AJAX changes advertising

Discussions

News: AJAX changes advertising

  1. AJAX changes advertising (24 messages)

    Ajaxian has posted a reference to "AJAX Counting Nightmares," by Eric Picard, which discusses the issues around ad delivery and AJAX. If AJAX refreshes content without an actual page refresh, then a reload - an impression - doesn't happen, and that affects how ad value is calculated. This affects the whole economy of websites using AJAX publicly. Why defend ads? Well, many sites - like TSS itself, for example - derive primary revenue from ad delivery, rather than selling subscriptions. As a result, regular users have full access to the sites, which normally increases the value of effective (and nonintrusive, one hopes) ads. If ad value is reduced through the use of "modern" page coding techniques, then those techniques will harm the sites using them - which means the techniques aren't very useful for external-facing sites. The alternatives are to avoid the techniques, to limit their application, or to define a new impression measurement. Avoiding the technique is certainly easy enough (after all, for most sites, that takes zero effort, since most sites aren't using AJAX yet.) Limiting AJAX' application is also fairly easy, but mitigates the whole point of having it in the first place - for example, a typical site might have 30K of "plumbing" in a given page, consisting of headers, menus, footers, standard static text, followed by 4K of dynamic content. An AJAX-enabled site would be able to replace that 4K of content, reducing bandwidth costs and speeding up the experience for the user by leaps and bounds. While updates to ads are possible when content is updated, the current methods for counting impressions don't work in this case, causing auditing problems. (At least, that's your Humble Editor's understanding. Corrections welcomed.) Defining a new impression mechanism seems to be the most effective decision here. From Mr. Picard's article:
    In general, everyone creating advertising scenarios in these new application types should be conservative and defensible in their approach. It's important from the advertiser's and application end user's standpoints. If advertising doesn't provide a useful impression for the advertiser, it won't perform and advertisers won't buy. We must avoid setting negative expectations, or the whole category will be set back. We also want to avoid clutter and ad overload for the consumer lest we lose all value from the category. We don't want another situation like the one we had with pop-ups and floating ads.
    AJAX is a great tool, one that's finally gaining much-overdue momentum. One can see ads tailored by data gathered during the browsing experience (i.e., hover over a link discussing database products, have an ad delivered that provides database backup service) and thus more effective for a given user, but without a change to the way ad delivery is measured, this can't happen. Ads are understandably frowned upon, in general, but if an ad can be made relevant to the specific user, then they're no longer nuisances but actual useful content. What do you think?

    Threaded Messages (24)

  2. I suppose it depends on how your hits are calculated. Even though there isn't a specific page refresh, an AJAX call will still result in a standard request to the ad server. The AJAX requests show up in the logs just the same as every other request.
  3. What about web stats ?[ Go to top ]

    So what about web stats ? Would that also run into trouble if a section of a page is Ajax enabled ?
  4. In know that the point of the article was on counting hits, but... what if advertisement also evolved and used AJAX to deliver ads. Ad companies could then have control of which and how many impressions they push to the client. I suspect this may already be happening.
  5. Re: AJAX changes advertising[ Go to top ]

    All the more reason that the "click through" model is probably a way better valuation than "impressions". I'd certainly never buy "impressions" these days, not with the ease of adblocking, etc. For example, I haven't seen a single ad on TSS for years, and I don't even see all those "new content" links on the right except when I'm on the wife's computer... - Don
  6. Re: AJAX changes advertising[ Go to top ]

    All the more reason that the "click through" model is probably a way better valuation than "impressions". I'd certainly never buy "impressions" these days, not with the ease of adblocking, etc.

    For example, I haven't seen a single ad on TSS for years, and I don't even see all those "new content" links on the right except when I'm on the wife's computer...

