Pubmatic Data Suggests Small Sites Command Higher Rates For Remnant Ads Than Large Sites.

Trackbacks/Pings (Trackback URL)

1. chris (trade2save.com)
   

   April 7th, 2008 at 11:10 am

   What this says to me is that specific/niche sites can have more relevance over ad content to a specific target market. This is reflects what Double Click set out to do. However, the larger sites have their place for major brands.

2. You all crack me up
   

   April 7th, 2008 at 11:40 am

   The real story here is completely hidden: Advertising doesn’t pay very well.

   Let’s say you’re getting $1 eCPM and you’re doing 100 MM page views a month. That’s 100,000 CPMs so you’re generating $100,000 per month. Depending on the kind of site you run, that may not even cover your bandwidth costs.

   Let’s also keep in mind that getting to 100 MM page views per month is a gargantuan task.

   Sorry to all the dumb VCs who invested in ad-supported start-ups. FAIL.

3. EH
   

   April 7th, 2008 at 12:35 pm

   #1: I think what it also says is that niche sites have more precise demographics by definition. A blog about walnut veneers has a higher percentage of veneer purchasers than Home Depot’s site, that’s all.

4. EH
   

   April 7th, 2008 at 12:36 pm

   #2: Neither does working at most startups.

5. Isaac
   

   April 7th, 2008 at 12:39 pm

   Let’s not forget that Pubmatic is a technology provider and is the one “Optimizing” all of these sites. They are not a data analyzation company, they are a Technology provider competing with Rubicon and the likes to win the rights to optimize and take a cut of a Publisher’s inventory.

   Based on this information it appears that Pubmatic may only well for small sites…

6. dmg
   

   April 7th, 2008 at 12:58 pm

   My take is that many small sites won’t bother with remnant advertising unless it’s at a higher cpm, as since they have limited inventory, the revenue will be completely insubstantial. Larger sites will more likely run remnant, at a lower cpm.

   Many small sites will reject a $1.00 cpm, period.

   So out of the small sites that *do* run remnant campaigns, the majority that do run them get a higher cpm compared to larger sites, but probably because most small sites simply decline to run remnants in general.

7. MikeT
   

   April 7th, 2008 at 1:22 pm

   @2: read this:
   http://www.redorbit.com/news/t.....index.html

8. MattW
   

   April 7th, 2008 at 1:49 pm

   @cracked up -

   If you can’t serve 100m page views for a heck of a lot less than $100k, you’re awful. Also, pages frequently come with a heck of a lot more than 1 ad. When they jam 6 ads on a page and average that $1, now they’re taking in $600k. I could serve 100m page views of a typical site for <$15k/mo. There are now two very good networks selling bandwidth for <= $15/Mb/mo.

9. Michael Bakovic
   

   April 7th, 2008 at 2:00 pm

   Of course smaller sites would charge more because usually there ads would be more targeted towards certain individuals.

   Check out my site for ways to make money online http://mikesmoneyclub.blogspot.com

10. justrum
    

    April 7th, 2008 at 2:21 pm

    I operate a small website ~200M pageviews per month, what point would remnant ads serve, ok maybe coffe funds.

    We have been approached to serve branded ads @ $45-55 CPM.
    http://www.cmecorner.com

11. Josh Crandall
    

    April 7th, 2008 at 2:44 pm

    We published a report - Netpop | Response - at the end of 2006 that looked at the effectiveness of ads presented on high-traffic and low-traffic websites. The trend was similar. Another analysis focused on the type of site - shopping, social networking, entertainment, etc.

    More info: http://www.netpopresearch.com/node/1649
    Press release: http://www.media-screen.com/press121206.html

    Relevant content, special interests and small sites are effective.

12. No BS
    

    April 7th, 2008 at 2:52 pm

    @2: You are absolutely right. But don’t expect any love from this bunch. They are already head over heels for the myth of making easy millions from advertising.

    @8: At a $1 eCPM #2 already took into account putting 3 ads per page since according to the data in the post, those 100M sites are only taking $0.31-$0.31 cpm PER ad. Putting 6 ads on a page is a bit excessive. There is only so much space above the fold.

    @10: Thanks for once again proving that the techcrunch comments are just for spamming your own business or website. The words small website and ~200M pageviews a month don’t even work in the same sentence. Typepad.com gets only 124M pageviews a month. Your site DOESN’T.

