Yahoo! Panama

The following is an excerpt from an article we prepared for Local Media Journal regarding Yahoo!'s new ad engine. It was published last month for our client advisory service. Given Yahoo!'s announcement today, we thought that a portion of the article would make a relevant post.

TKG thinks the speculated changes would have a positive impact on Yahoo!'s paid search business. Here's why.

CBS MarketWatch reported that "according to Thomas Weisel, in 2005, Google generated $90.60 in revenue per thousand queries in the U.S. and $89.93 outside the U.S. By comparison, Yahoo generated $41.18 per thousand queries in the U.S. in the fourth quarter … " There are many inputs into a revenue-per-search calculation — for example, query ad coverage and bid price — but a portion of this revenue difference is the way the two ad systems work. In essence, Google gets more clicks on the ads it runs.

We've seen a wide range of estimates, some that suggest this particular portion accounts for 50 percent or more of the difference. Assuming the new "ad interface" includes improvements in ad targeting, we could calculate the exact difference between the two systems once and for all.

Today, YSM advertiser positions are defined by bid price (assuming they make it through an editorial process). The more an advertiser bids, the higher its position in the search return. In some cases, this causes a problem called "squatting." This refers to an advertiser that appears relatively high in the paid search result, but does not generate many clicks because its ad is not as relevant as others in the search return. An ad targeting change would aim to correct this problem and therefore raise the RPS.

The new system apparently will function in a similar fashion to Google's paid search algorithm. It will generate an advertiser position based on the expected revenue yield. Advertiser yield is defined as a function of bid price and clickthrough rate. For example, if an advertiser is willing to pay a lot per click but doesn't get any clicks on its ad, its position drops. This is how Google AdWords works today.

The Bottom Line: We're going to be hearing a lot about this project. It's a massive undertaking and quite an engineering feat to move a live system that processes YSM's query volume to something entirely new. The system will need to be operational for a while to see where it all shakes out, but here are a few things TKG expects to see:

• Advertiser Complaints. Anytime a change is initiated that alters an advertiser position, you are going to hear about it. Let's not forget that there are advertisers that benefit from the current system. In particular, those that get "seen" a lot but do not get as many "clicks" will not like this change. For most advertisers, however, the change will be positive and they'll see it in their campaign metrics (better conversions from writing more relevant and targeted ads).

• Problems During Rollout. During test periods in various markets and the final rollout, much will be written of problems. The success stories will not be as prominent as information leaks out. This won't be because successes don't exist, but because a change like this one has long-term implications. The longer this system runs, the more it improves.

• Yahoo!'s Margins Will Increase. Specifically, the change will positively affect margin contribution from YSM to Yahoo!'s overall results. Why? With basically the same resources and equipment (avoiding a lengthy discussion of capitalized expenses) the same system will start to generate a higher RPS. A higher RPS will flow directly to the bottom line. The caveat here is the competitive environment around traffic acquisition costs. This is the amount of revenue share YSM gives partner networks in exchange for traffic.

• Yahoo! Search Stands to Gain Market Share. This is a distinct possibility. To illustrate, we chose a search with the maximum amount of screen ad coverage on both Google and Yahoo!. This search result shows a query on Yahoo! Search for "car insurance rates." Notice that the Y! internal link and two organic search returns are visible above the fold (the portion of the screen that users can see without scrolling). The rest of the page, minus the header and the search box, is determined by advertiser bid price. Therefore, most of what a user sees after completing a search, in this example, is controlled by advertisers.

On Google, the search real estate is determined by advertiser relevancy (a function of bid and user click rates) and then page rank. So, we're suggesting that YSM's change will increase the user perception of relevancy for a Yahoo! search where ads are present. Assume the difference between Google's current relevancy-based ad return and YSM's bid-based return accounts for 25 percent of the RPS variance discussed above. If you could increase the utility of your product by 25 percent, how much is that worth? If the 25 percent is actually 50 percent of the difference, how much is that worth? We'll see once the new "interface" rolls out.