November 6, 2006

Turn to Launch Hybrid CPA Ad Network

Marshall Kirkpatrick

30 comments »

San Mateo startup Turn will unveil an interesting new advertising platform tomorrow at the Web 2.0 conference. It combines Cost Per Action bidding with a wide range of user, site and ad performance analytics. The company calls itself the world’s first automatic targeting, bidded CPA ad network.

I often write that new ad networks bore me as a prefix to profiling an interesting one, but lately I’ve been seeing enough interesting new approaches to online advertising that I’m going to confess a newfound interest in the field. To say there is room for innovation in advertising would be a huge understatement.

The Turn network is taking an approach that mixes cost-per-action (CPA) bidding and analysis of 60 contextual, behavioral, demographic and performance factors in determining which ads to serve and when. The company intends to do all the ad placement heavy lifting in the background so that buying ads is made supremely simple and yet unusually effective. That’s what they intend anyway. It reminds me of RightMedia’s dynamic serving of whatever subnetwork’s text ads will pay the highest at a given moment and BrightRoll’s selection of video ad formats based on ongoing conversion performance.

Turn was founded by Jim Barnett and John Ellis, formerly CEO and CTO of Alta Vista. As the company launches it has 5 million ads in its inventory from 1 thousand advertisers, run by 30 publishers with 11 million uniques visitors per month. The company has taken $18 million in venture funding from Norwest Venture Partners, Trident Capital and Shasta Ventures.

Buying ads in the Turn Network doesn’t require keyword selection and management. Instead, buyers identify actions they want their audience members to take and how much they would be willing to pay per time those actions are taken. That action might be a site visit, email sign-up or completed transaction.

When visitors come to a site, data about those users, contextual analysis of the site, of the ads and of every ad permutation’s success in that and related sites are all considered in determining each ad’s probability of success. That probability of conversion is then considered relative to the price being paid on a CPA bases. The CPA bid divided by the probability of the action being takes equals an ad’s effective revenue per thousand impressions. And thus an ad is served!

Ad networks must all look shiny, new and promising when they launch. Only time will tell how effective this network’s execution will prove to be. It’s time someone made a mixed CPA and analytics based approach available.

  • Sphere It

Trackbacks/Pings (Trackback URL)

  1. Adobe Activescript Open Source, Turn (beta) « Technically Speaking
  2. Turn.com Hopes to Challenge Google’s AdSense | Marketing Pilgrim
  3. Cost Per News
  4. TechCrunch Japanese アーカイブ » Turn、ハイブリッドCPA広告ネットの発表へ
  5. Global Nerdy
  6. Turn Launches Hybrid CPA Ad Network » Dee’s-Planet! Blog
  7. Venture Beat Contributors » Yahoo just needs to fix one thing: Monetization
  8. fuery.com
  9. Google'ın PPA reklam ürünü | 6 Patlar Teknoloji
  10. L’implantation du Coût par acquisition au Canada: rêve ou réalité? « Katheline Jean-Pierre, Stratégie Web et développement des communautés.

Comments

RSS feed for comments on this post.

  1. Jeff O'Hara

    Ad networks are great and all, but i think a lot of them are going to have problems filling ads for their publishers. They are all going to have a problem taking share away from adsense.

    Just my 2 cents.

    -Jeff O’Hara
    http://blog.zemote.com

  2. wyatt

    First of all CPA is not advertising, they are two different things right? CPA takes creativity out of the most creative industry in the world. There is big value, in buzz, branding, and awareness.

    It is funny to hear companies “gunning for Google”. CPA is a switch Google has hinted at turning on, and I imagine their distribution would be much easier.

    I visit TechCrunch everyday, and I can say that I have never signed up for a service of any of the advertisers on TechCrunch. But, do I know who they all are? Yes. Do I think they are legitimate? Yes Would I use their services if I needed them? Yes.

    That has value!

    Imagine this….What if all the junk direct mail (snail mail) you got in your mailbox told you to mail it back, if you didn’t like the offer, and weren’t going to buy anything. Wouldn’t that be absurd? That is what networks like Turn and CJ are asking the their publishers to do with their audiences. I think there is value in these networks, but they are certainly not a first choice or an end all be all.