    - Don
    Don, I fully understand. Ads are annoying, and it's hard to imagine cases where they aren't. However, I'd suggest that they are annoying precisely because they're not targeted very well. Personalization helps, but still doesn't tend to rely on data mining techniques that are sufficient to truly target a user, and thus contribute value instead of distraction. An ideal "ad" is one that doesn't really stand out as an ad, but stands out as additional (and sponsored) relevant content. Again, looking back at the original post: consider a news posting on databases, in particular on backing up databases. A "good ad" here would be one that offered up a set of backup solutions, weighted towards database backup/restore solutions. If you were reading over the article - indicating interest - the products would have relevance to you. Further, if there was a link to... oh, let's pick Apache Derby for an example. If a link to Derby's site existed in that news post, the product ads could further look for backup/restore products that supported that particular database. That's a very targeted ad, and while it still needs metadata in the ad mechanism itself (i.e., "this ad content will appeal to...") the burden is much lighter than an individually targeted ad would be (one that relied on data collected specifically about you). Now, looking at the larger picture, it's important to know that those ads that you're skipping out on actually can affect TSS in the long run.. in all honesty, by removing the effect, what you're doing is making it harder for TSS to pay its bills. If it's the golden goose for you, you're not feeding it. Have a heart, willya? :)
  7. Re: AJAX changes advertising[ Go to top ]

    Hi Jo,
    Now, looking at the larger picture, it's important to know that those ads that you're skipping out on actually can affect TSS in the long run.. in all honesty, by removing the effect, what you're doing is making it harder for TSS to pay its bills. If it's the golden goose for you, you're not feeding it. Have a heart, willya? :)
    I sympathize with this, but animated ads is where I draw the line. I am perfectly fine with static ads (I hardly ever notice the ones on my Gmail account, which I stare at hours every day) but as soon as an ad starts showing some animation, I immediately block it with AdBlock. Want more ads clickthrough? Ban Flash ads. -- Cedric
  8. Re: AJAX changes advertising[ Go to top ]

    Hi Jo,

    Now, looking at the larger picture, it's important to know that those ads that you're skipping out on actually can affect TSS in the long run.. in all honesty, by removing the effect, what you're doing is making it harder for TSS to pay its bills. If it's the golden goose for you, you're not feeding it. Have a heart, willya? :)

    I sympathize with this, but animated ads is where I draw the line.

    I am perfectly fine with static ads
    Yes Jo, sorry that my post was so terse. In fact, I do actually swap off adblock quite a bit, so I do see what's going on. Corry Doctorow hit on a good point at the Red Hat Summit last week. He noted that all his books (he does Science Fiction) are available free online. Why? How can someone make a living as an author by giving away the books for free? Simple (to him) - because his number one competitor is obscurity. By making his books free, many more people read them than would otherwise, they talk about it, blog about it, etc. This in turn generates a lot of actual physical book sales for those who want the hard copy for whatever reason (collecting, offline access or simply karma). The parallel here is this - adblocking might prevent you from earning revenue from someone like Cedric, but because of Adblocking Cderic continues to be very active on this site. This attracts lots and lots of people (it seems that a lot of my google searches are something Cedric says on TSS, check your logs :) Note that I pick on Cedric here, because no one has ever logged into TSS to hear me remind yet again that JDO was a, uh, forget it ;). Don't be disenheartened by the changing business model, find ways to exploit it.... - Don
  9. Re: AJAX changes advertising[ Go to top ]

    +1
  10. and what about 'pay per click' model? Dmitry Coldbeans
  11. Re: AJAX changes advertising[ Go to top ]

