Google TAC Deception and Marketplace Margins

Google's better-than-expected 3rd Quarter earnings last week was welcome news to a nervous market and a valley on the verge of panic. There was a lot of juicy data to dig into, but one Q&A that caught my eye was:

Q: A little more clarification to improvement in margins. Was that through better AdSense deals, or through expense controls?
A: Across all categories of expenses people have been very diligent.....I am not aware of a change on the ad partnership side that would result in a margin change.

This caught my eye because I've been researching marketplace margins for some time, including GOOG. While there may not have been a noteworthy AdSense partner margin change this quarter, the longer-term trend is eye-opening. In fact, if my calculations are right, Gross Margin for GOOG Network Revenues has dropped around 20-25% since Q3 2006. Given how slim those margins already are, that's a shift worth watching. You wouldn't see that if you focused on the deceptive Traffic Acquisition Cost (TAC: roughly the inverse of GOOG Network margins) chart GOOG shares each quarter:

That chart looks like TAC is dropping and, thus, margins are rising. However, TAC is largely driven by GOOG Network partner payouts -- but the chart compares to all GOOG revenues, not just those from GOOG Network partners?!?

Given that my research has been focused on Google's Network, rather than their proprietary properties, I charted GOOG's Gross Network Revenue, Net Network Revenue and Gross Network Margin below. I didn't include proprietary revenues and I only included TAC from AdSense Partner payouts -- providing a better apples-to-apples comparison of revenues and margins.

The result, contrary to GOOG's official TAC slide, this chart shows GOOG Network gross margins dropping over time -- from 25% in Q3 2006 down to 20% in Q3 2008. Gross margins of 25% are already considered low in technology circles, shave off another 5% and I start wondering if something important is happening in GOOG's Network (e.g. competition for publishers).

It also makes me question what the market is telling us about appropriate margins for online marketplaces -- should they be 50%+, like many other technology businesses. Once you take GOOG proprietary properties out of the mix, AdSense is essentially a marketplace between advertisers and publishers -- with GOOG providing an opaque venue. Publishers are getting 80% of each click and GOOG is getting 20% of each click. With over a billion a quarter flowing through the marketplace, that's a strong datapoint for marketplace margins.

Are there other marketplaces that could provide a sanity check to GOOG's numbers? The largest is probably eBay, although EBAY provides much more marketplace transparency for its participants. More transparency might suggest lower margins, because it's harder for EBAY to hide their cut the way GOOG does. The following chart confirms that suspicion -- EBAY's Gross Marketplace Margin has grown by almost 50% (unlike GOOG's dropping Network Margin), but overall magnitude is smaller at 9.3%.
Therefore sellers are getting 91% of each sale and EBAY is getting 9%. With over $60 billion of Gross Merchandise Sales projected for 2008, that's another strong datapoint for marketplace margins. Note, although EBAY may use other terms for gross or net revenue, using Gross Merchandise Sales provides the best apples-to-apples comparison to Gross Network Revenues in an advertising marketplace like Google -- one marketplace sells physical goods, the other sells advertising.

That gives us two substantial datapoints for online marketplace gross margins: 6-9% and 20-25%. Do you know others? Is my math wrong (probably)? Should they both jump their margins to the 50%+ expected of technology companies? All thoughts appreciated...