  3. T.J. Crowder

    To wyatt: Very good point about CPM having value. And even absent the consumer ultimately saying “Hmmm, I need a widget — I’m always seeing Widgets’R'Us on TechCrunch, I’ll go check them out” there’s the old (and probably apocryphal) stat that it takes hundreds or even thousands of impressions before someone clicks. If 90% of those were on Site A, but the user ultimately clicks the ad on Site B, CPM rewards the site doing the bulk of the work where as CPA just rewards the site where the final click occurred. And yet, advertisers need some way of knowing where their ads are being effective and where they’re not. So we see a rise in CPA models. Hopefully we’ll see more blended models as our understanding of this medium improves. Turn’s pages for publishers suggest that they do blend their model a bit, combining eCPM with their CPA stuff; it’ll be interesting to see how it works out…

    Separate from the CPM vs. CPA thing, though, there are sites for which contextual analysis doesn’t work well, particularly those requiring login or which feature largely non-textual content. I run a site for playing board games which has both of these problems — we need you to login (because, after all, we need to know which games to show you) and board games are largely graphical rather than textual. We don’t feel we’re getting as much out of AdSense as we’d like as a result and we’re looking to move to services which let us describe the site more fully to potential advertisers rather than relying so much on the analysis of the site by a third-party (e.g., Google or Turn). So for some sites, the old-fashioned “give us keywords, we’ll give you ads” is still a useful structure.

    – T.J.

  4. Gabe

    What would really help is:

    1. An few examples of sites running their ads. Are there any?

    2. A concise, clear, and specific description of the key difference(s) vs. AdSense. Really short.

  5. George

    Well, AdSense has lately been having a problem with not having enough publishers to place all of its ads, so might be a good market for a more effective advertising solution. Best of luck to Turn - I’d love to see a few more advertising platforms on the market (since Panama has been delayed so many times, it’s not even funny anymore).

  6. Jawad (Shuzak)

    Google has a vast distribution channel that others lack. It is simply futile for a startup to be gunning for a corporate giant with a treasure chest in billions.

    “What all great companies have in common is that they have nothing in common…the leader is the leader because he did something remarkable. And that remarkable thing is now taken - it is no longer remarkable when you do it.” - Seth Godin, All Marketers are Liars

    As a startup, when you compete with a giant, you are almost always walking in its shadow. The problem with catching up with competitors is that you are always catching up.

    Jawad (Shuzak)

  7. Jeff

    There’s a lot of problems with their business model that aren’t readily apparent to people who aren’t in the CPA industry already:

    1. The great majority of successful CPA traffic is driven from emailers and affiliates who bid on paid keywords, not from websites that load banners or text links (those only drive about 10% of any network’s traffic).

    2. There seems to be very little concern about whether the advertiser is actually tracking the action. Obviously the algorithm they are using will not display ads that aren’t converting, but if I was an advertiser it would certainly be tempting to sign up and not worry about placing the tracking code on my site to see how much free traffic (not to mention a impressions) I could get. Successful CPA networks monitor this thing actively and get feedback from affiliates who test out the system.

    3. CPA works when advertisers and publishers have a close relationship and work together to tailor ads and timing. Turn’s system doesn’t allow for any close relationship at all with the publisher. It’s as if Turn decided that pay per click and CPA could work the same way; but PPC is fine when the publisher is essentially anonymous, not so with CPA.

    4. Overall, I think this model is built on faulty logic, again because it was built as though it were a PPC or CPM network. CPM advertising puts all the risk on the advertiser (who pay for impressions in hopes that customers click and convert); CPA advertising puts the risk on the publisher (who gives up inventory in hopes that someone clicks and converts). Publishers are usually only willing to take CPA if a) they aren’t established enough to demand CPM, or b) they know that they can use their own marketing skills to create ads that are relevant to their traffic, or at the very least choose the most appropriate ads. In Turn’s system, the publisher is giving up all control of the ad (as they would in CPM or PPC) but still has to accept the risk of CPA.

    Commission Junction, one the world’s largest CPA networks, recently told its publishers it was going to change the ad distribution to a system that prevented them from controlling the creative. The publishers, who understood their profits would suffer if they couldn’t control the ads, revolted and forced CJ to reverse its decision. Turn is trying the same system, but also removing the ability for publishers to use email or PPC arbitrage to drive the traffic.

    Overall, the model just doesn’t make sense. CPA works because the publisher has skills as a marketer and therefore accepts the risk of CPA. Behavioral and contextual targeting is great, but its only necessary if the publisher isn’t willing to spend a few minutes to choose (or create) their creative. Their technology may have some application as part of a larger CPA network, to target newbie publishers who don’t know how to drive converting traffic, but as a stand alone offering it seems doomed to fail.