13. Abdullah Anwar
    

    April 7th, 2008 at 2:55 pm

    Smaller networks do make more money since they are targeted and it makes if difficult for larger networks to target since there are people of all types and therefore the CPM rate on smaller networks is higher than on larger networks.

14. Ben
    

    April 7th, 2008 at 3:10 pm

    I was in a marketing class at the local high school last week (no not for a remedial class in spite of what my marketing team might say) They asked me what the one thing they needed to learn in math for Marketing.

    I told them learn statistics, specifically the difference between causality and coorelation. It would seem that this would apply to this post.

    While I am sure there is a coorelation between small sites and high CPM, I doubt you can show causality.

15. Anand Rajaraman
    

    April 7th, 2008 at 3:40 pm

    This is very interesting data. Coincidentally, I just posted a fairly detailed analysis of social media ad CPMs (http://anand.typepad.com/datawocky/2008/04/affinity-and-he.html). The analysis teased out two factors that seem to explain a lot of the CPM variation (from $16 for TripAdvisor to pennies for Facebook). It’s not size that matters so much as the ability to target; it just so happens that the sites with higher traffic in Pubmatic’s set also tend to be have lower targetability, but that doesn’t have to be true in general.

16. You all crack me up
    

    April 7th, 2008 at 3:59 pm

    @ 11 - Thank you for having a brain. Seems like it’s something most people on TechCrunch might want to look into obtaining. Unfortunately brains are not found on Twitter, SecondLife, Facebook, or other idiotic places the regular TC reader visits.

    There’s some VC that had a blog post about this a while ago. It’s very, very difficult to build an ad-supported business. Too many people are enamoured with being “media” companies.

17. dotw
    

    April 7th, 2008 at 4:04 pm

    I tried pubmatic and saw no evidence of this at all. I save a “small site” by their standards, but saw CPMs was below $1. In fact, I do better chaining networks together myself.

    And their analysis is not that earth shattering. The more page views you have, typically the content is less focused and more general. As are your visitors. This can bring in lower CMPs. Additionally, a large site will always have excess inventory to monetize and sometimes these are monetized at very low cpms (ie adsense). The sheer number of which will drive the Average CPM down. Not rocket science.

18. chris m
    

    April 7th, 2008 at 4:24 pm

    @11: I think @10 was using M as a unit in the same way it’s used in CPM, in other words to mean one thousand - not one million. So in other words his site has 200K monthly uniques.

    @10: $45-55 CPM sounds great - what kind of commitment were they willing to make in terms of number of pageviews, duration of contract, etc.? Your site seems quite targeted.

19. Mike
    

    April 7th, 2008 at 5:59 pm

    I don’t think it’s small vs. large but the depth of the visit. Sites like MySpace that have say 30 pageviews per visit are naturally going to have lower CPMs than sites with 5 page views per visit.

    The deeper the user is in the visit the less likely a user is to click on an ad. Even CPM ads often back into a CPC or a CPA. Advertisers with big networks will pull ads from sites that aren’t performing. This is also the reason a site can do well with a new ad network in the first few weeks and then see a dramatic drop in CPM as poor performing ads (that may have been high paying) are pulled.

    Also, sites with multiple ads per page will often see the individual CPMs on each ad unit decrease as a user can’t perform an action on all of the ads. On the other hand, there may be a balance that nets the site more money - this isn’t to say 1 unit is better than 2. Just don’t expect 5 ads to achieve 5x the original 1.

20. Mike
    

    April 7th, 2008 at 6:04 pm

    Should note that… I don’t think it’s *entirely* depth of visit - but it plays a noteworthy role in influencing CPM.

21. Bob Bowden
    

    April 18th, 2008 at 12:41 pm

    What CPM rate would Wikipedia command? With 3 miln pages and top pageranks, etc?

22. Andreas
    

    April 29th, 2008 at 3:21 am

    Great analysis - does anyone have figures for Germany? Thx

23. Dennis
    

    May 1st, 2008 at 11:12 am

    Thanks for this good data. There’s somethin to think about, and I guess the truth is behind the scenes. Some arguments can be found in these comments here.

24. UtahLuxury.com
    

    May 4th, 2008 at 9:57 pm

    Unless you can catch a ride with Federated Media I think it best to sell ads on your own (like we are getting ready to do). If you want to be serious about business then work up some demographics, hire some sales people and start selling your space. Using these ad networks like we all do is shameful waste of our time.