More evidence that ad CPMs are higher in the Tail

Following up on my post about how niche social networks can command higher  ad CPMs than the big generic sites such Facebook and MySpace, here's some new data from PubMatic that makes a similar point about web publishing:

Not only do small (Long Tail) publishers montetize their content at 3-5 times the rate of the larger publishers in PubMatic's survey, but they're improving in the current environment while the big publisher decline.

Segment Definitions

• Small Web site segment: Less than 1 million page views per month.
• Medium Web site segment: Between 1 million and 100 million page views per month.
• Large Web site segment: Over 100 million page views per month.

Comments

[Paul Dettman]

Hi Chris, You had me the first time on this. I know as an economist you have to search for the evidence, but I think this feels intuitively true. If you are reading a niche publication, you are far more likely to respond to advertising targeted at your niche. If the same person reads a large circulation site/TV show/newspaper then they would expect less targeted ads. Because the broad ads have a huge audience, they would expect to pay less 'per impression' which makes no sense in the old media, but is still intuitive. It may cost thousands for a full page in the WSJ, but then many hundreds of thousands of eyes will see it.

The lesson here is that if you run a small site, you can charge a higher CPM but you will (almost by definition) get less clicks in total. There is room in the market for both: MySpace will attract Coca-Cola and MacDonald's, and my niche politics site may attract (for example) an advert for a short-run memoir. The publisher of the memoir will see that MySpace's CPM is lower, but the quality of those clicks (for them) will be poor.

Thanks for making me think about this!

[Rajeev Goel]

Chris,
I agree with your post on how niche social networks can command higher advertising rates than larger, uber-networking sites as your own data shows. I think there are two phenomenon in play here:

1) The general purpose social networking sites (Facebook, MySpace, etc.) have a much broader and therefore untargeted audience then niche social networks which are organized around some galvanizing thread (age, gender, interest, or other demographic attribute). As we are in the early stages of experimenting with social media monetization on the web, it will be some time before the general purpose sites can deliver targeted performance like a targeted social networking site can.

2) I suspect that user sessions on general purpose social networking sites are longer than most non-social networking user sessions and even than targeted social networks. These shorter user sessions leads to a higher probably of click-through or other ad performance, resulting in higher monetization rates.

I'm sure there are more causes for this as well.

- Rajeev

[Eric]

Niches naturally provide a greater fit, no doubt, than large mega-sites that work better for general consumer products.

I'm also glad that this trend exists - as of now.

But don't Google and Yahoo ads already target narrow word clusters and keywords? Why is this sort of advertising - which I have often personally found ineffective - so ineffective now? How do you expect Google, Yahoo, Myspace, etc to respond?

Consider me curious.

[Rajeev Goel]

Eric, I think you have hit the point exactly - Google, Yahoo, etc. are best at doing contextual targeting. The challenge is that contextual targeting may not be working well in social media. I suspect that use of explicit and implicit demographic data will ultimately be best, perhaps combined with contextual and other targeting techniques, but Google, Yahoo, Fox, etc. are still learning how to do this. Just my sense - i have no data to back this up.

- Rajeev

[Odysseas Ntotsikas]

Hi Chris,

I don't see this as a matter of Short vs Long Tail. It looks more like a matter of niche highly targeted sites vs the mainstream generic portals. It is normal for highly targeted sites to get a premium because a) they usually run very contextually related ads which result in good CTRs b) their supply is limited so in relation to (a) and in order to be profitable they need to sell at a premium price.

Odysseas,

[Marshall]

If the long tail has as much or more audience than the big sites, then why isn't the average close to the long tail ? The data for February imply that the Large and Medium web sites, together, have 8.6 times the weight of the small ones, and the data for all three months imply that the large web sites have 61% of the weight, the medium have 24% and the long tail only 14.6%.

So, what was the weighting used to produce these numbers ? A total cumulative audience of 61% for the largest websites seems too large to me

[Marcelo Barcia]

Dear Chris, " The Long Tail" is great sucess in Brazil. Congratulations!

Barcia

[Kiran Pillai]

Its more than natural that 'small' websites would attract more relevant audience than a generic major like Facebook or Myspace, as users who are specific about the information they seek would be attracted to a niche source than a whole storehouse of vague information.

Are we seeing a trend of Social Media Monitoring overtaking CPM / PPC campaigns here?

[kobi]

it's really cool statistics.good work

[NoDomainsForSale]

Ad revenue figures are becoming less important in valuing online properties, says Rick Latona in a May 26, 2008 article. Revenue may make a property sellable, he argues, but with Google continuing to cut payouts for click throughs, valuation factors will shift away from revenue rates.

Yes, long-tail sites can enjoy better CTRs and twice the sales conversion rates of search advertising, but as Google takes larger and larger shares of the pie, then many publishers and site owners will find themselves at a crossroads.

