louisrosenfeld.com logotype

Home > Bloug Archive

Feb 07, 2002: The Value of "Pay-For-Performance Search"

In the last Bloug entry, my pal Scott Brylow brought up an interesting point about overture.com, the site formerly known as goto.com.


Overture is up front about its business model, namely, charging clients for placement in search results. In a nutshell, the client that bids the highest gets listed most prominently.


My background is in information science and librarianship, so naturally such blatantly commercial approaches make me break out in hives. Do users really benefit when their queries retrieve the sites with the greatest financial backing, rather than the ones that are most relevant?


But Scott claims it's working, and that overture.com has been able to attract investment in these post-bubble days. (I typically believe whatever Scott tells me.) So what's behind overture.com's success?


If any of you are overture.com fans, please tell me why (and I ask purely out of respectful curiosity, not because I think you're a nutcase). Even if you're not an overture.com user, post your favorite explanation here.


Oh, and I do have my own theory. Overture.com is still a small directory. Naturally this means results sets are smaller, and perhaps users get the impression that small result sets mean "filtered" results. In overture.com's case, such "filtering" has nothing to do with quality, but users may have the impression that fewer results mean quality.


If I'm right, then overture.com's business model won't scale if it continues to attract clients--users will eventually retrieve too many results and the sense of culling, artificial as it may be, will vanish.


But I'm probably completely wrong. Still, I'd bet that overture.com's current success has more to do with its IA and UI than the company's financial backers might realize, or be comfortable with...

email this entry

Comment: skebrown (Feb 8, 2002)

Hi Lou -
I manage my companies keyword purchases via Overture and I am not sure if I agree with the practice --but it works in producing quality user sessions! Also, Overture has strict listing guidelines and I would say in most cases, the user benefits.

Overture, with pressure from its affiliates, must ensure keyword accuracy. Each keyword used in Overture must have relevancy to the URL that is associated with it. They have humans reviewing to ensure this --as I have found out while trying to alter one of our keyword URL's.

So, as long as Overture is reviewing with strict eyes, they will produce relevant results. I think this is why they are successful.

ske

Comment: Avi Rappoport (Feb 8, 2002)

Overture seems to be better at shopping search -- I was just upgrading RAM on my machine and Overture's results were significantly better than Google's. That's only one data point, but I've heard this from other people as well.

I like the idea of different emphases for different engines.

Comment: Rich Wiggins (Feb 8, 2002)

I have always sung high praises of Goto.com. Their CEO had a great quote a few years back:

"We were looking for a way to organize our directory that scaled. Nothing we thought of scaled. But the market scales infinitely."

Paraphrased, but fairly close. I can look up the exact quote in one of my past presentations.

I think it's beautiful. They rank based on dollars paid and they are totally up front about that fact. So what's to quarrel with?

It is no different than the Yellow Pages having full page ads from roofers and lawyers and hospitals that are willing to pay for position (and space). Or ads during the Super Bowl. Or on the walls of Nascar tracks and minor league ball parks.

Am I the only person who ever picked a house painter because they had a bigger ad in the yellow pages, assuming they had more money to spend on ads because they did more business?

The moral issue comes in when a search engine blends positional selling with editorial copy, so the hapless user doesn't know that $$$ were part of the ranking algorithm.

But guess what, a lot of "directories" that we trust aren't pure. Years ago I stayed at a ma-and-pa motel near Orlando. Turned out to be a very clean, quiet, comfortable place that I found when other choices were full.

I asked the owners why they weren't listed in the AAA guide. They told me that because they didn't have a lounge (for religious reasons; this particular ma and pa were pretty conservative and didn't want to serve liquor) they could never get more than two diamonds from AAA. And they told me that AAA charges an annual "inspection fee" to be listed.

If the AAA guide disclosed how much that annual inspection fee was, and next to the EDITORIAL listing for the Hilton in Orlando or the Doubletree in Houston, revealed how much the hotel paid for that ad on the facing page, consumers would have much more confidence
in the guide. Or maybe less confidence, but a more realistic assessment.