    Hey - we could use Javascript AOP to send an AJAX request to reload the ad prior to the AJAX request to load the actual content. And we should actually take it a step further and continously reload the ads via AJAX to drive _more_ revenue. It's not - as the author claims - about limiting AJAX, it's about taking AJAX to its limit (and mention it whenever possible). Amen, uhm - AJAX! Think positive :)
  12. Click through is a bad way of measuring ads because it allows for advertisers to implement extremely low cost "branding" campains without fair compensation to the host site. These campains use ads designed _not_ to be clicked but still have the positive effect of getting the company name and image out there. How fair would it be for your ad-supported site to serve up thousands or millions of ads like this? Impressions is an appropriate measure of advertising on the web... If defined as "ad display count". What is inappropriate are the rules around when an ad can change (historically on a full page refresh). Understandably advertisers don't want people creating pages that do nothing but dynamically load ads for the purpose of artificially increasing the impression count (and the site's revenue). However, I think it's completely reasonable to dynamically load a new ad on any _user initiated_ interaction that would have resulted in a page reload when using non-ajax techniques. This would be equally applicable to Flash based sites. I feel most people would agree this policy is fair and equal to the site, advertiser and ad services - everyone involved. The real problem of course is in enforcement. But isn't that a problem that already exists? There are so many ways a site could artificially increase impression count right now that it would be hard to implement any automated system to catch cheaters. So what really needs to change is the contract between host sites and ad services so honest sites can take advantage of AJAX and keep their ad revenue too. Technically, this is an easy change as it only requires modification of ad service's terms of service (contract). Realistically, this will take a long time and a lot of pushback by the community before a change is made. Ad services and advertisers are comfortable with the existing terms of service and are unlikely to change quickly or easily. And they hold the purse strings so if you want to get paid you've gotta play by their rules. Hopefully some big ad service will "get it" and be the first to make the change and the others will fall in line.... Google ad-sense anyone? (I haven't read their TOS/contract, so maybe they are already "get it".) Rob
  13. I think it's completely reasonable to dynamically load a new ad on any _user initiated_ interaction that would have resulted in a page reload when using non-ajax techniques. This would be equally applicable to Flash based sites. I feel most people would agree this policy is fair and equal to the site, advertiser and ad services - everyone involved.
    I disagree with this. At InfoQ.com Enterprise Software Development community we've been building our own ad hosting system and as our site is heavily ajaxified we've had to deal with a lot of these issues and make some interesting decisions. We've ajaxified a lot of stuff that is typically 'wasted' pageviews as far as advertisers are concerned, such as login/logout, message preview, and message reply (no refresh!). When users interact with the system to do these things, they are not looking at ads, they are trying to get somewhere/do something. So those impressions IMO should not be counted against the advertiser as it wastes their money (even though every other site does this). So on InfoQ we do not refresh an add on any of those actions. There was only one particular ajax action that we thought merited refreshing the ads too, that was navigating the tabs on the rightbar (thanks Alexandru Popescu for a great implementation). This is because indeed this would cause a page refresh on most sites, AND users when changing site sections are in browse/read mode, which makes them more likey to look at an add, rather than logging in/posting a message. I think the ajax revoluion is great, it is going to drive down 'wasted' pageviews which is good for the advertiser and forces the publishers to be honest and just focus on great content. Sites hosting ads will definitely need to find ways to position their offerings differently to still do well. Floyd
  14. If any thing this gives you the opportunity to give higher quality impressions, as they're on the more important pages. Nobody says that you have to make a de facto rule that every click reloads the ad pane. But, rather, you do need to emplace the mechanism that any click CAN reload the ad pane, and you simply need to determine whether it's actually appropriate or not.
  15. I think it's completely reasonable to dynamically load a new ad on any _user initiated_ interaction that would have resulted in a page reload when using non-ajax techniques. This would be equally applicable to Flash based sites. I feel most people would agree this policy is fair and equal to the site, advertiser and ad services - everyone involved.


    I disagree with this. At InfoQ.com Enterprise Software Development community we've been building our own ad hosting system and as our site is heavily ajaxified we've had to deal with a lot of these issues and make some interesting decisions.

