AOL Search – a case of sincere flattery, but a bad business decision?

First, let me congratulate the AOL team on its latest acquisition, Tacoda. I think that since Randy and Ron have taken the helm, the acquisitions to continue to fill out the advertising infrastructure have been logical and done at sane valuations. At this point, the non-search related portions of AOL are as strong as any of the other majors, and you could even say that continues to be the best-in-class player out there in the display ad network arena.

That said, while there are a number of really good things occurring, the recent decision to clone the Google search UI is one that I think needs a major revisit.

A brief history of AOL search for the uninitiated: AOL did a deal that, in Mid 2002, swapped out the prior Inktomi-powered search experience with a Google-powered one. The deal was a seminal one for Google, garnering them a lot of US search traffic at a critical time in their history (remember the more dominate position that MSN and Yahoo had in search in prior years).

The nature of the deal involved really favorable economics, as well as warrants in the pre-IPO Google. The deal did, however, have a number of constraints on what AOL could do to customize the Google search experience. The deal was a good one – AOL had much more flexibility than any other Google partner to-date, but even with that flexibility, it was probably correct to paint a picture of AOL being able to build a creative veneer around the core Google results, but not much else.

Well, the search team, led by Jim Riesenbach, Gerry Campell and Mark Canon, actually did a really great job, given the cards they were dealt. They did a number of things to allow for creative monetization of high-value searches, and built an underlying architecture that allowed real-time mixing of a number of external feeds, published in a customized template for each category of search terms. They were able to serve up the results with very little latency, given all the pre-processing they were doing. Overall, a nice piece of work. They also championed the selective use of editorial results in the header of popular searches (later copied by others), giving users really good payoff in the top half of the initial results screen displayed.

The growth in search-related ad revenue was pretty dramatic under their watch, but they knew that they had to test moving further out on the efficient frontier of “veneer design”, if AOL was ever to establish its own unique voice in search.

Their last major release (started long before their departures) was AOL FullView, which built on all their prior learnings/feedback on everything from consumer lukewarm acceptance of clustering, to the navigational issues of excessive reliance on horizontal tabs to toggle between result set types (text, image, video, etc.)..

Here is a sample of what one of the FullView templates looked like for an artist-related search (in this case, Dave Matthews):

AOL FullView Template for Artist-related Search

To me, it resembles more of what has morphed over the last several months – delivering a composite page with lots of rich content payoff above the fold. Compare that to a search for the same artist search today on AOL. using the new “look like Google” philosophy.

AOL Search As It Exists Today for Same Artist

It is drab for a content-rich search like a music artist, and is a pretty forgettable experience. Note the heavy use Google branding at the top and bottom of the page. I understood the desire: If consumers felt that they were getting “Google results” from any search originated from the AOL network, maybe they would reduce their tendencies to use the Google site directly. This search traffic capture objective, combined with the newest Google deal (done as part of the Google investment in AOL in 2005) to allow AOL to start selling their own version of adwords to major advertisers, certainly has intellectual appeal at some level.

The challenge with this approach is fourfold:

1. We lost the impact of some of the tips and techniques we had developed as a “veneer” player that maximized the revenue we could get from high-value searches – maybe that will be offset by the private label adwords, but maybe not…

2. Over time, the average consumer is going to develop more of a tendency to just go to directly – I certainly have begun doing that. The value of AOL as part of the value chain in the eyes of the consumer is getting pretty minimal.

3. It under-leverages AOL’s ability to deliver unique rich content assets from within the TWX network.

4. It is, by its nature, defensive in nature – how could you possibly grow search share materially this way?

I do think AOL has some great assets, and its video search is best-in-class, thanks to the Truveo acquisition. A great video search, however, will not have enough of a material financial impact to offset the gradual decline in core algorithmic search that this new Google-clone approach portends.

It’s time to revisit the decision before Google further erodes AOL’s position here. FullView may not have been the answer, but this current approach is definitely not a winning playbook.

Update: Saw the latest TWX earnings announcement this morning. AOL ad revenue up only 16% year-over-year (they did warn about this earlier in the quarter), but certainly better that some (YHOO). Under the assumption that continued to do well (since it is less tied to AOL pageviews), I am not sure how the rest of ad/search revenue did (I am assuming lower than the 16% average). I had hoped the paid subscriber declines would slow on a percentage basis, and the 1.1MM sub loss in the quarter pains me – it creates an environment of more OIBDA/cost pressure going forward. Here’s hoping that a lot of them were converted to free AOL users – I think we’ll hear about that “save rate” on the earnings call, since it is a critical leading indicator of the new strategy.