Do we really think that PC magazines have a wall between editorial and advertising? Or stereo magazines? Or the travel section in the NY Times? Would that it were true!

So kudos to Overture for full disclosure!

/rich

Comment: Jeff Stuit (Feb 9, 2002)

I agree with Rich on this one - as long as Overture is honest about the practice, charging for placement is a fine way to prioritize links. In fact, I think that Overture does this extremely well - showing how much an advertiser will be charged if I click on a link tells me something about how much a firm wants my business.

I don't think it's a matter of whether Overture's approach will scale or not, it'a a matter of how skillful they are at managing their catalog of advertisers, while maintaining a valuable experience for visitors. Overture might not end up being a huge business, but I'm sure that with good management, they can be succesfull.

Comment: Lou (Feb 10, 2002)

Rich and Jeff, you sway me... just a little. I've we're using the Web as the equivalent of the local yellow pages, i.e. as a source of information on providers of niche goods and services, then pay-for-placement seems reasonable to me.

But would it ever make sense to use this approach for the kind of topical queries that we use, say, Google for? I don't think so. But maybe that's not Overture's model anyway.

I still wonder if the small retrieval size has more to do with users' impressions of quality than does the pay-for-placement "algorithm".

Comment: Lou (Feb 10, 2002)

Of course, this item provides more anecdotal grist for the "figure out which algorithm makes sense where" mill.

So...

We now might say that popularity/citation-based algorithms make good sense for ranking results of web-wide searches. I.e., Google.

And a pay-for-placement algorithm may make sense for shopping situations. I.e., Overture.

What other algorithms (or combinations thereof) make sense where?

If we could nail down the answers right here in Bloug item #71, we'd all be filthy rich bastards in the not-too-distant future...

Comment: Chris Farnum (Feb 12, 2002)

Algorithms that spring to mind (from current search technologies):
- relevance by occurance of keywords in text, title, meta tag, xml tag...
- by statistical co-occurance of search terms (i.e. Autonomy)
- number of links/pointers (a la Google's original model)
- human indexing according to a controlled vocab/thesaurus/taxomony
- geography (when I retrieve weather via my Palm VII it grabs the report based on the location of the nearest tower)
- popularity rating
- ratings (of users or experts)
- personalized profile (past searches, purchases...)
- up-to-dated-ness

So what else might be lurking out there?
- random (could be good for entertainment)
- astrological
- degrees of separation from Kevin Bacon...

Comment: Andrew (Feb 20, 2002)

Google announced today that they'll be starting a pay-for-position program immediately. It's a little unclear as to how they'll implement this: listed with non-pay results, or off to the side in the right-hand collumn?

Anyway, they're now doing what Goto.com is doing.

See http://seattlepi.nwsource.com/business/58947_google.shtml

Comment: Lou (Feb 20, 2002)

Might not be so bad: Google won't rely solely on PFP for ranking. So the overly commercial aspects will be watered down a bit by the popularity algorithm's impact.

Comment: Eric Scheid (Feb 20, 2002)

My reading of that article says that google won't be messing with the search results per se, just the "sponsored links" part.... "The company has been listing results from advertisers under a similar format since 2000, but it previously sold space under a fixed pricing system, which prevented sites from boosting their rankings with more money."

Add a Comment:

Name

Email

URL (optional, but must include http://)

Required: Name, email, and comment.
Want to mention a linked URL? Include http:// before the address.
Want to include bold or italics? Sorry; just use *asterisks* instead.

DAYENU ); } else { // so comments are closed on this entry... print(<<< I_SAID_DAYENU
Comments are now closed for this entry.

Comment spam has forced me to close comment functionality for older entries. However, if you have something vital to add concerning this entry (or its associated comments), please email your sage insights to me (lou [at] louisrosenfeld dot com). I'll make sure your comments are added to the conversation. Sorry for the inconvenience.

I_SAID_DAYENU ); } ?>