answer then is to have the TweetGlide server load the page to get the title once each time the ad is placed in the feed (in case it has changed since the last time), and then output the title in the feed so that the client app doesn't have to load the page to get it.
Looks like I've got to "print a retraction." Apparently it's a design flaw with unintended side effects.
October 31st, 2009 at 1:13 pm
Just another example of folks messing up the internet experience for their own gains. You see tip of iceberg Grasshopper! Let's also ambush those who stick those exit boxes in our faces as we leave their scammy, spammy sites and do a thang on their stanky legs for that amateur kidnapping of your time thing they do. Was it worth getting the search engines to knock you down as they just did for being the traffic whores you have just become?
October 31st, 2009 at 1:22 pm
David,
If your read the last two comments, you'd see that this was a side effect of a design flaw, and that they've said they're going to fix it.
If the exit boxes you're referring to are the JavaScript alerts, I agree that those are an annoying waste of time since it's impossible to leave the page without clicking to dismiss them.
Personally, I don't have a problem with exit hover ads, which are just HTML content displayed within the page, which doesn't prevent you from leaving. They're an effective tool to grab attention and deliver a small, focused message and increase conversions without forcing people to jump through hoops.
October 31st, 2009 at 2:41 pm
Thanks Andy for the retraction. When I read your email yesterday I was some what dishardened.
I like tweetglide and telling everyone who'll listen about it. I appreciate you taking the time to find out what the culprit was and also for Mike addressing your concerns.
Thanks for being a watch dog.
October 31st, 2009 at 11:09 pm
I just saw this in the release notes for the latest update to the app:
"* Fixed: Not loading content pages with titles in the ads glide."
If I'm reading that correctly, it sounds like this issue has been fixed already. The "phantom" hits coming into my site appear to have stopped to. Kudos for the quick response.
November 1st, 2009 at 8:00 am
Very valuable discussion here. So now the problem is fixed I would like to ask: shall I buy it or not? I have still a bit doubt.
November 2nd, 2009 at 10:30 am
I'll share my thoughts on two questions: whether to use it, and whether to pay for the account upgrade.
* Even if you're not sure whether you're going to use it long term or not, I recommend downloading it now, getting your referral link, and talking about it (with your referral link). Why? Because the people who start getting referrals first are the one who'll build huge downlines and get tons of free advertising. The longer you wait, the less potential you have for building a downline.
* I paid to upgrade. One of the main reasons I see for paying is because your ads will be twice as big and much more eye-catching. I have no stats to prove whether or how much more effective that'll make them, but it seems like a no-brainer that they'll do much better. Whether you think it'll be worth the 1-time investment is up to you.
(Disclosure -- the link above is my referral link. If you use it and download the app, you'll be in my downline, and if you pay for an upgrade, I'll earn a commission).