    We've ajaxified a lot of stuff that is typically 'wasted' pageviews as far as advertisers are concerned, such as login/logout, message preview, and message reply (no refresh!). When users interact with the system to do these things, they are not looking at ads, they are trying to get somewhere/do something. So those impressions IMO should not be counted against the advertiser as it wastes their money (even though every other site does this). So on InfoQ we do not refresh an add on any of those actions.
    Floyd, as Will Hartung already pointed out you've made a business decision to provide a higher quality of impressions to your advertisers. This is a very commendable goal, but a separate issue. As you've already pointed out "regular" sites "waste" advertiser impressions on these get/do something pages already. It would be unfair to impose additional restrictions on AJAX'd sites just because they use a different mechanism to refresh the page. I hope you can understand and agree with this. Although personally, I think it would be a Good Thing (tm) if people followed your lead. Allowing ad changes on any _user initiated_ server request (ajax or standard full page) is a pretty easy line in the sand to "see" and understand and one that doesn't change the rules too much.
  16. However, I think it's completely reasonable to dynamically load a new ad on any _user initiated_ interaction that would have resulted in a page reload when using non-ajax techniques.
    It's even more complicated than this; with Ajax "push" (or "COMET") such as with ICEfaces ads could be updated even while the user is idle, or in response to mouse-over events. Further, the user could interact with an HTML-based ad (assuming it used Ajax) without loading a new page, possibly encouraging people to do so (because it's a more lightweight activity). Ajax opens many new possibilities for advertising, so it's well worth updating the metrics to account for it.
  17. Ted, I agree with what you've said regarding the possibility of "pushing" ads using ICEfaces or reverse-ajax, as it's being called these days. But I think you may have missed a few key parts of my statement.
    However, I think it's completely reasonable to dynamically load a new ad on any _user initiated_ interaction that would have resulted in a page reload when using non-ajax techniques.
    • user initiated - I wouldn't call push user initiated
    • interaction - So it doesn't happen unless the user does something. Yes, this could be mouse-overs but then their is part 3 of my statement
    • that would have resulted in a page reload when using non-ajax techniques.
    I think in aggregate those clauses fairly describe what most people would accept as a user request to load a "new page". Just because the "new page" is partially new content and partially old content and loaded by a new means called ajax is not as important... to some extent it's still a "new page" pulled from the server and should be ok to display a new ad. Now, can I point out ways that people could beat the system.. sure. I could create a page that contains banner ads from 5 different ad services, and minimizes those ads so visitors don't even see they are on the page (and thus keep coming to my site but aren't bothered by ads) and reloads a new ad every 30-240 seconds or so with a random delay to it so it's not as obvious that it's being done via javascript instead of the user loading a new page. Would an ad service TOS/contract prevent me from doing this? NO. Would it take a little analysis and time for the ad services to catch this abuse. YES. Would it take them long to find the abuse? Probably not. Would doing this even require AJAX? NO. I'm just doing image src updates which have been around for ever and don't use XmlHttpRequest (the core of this "new" stuff everyone is calling AJAX). My point is that listing ways in which AJAX, etc could be used to abuse ad services isn't really helpful. I can abuse the system now using standard JavaScript w/o XmlHttpRequest or reloading the page. Even the existing TOS/Contract wouldn't prevent a skilled person from creating a page to cheat the system and get away with it for a little while. The only way to prevent abuse is to monitor to catch the obvious abusers and use a more manual process to catch the crafty ones. The problem is most existing ad service's TOS/contracts unfairly limit site owners who want to refresh ads in a legitimate and fair way because they want to AJAX their site and by changing the TOS/contracts legitimate users would be less hindered. I think my statement clearly explains the legitimate times that a site should be allowed to change an ad and is sufficient, except perhaps to those who are looking to abuse things. But those people will abuse the system anyway.
  18. Ted, I agree with what you've said regarding the possibility of "pushing" ads using ICEfaces or reverse-ajax, as it's being called these days. But I think you may have missed a few key parts of my statement.
    However, I think it's completely reasonable to dynamically load a new ad on any _user initiated_ interaction that would have resulted in a page reload when using non-ajax techniques.
    • user initiated - I wouldn't call push user initiated
    • interaction - So it doesn't happen unless the user does something. Yes, this could be mouse-overs but then their is part 3 of my statement
    • that would have resulted in a page reload when using non-ajax techniques.
    I think in aggregate those clauses fairly describe what most people would accept as a user request to load a "new page".
    Absolutely; you have accurately described what we need to extend current advertising terms of service to Ajax applications. Each mouse click is a legitimate opportunity for updating the ads whether or not the entire page is refreshed. On a technical note, this requires Ajax that is integrated across the entire page (again a capability that is emphasized by ICEfaces). When the user interacts with an Ajax component, the ads should update, but the page developer should not be required to keep track of which user events update which ads. It should simply be a matter of including an "ad" component in the page, and the ad component should update itself on each user interaction. (This is easy with ICEfaces because the component tree of the entire page is processed and only the minimal update is applied to the browser via Ajax; when the user event is processed, the ad component gets a chance to run just as the application components do.)
