Advertisers Escalate Banner Ad War

AnonymousComrade writes: "In today's Newsbytes, there is an article about MediaBEAM GmbH, a German company that say they have developed Web server software that can detect whether a home browser is blocking banner ads or pop-ups. If the Web server detects blocking software, a message appears on the screen advising the 'free-loading' surfer that he has two choices if he wants access to the Web site's content: pay for it or be exposed to the ads. This sounds strange to me. Can they really include something in the download (Java or JS, I assume) that detects whether an ad picture has been downloaded or not? What if you have blocking S/W that not just blocks the download of the ad picture, but also modifies the HTML on-the-fly (a la the Proxomitron). Can they really distinguish this from a remote ad server that just isn't responding? And how long will it take before ad blocking S/W is updated to block this blocking-detection mechanism?"

Comments

And what about text/speaking browsers?

[Masem]

Some browsers don't support JS, and cannot download images (eg: lynx, or browsers for sight-impared people). At least in the latter case, there's a legit reason for them to not have images.

In addition, in today's age of worms upon virus upon other nasty things, there is a sufficiently significant (probably around 10%) of users that have turned off Active Scripting in IE or the equivalent in NS to avoid such problems. I very much believe that these users have more of a right to keep this off than an advertizer has to force you to look at an ad.

Arms race prediction

[DickBreath]

You can do this purely on the server side. No cookies. No by IP address. No javascript. No Java. Possibly without even using frames.

First, with frames. Send them a tiny page with frames. A frame for the ad(s) and a frame for the content. The tiny frame containing html contains custom url's for both the content and the ads. That is, the "session id" is embedded in the url's, without using cookies.

When the browser requests the url for content, the content stream is stalled until the ads are downloaded, or at least started. Since ads and content are tied to the same "session", you can tell which content goes with which ads. But you don't penalize others behind a NAT. The server can still be load balanced because a database keeps track of the sessions -- which can be very short lived. So even multiple servers can be used, as long as they share a common database -- or some rpc mechanism to ensure evil ads have been served before "unstalling" the stream for the content.

I said do it without frames. Simply send the html stream of the main page. Ads appear in the stream before content, which almost means necessarily "above" the content. When the stream gets right up to the point where it is to start delivering content, you stall the stream until the ad image(s) are at least requested.

Possible problem: are there any browsers that cannot request the ad image while the main content page is stalled? i.e. non multi-threaded?

Possible countermeasure: when your junkbusting proxy detects ads, it must deliver fake ads to the browser (or better, rewrite the content stream so that there aren't even ad spaces in the content), and it must make a pretense of requesting ads from the server. The proxy would continue to suck down the ad images until the content is delivered -- then abruptly close the ad stream connections. This way, if the server isn't willing to unstall the content until the ads are fully delivered, all you wasted is the bandwitdh to get the ads, but you don't see them. If the server is willing to unstall the content as soon as ads are requested, then you drop the connection on the ads asap. Using such a proxy, the server is unable to detect that you didn't actually see the ads. You at least went through the pretense of downloading the ads.

I don't see any counter-counter-measure that the evil advertisers could employ. From their point of view, you are a normal browser, downloaded both the ads and content. How can they further tell that you can't see the ads without going to more invasive techniques like Javascript?

I've often wondered about using javascript to deliver the content. You send down a javascript program that writes new content into an <ilayer>. But the javascript can be obfuscated. Even the "content" can be compressed with the javascript effectively unzipping it as it writes content into the layer. This almost certianly requires real javascript running in the browser to render the content part. The javascript could attempt to detect that the ads have been rendered first.

Now the counter-counter-counter-measure. Let the javascript and rendering happen in the ad busting proxy. The proxy is designed so that its rendering engine renders a data structure in memory. You then run filters on this data structure. Pattern matching. (Lisp anyone?) It's like a regular expression, but without the same kind of syntax. You do the recognition on the final page, which is expected to be structured a certian way. The ad, which falls in a familiar place is removed, and then new html is written from the in-memory data structure rendering of the original html. The new html is sent from the proxy to the end user.

Then what about a counter-counter-counter-counter-measure? Well, the evil advertisers could start sending you the content as a java applet. The applet contacts the server via. a non http stream and gets the content through a secret non-standard means. But only if the ads were delivered.

But then the counter-counter-counter-counter-counter-measure is to run the applet in a faked environment that fools it into connecting to the server and think that the ads were deliverd on the local page.

But then the counter counter counter counter counter counter measure is to serve both the ads and content together in a single big applet which uses a proprietary non-standard means obtain both the ads and content over a special stream from the server. Any attempts to circumvent this is a violation of the DMCA. They automatically record your IP address, look up your location, and to an XMLRPC call to the local FBI office's server to send goons to your door.

Then the counter counter counter counter counter counter counter measure is to stop visiting such sites. (And to bitch and complain on slashdot.)

Then the counter*8 measure is to lobby for custom legislation that requires you to browse to their web site if you were a regular visitor before, and to watch their ads. Alternately they can send party comrads to your home to force feed you the ads.

Finally, you must download and install their TeleScreen(tm) applet which uses your computer's usb camera and microphone to give them two-way telescreen access to your home to ensure that you are watching your dialy minimum recommended allowance of ads, as determined by federal standards.

What is the counter*9 measure?

Current web advertising unacceptable

[Chris Johnson]

The current state of web advertising is comparable to TV advertisements that come out of the TV, walk over to you or follow you into the kitchen or bathroom and tug on your sleeve whining 'buy me!'.

That's bad advertising, plain and simple. It's been _proven_ through marketing research that if you get too annoying you unsell your product, most notably that's been proven through market research of particular sorts of TV advertising.

We're not talking about 'just let these poor people make money, will you?', we're talking about enabling them to hose themselves through severely stupid and bad advertising. The people using ad blockers are doing advertisers a FAVOR- that is valuable data, that information. They often accompany this with other valuable data- announcements that "if it was all quiet well-behaved banner ads that didn't blink or flash or move a lot, we wouldn't feel compelled to be doing this". That's valuable information. Since when is a random consumer's browsing history more valuable than an outright, impassioned statement of that consumer's preferences on how they want to be courted, advertising-wise?

Intrusive web advertising can be compared to billboards: the people attempting to use it can make a big fuss about how it's a moral imperative that they should be allowed to do this, but it's not only a lie, it's not even a healthy or useful thing to be doing. They are wrong in wishing to do it. If they are allowed to do it they will actually harm advertising in general- though this does create a window of opportunity for well-behaved advertisers, as well as substantially driving down the costs for well-behaved advertisers. Still... if you don't actively hate the entire field of advertising, it's hard to justify these abusive, useless practices, which harm advertising in general.

David Ogilvy considered advertising the art of 'speaking well about' things. Abusing people to the point that they are blindly, acutely hostile to anything resembling advertising makes it that much harder to do it properly and sensibly.

Just as restrictions are placed on the use of roadside billboards, I would like to see this abusive web advertising restricted by regulation and government oversight. It's plain that these people cannot and will not behave or police themselves.