  8. alan patrick

    Very interesting comments - very perceptive. As to Turn, its going to be hard to differentiate themselves but a hybrid plus great analytics seems to me like the best approach..

  9. Matt

    These are the founders of AltaVista. These are not a couple of no-name startup entrepreneurs trying to develop an ad network. It’s easy to lump Turn into a category with lots of others…but, that would fail to recognize this important difference. I expect these guys will be very successful.

  10. Patricia

    @ Jeff, thank you for your perspective. I read this article this morning and was like, “wha?”. I am going to re-read your comments later to fully take them in. Thanks for posting them!

    Can anybody explain how this affects or doesn’t affect interactive advertising? I guess I’m having a hard time finding where all of this falls into place compared to what’s going on now.

  11. Basicity

    Wow, what a great four-letter domain name for a company! Wonder what they paid for that domain.

  12. asdfsd

    asdfasdf

  13. Jeff

    @ Matt — I think you’re falling into the trap of believing that just because there are big names attached to a project that it will succeed. Google publicly announced its own CPA offering a few months back, and some fairly respected journalists predicted they would destroy existing CPA networks. Google has since backed off that model and no one talks about it anymore; the reason it didn’t make sense to me then was for some of the same reasons I mentioned above — PPC and CPA are completely different beasts.

    I have no doubt that Turn will draw some advertisers, but just look at this system from a publisher’s perspective — assuming you can’t attract CPM advertising (which all publishers want first), then you have the choice of reliable AdSense that pays for clicks, or existing CPA networks that allow you to create your own links and ads and have a trusted tracking system and close relationships with your merchant partners, or you can decide to try Turn which gives you no flexibility, only pays you on conversions, doesn’t allow relationships with the advertisers, and only offers the promise of some new technology that targets ads better. Why would they switch? Certainly not because the company was started by former Alta Vista execs.

  14. waldo

    Just for clarification, these are not founders of altavista. They are the dudes CMGI hired to stem losses and package it for sale to Overture. May their luck hold on this one.

  15. Sherwood

    “Buying ads in the Turn Network doesn’t require keyword selection and management. Instead, buyers identify actions they want their audience members to take and how much they would be willing to pay”

    — Umm, keyword selection is the process of selecting your audience. So these ads are basically being sent into the network, sucked through the straw wherever “demand” exists.

    Sounds like a recipe for click fraud - or in this case, action fraud.

  16. David Mackey

    Pretty innovative idea. All that processing overhead though - how long will it take ads to load?

  17. Stuart

    Not ANOTHER ad network, I’m forgetting all my logins :P This does sound interesting though, will give it a go.

    Stuart
    http://www.earnersblog.com

  18. Matthew

    CPA has been around for a little while now. I read a nice overview on mercury news (http://forums.mercurynews.com/n/mb/message.asp?webtag=kr-mercurybiz&msg=175.1&ctx=0). But Turn does seem innovative despite this. I have not seen anyone else offer blended targeting of both Graphical and Text ads. Also just fyi, I signed up as an advertiser last week and loaded a campaign (with about 25 ads) into the turn network. I have the same ads running through a different network as well (not CPA) and am glad to say that Turn’s placement of my ads was generally far more accurate (Yes you folks at Turn - Although you won’t tell me where you stick my ads, I have my ways of finding out :-) - Matt

  19. Chris

    Patricia (stylediary.net) wrote:

    > Can anybody explain how this affects or doesn’t affect interactive advertising

    As a relative latecomer to this thread, I just saw Patricia’s comment above, and Matthew’s post (previous) pointing to an overview article, which I promptly perused. Patricia — I found this article to be an excellent description of how this affects the world of interactive advertising. It’s well worth a read. I recommend it. Thanks to Matthew.

    PS. The article is still live, I believe, as of this posting. It’s at:

    http://forums.mercurynews.com/.....;msg=175.1

    They do take it offline at some point, right?

    Chris

  20. Adrian

    Jeff, are you sure you don’t have a vested interest in the current CPA network model? Some of your points don’t quite hold water.

    - CPA networks don’t allow relationships with advertisers
    - CPA networks only pay on conversions
    - The publisher won’t care if tracking doesn’t work properly. ECPM will drop, therefore another offer will take its place

    There is more.