    My point is that listing ways in which AJAX, etc could be used to abuse ad services isn't really helpful. I can abuse the system now using standard JavaScript w/o XmlHttpRequest or reloading the page.
    If anything, it's harder to use Ajax to abuse advertising services because XMLHttpRequest connections are limited to the originating server.
    I think my statement clearly explains the legitimate times that a site should be allowed to change an ad and is sufficient, except perhaps to those who are looking to abuse things. But those people will abuse the system anyway.
    It's certainly reasonable to use Ajax to change ads when ads would have been changed with a traditional web application. However, is this all that is legitimate? What about pushing down new ads not to abuse the advertising system (it might be abusing the user), but just to show the user a new ad when they've been on the same page for a long time? A sequence of ads could be designed to build on each other, or could be based on the scrollbar position. There are many new possibilities opened up by application-initiated advertising. Moreover, this allows page design to not require user events to proceed -- web articles need not be broken into "pages", for example.
  19. Further, the user could interact with an HTML-based ad (assuming it used Ajax) without loading a new page, possibly encouraging people to do so (because it's a more lightweight activity). Ajax opens many new possibilities for advertising, so it's well worth updating the metrics to account for it.
    Another reason to turn Javascript off.
    What about pushing down new ads not to abuse the advertising system (it might be abusing the user), but just to show the user a new ad when they've been on the same page for a long time?
    Whenever I encounter a self-reloading ad, I leave the website and never come again. Also, there are parts of the world where people still pay for the traffic. For them, ads are like advertising calls or text messages sent to their cell phones, not just annoying but costly too.
  20. Whenever I encounter a self-reloading ad, I leave the website and never come again.
    Do you do leave the site because the self-reloading ad affects the performance of your browser or because you don't want to see the ad? The worst case is when the entire page is self-reloading to refresh the ads, and you lose your place in the article because of it. If the ad updated itself without affecting your reading experience, would that warrant leaving the site?
  21. Click through is a bad way of measuring ads because it allows for advertisers to implement extremely low cost "branding" campains without fair compensation to the host site.
    Click thru' is a bad way for the Web site owners that want to maximize revenue if they have volume but not quality. But click thru' is almost always much better for the paying customer (i.e. the advertiser). One of the hardest things in marketing is to measure results. Paying per click rather than paying for 'banding' advertising is a much better use of a budget. There's many ways of preventing advertisers from producing ads that are not-for-clicking and just for free branding (the most common is minimum click-thru rates). PJ Murray, CodeFutures Software Data Access Objects and Service Data Objects
  22. Click thru' is a bad way for the Web site owners that want to maximize revenue if they have volume but not quality. But click thru' is almost always much better for the paying customer (i.e. the advertiser). One of the hardest things in marketing is to measure results. Paying per click rather than paying for 'banding' advertising is a much better use of a budget.
    Click through has always been unfairly biased towards the advertiser and the ad-service companys. Of course it's a better use of the advertisers budget to display 1 million ads and only pay for 10,000! All advertising in the physical world is impression based, why shouldn't online advertising be the same? Does a newspaper or TV ad result in an immediate click to the advertisers website? No and those advertising mediums are much more expensive. Do you think any of the major TV advertisers and newspapers would agree to a minimum ad response rate before being paid? Heck no. Try going to a billboard advertising company and telling them you want to put up an ad on the highway but you will only pay them if at least 10,000 people buy something from you and directly mention the billboard when they buy. They'll quickly show you the door. Online banner ads are the billboards of the internet. Any ad has a residual branding effect, including banner ads. Banner ads may even be more successful than people initially give them credit for. They may have a subliminal effect simply _because_ you don't take notice of them anymore. Your purchasing habits might be altered because subconciously your mind remembered all those banner ads while surfing and you bought the new hot gadget at BestBuy as a result. Should the sites that showed those ads not be compensated because you didn't decide to visit the advertiser's website at that moment? The minimum click through you speak of is a guarantee between the ad service and the advertiser, not the ad service and the website. So if the advertiser creates a lousy banner ad that no one clicks on or is selling a product no one really wants their click through will be very low. The ad service guaranteed them a minimum of 10,000 clicks so the ad stays in rotation until it hits the minimum. This means my website is spending some time displaying bad banner ads instead of displaying other ads that could generate revenue for my site. I, the website owner, am being cheated out of fair compensation for displaying these ads just because the advertiser has a bad ad or a lousy product! Everything is heavly biased to the advertiser and ad service and not the website owner. Unfortunately, this is the sad state of affairs with online advertising and unlikely to change in the near or long term future.