See:
http://www.ricklatona.com/
2008/05/26/what-makes-a-domain-sellable-part-4/

[Andreas Ahlenius]

Hi

I think this definitely applies for general and niche sites from all over the world. However when Facebook released their new platform for advertising last year they kind of changed all this. They made a large untargeted web site into different smaller web sites or market places If you like, when they made it possible to target ads by age, sex, location, relationship status and interests. I think this is the future of advertising, making short tail into long tail.

[david cushman]

The clues to this are in Eric Schmidt's assertion that the mobile internet will make more ad revenue than the fixed line web within a few short years. That (I'm sure) is not because he expects it to carry more interruptive banner ads and links to click - it is because he expects a new more personal form of messaging focused around cost per action to emerge to make the most of the new medium.
Fixed line internet metrics are hung up on broadcast models (cpm)- these clearly don't fit how social networks (which are few-to-few networks) function.
Even the biggest social network is actually a cluster of long tail self-formed groups - so you need a new form of 'advertising' to join in the creation of value in these networked models.
Read my white paper (marketing and advertising solutions for a networked world) for a few suggestions.
It's on my blog: http://fasterfuture.blogspot.com

[val]

Chris,
While the gist of your argument makes intuitive sense I feel that there are several effects here that are overlapping and you cant really understand the underlying trends without disentangling them.

Effect 1: The main issue is that as you go down the long tail the following metric frequently decreases: CPM per unit of effort to obtain that CPM. So in this case we can see the unit of effort as related to the number of hours a blog poster works on his blog per day or week or other time period. So while it is true that CPM might increase, the value of the blog per unit of work decreases.

Effect 2: As we go down the long tail the tendency is to go from general topic blogs and sites to more niche sites. CPM is of course higher on niche sites than on general topic sites. But the increase of CPM from niche effects is different from Effect 1 above, especially because we can have long tail blogs that are not necessarily niche based in such a way that advertisers can take direct advantage of their niche value.

While it is true that niche content creators in all mediums are putting increasingly more effort into the content creation process to keep the interest of their audiences, and that this increase in effort/cost is counteracted by the decrease in technological costs and efficiencies in content creation from new social and non-social technologies, the increased "value" from the long tail may only be for the advertisers (within inventory limits) and not for the content creators. Of course many content creators do it "for the love" which is absolutely fine, but in many cases the drop-off in #impressions/hour worked may be much more than the 3-5x increase in CPM rate.

Im not saying all this cant change, just that we need to tease apart all the underlying factors to understand the nature of the effects.

I have an interesting alternative method to monetize the long tail- drop me a line if you would like to hear it.

[manele noi]

manele noi

[manele noi]

www.top-manele.org manele noi

[manele noi]

www.top-manele.org manele noi

[manele noi]

manele noi

[Laura]

An apples to apples comparison might be looking at targeted ad revenues across all types of sites (sm, med, lg).

Furthermore, the report begs the question: does a "small" site scale into a meaningful business? Regardless of whether the CPM's garner a premium, is it even profitable?

Finally, the study notes the following: "The pricing data reflects net publisher monetization via ad networks and excludes ad networks' share of ad spends as well as inventory sold directly by publishers to ad agencies or advertisers."

This is self-serving research since a smaller site wouldn't necessarily have the resources or expertise available to sell directly to ad agencies or clients. Hence, it is an advertisement for buying the PubMatic service (come with us and we'll monetize your efforts). Most large publishers have a direct sales force and pride themselves on not filling the site with "ad network" inventory. Salespeople are valued by the direct relationships and business that they generate with clients.

So it seems to me that a substantial amount of the ad revenues for the big sites must be left out of this report.

[Dave Allen]

Chris,

I've been following this discussion closely as I am uncertain of the future of, as you say "the one-size fits all" mentality of the large social networks especially when it comes to marketing to those immense audiences. After being a panelist twice in two weeks where these topics were discussed, one in Portland, Or, called LISA08 (Lessons in Social Advertising) a "push" discussion, and another at the Grammy organizations' Hawaii MusicTech in Honolulu where we discussed with musicians the idea of having their fans be part of the "pull," I got to thinking.....

I thought I'd share a post I wrote about the subject of Social Media and Advertising as it's a precursor to a white paper and also includes information and quotes I've pulled from your Long Tail blog posts.

http://www.social-cache.com/2008/06/on-social-media-blogs-and-advertising

[seo]

But don't Google and Yahoo ads already target narrow word clusters and keywords? Why is this sort of advertising - which I have often personally found ineffective - so ineffective now? How do you expect Google, Yahoo, Myspace, etc to respond?

[kabin]

thanks...
Kabin
Konteyner