  23. Banner ads may even be more successful than people initially give them credit for. They may have a subliminal effect simply _because_ you don't take notice of them anymore.
    Sure, this is true. But you can't measure branding. But that very minor compared with the (rare) opportunity for a marketing person to measure the real ROI on an advertising campaign or to link a product sale directly to an advertising campaign. That's a huge issue for marketing departments, so it makes a lot of sense to use CPC and to have measurable results. The real issue is a fair price for each click thru and sites having the right to refuse adversting that has a poor click thru rate.

    The minimum click through you speak of is a guarantee between the ad service and the advertiser, not the ad service and the website.
    Actually, I think we're talking about different advertising services. The one I'm referring to is stacked in favor of the Web site because an ad is not displayed unless it has a minumum click-thru rate. Which service are you talking about? Sounds good to me (as an advertiser that spends thousands each month on Web ads).
    I, the website owner, am being cheated out of fair compensation for displaying these ads just because the advertiser has a bad ad or a lousy product!
    I agree with you there. You should not have to pollute your site with poor quality ads. But you always have the option of refusing to display an ad.
    Everything is heavly biased to the advertiser and ad service and not the website owner.
    Which is exactly how it should be. They are the paying customers, after all.

    Unfortunately, this is the sad state of affairs with online advertising and unlikely to change in the near or long term future.
    Actually, one interesting development is that leader in CPC - Google - has recently introduced CPM advertising for some selected sites. Using this system, I've seen my cost-per-click increase by a factor 20 on most of the sites that I use Google's CPM. So I'm thinking of sticking to my old rule of only using CPC advertising. PJ Murray, CodeFutures Software Data Access Objects and Service Data Objects
  24. The solution is Presence[ Go to top ]

    It is true that AJAX client applications have changed online advertising by making the page impression meaningless. It is obviously not sufficient to leave it up to individual websites to decide what constitutes a page impression or how important each click is. I believe that the only fair way to charge for advertising in any client side application is by using presence. Just like Instant messaging client detect when a user is inactive, away, or offline, a clever presence enabled advertising server could detect if an ad is on a page that is in the foreground and if so if the user is active or if they have gone out for lunch. Advertisers could then be charged for number of unique users and the average amount of time that the ad has been displayed for. No system is ever going to be able to perfectly detect how much attention an ad recieves but I think that this system comes about as close as is possible.
  25. Re: The solution is Presence[ Go to top ]

    Back In The Day, that high flying 6 week period when "Push" content was "the thing", we had a push client, and it did precisely this. It essentially "knew" when someone was at the machine, rather than the ad displays simply running. We could record "active" impressions vs "passive" impressions. We simply tracked mouse movements and what not, and then gave it a reasonable time out (30 secs, 1 minute, whatever) for the user to go idle. I don't know how much activity can be tracked with a browser (I assume all of it). But the real issue at this juncture is communicating to the advertisers the capabilities of this system, and being able to charge appropriately for these impressions. An "active" impression should cost more than a "passive" one. And, of course, you need to have relationships with the advertising providers. Finally, as an aside, my biggest complaint about ads is simply performance. For my main browser, I don't even have flash installed. Send me a flash link and I'll open up IE. And all of my GIFs are set to animate once. (Message being dump flash ads, and put your message on the last frame of the GIF.) But I loathe sitting there staring at an empty page, little spinner thing in the corner spinning while the status is stuck at some ad provider taking their sweet time to pony up their graphic. This is teh suk! At least with Ajax ads, they shouldn't delay the page from loading while they fetch their eye candy in the background.