In October, across our clients, Google picked up 3 points of ad spend share, reaching 79% share. This is the highest point Google has reached since we’ve begun reporting these figures, and the 3 point increase is the largest single month share gain seen this year. Google’s gain came at Yahoo’s expense: Yahoo’s share fell from 19% in September to 16% in October. Microsoft held steady in distant third, maintaining a 5% share.

To describe where these numbers are coming from, our agency manages paid search for over 100 online retailers, including some of the web’s largest brands. About 85% of our clients are B2C retailers; about 15% are B2B retailers. We manage search for clients using proprietary portfolio bid management software. The math driving our system has roots in my doctoral research at MIT, and we continue to hone our technology via ongoing collaborations with faculty from UVA, U. Rochester, and U. Santa Clara. Our bidding system doesn’t establish a priori budgets for each engine. Rather, our clients instruct us to buy clicks which generate sufficient sales to meet their profitability metrics, and then buy as many of those profitable clicks as we can.

Because of our econ-based bidding approach, a change in the percentage share of our clients’ aggregate advertising budget across the engines reflects differences in click performance and click volume between the engines.

Secondary Metrics

Back in July over at SEL, we discussed several secondary metrics characterizing the clicks we buy from Google, Yahoo, and Microsoft. (Article here: Google, Yahoo, Microsoft: Year-To-Date PPC Report Card) A few months have passed since then, so let’s revisit some of those.

This graph presents the monthly click share of each engine. Unlike the prior graph, this metric doesn’t take into account click costs, just click counts.

While this resembles the ad share graph, Yahoo’s click share is slightly higher than Yahoo’s ad spend share, indicating Yahoo doesn’t monetize their traffic as well as Google or Microsoft. This shows up more clearly in the next graph, which shows the average CPC (cost per click) across our aggregate client base by engine by month.

Again, our staff and our portfolio bidding technology strive to buy as many profitable clicks as we can for our clients. The word “profitable” is really important in the prior sentence. It explains why our agency will pay Microsoft 11% more (on average, in aggregate) for a click than we’ll pay Yahoo, and why we’ll pay Google 24% more (on average, in aggregate) for a click than we’ll pay Yahoo. The difference is simple: Google and Microsoft clicks sell more for our clients than Yahoo clicks.

The next graph presents SPC (sales-per-click), for sales directly tracked back to a paid click, by engine by month.

SPC is the product of two other metrics: conversion (the odds a click generates an order, defined as orders over clicks) and average order value (the size of the average order, defined as sales over orders). Yahoo has both lower conversion and lower AOV than Google and Yahoo.

Here’s that AOV graph:

The exact size of these lines doesn’t matter, as their absolute size is driven by the composition of our client base. If we served a greater proportion of B2B retailers (who often have higher average orders than B2C retailers), these three AOV lines would move up. What does matter here is the relative size of the AOV numbers: for our clients, Google clicks generate significantly larger orders than Yahoo and Microsoft clicks.

Finally, while these engines sell (most) of their clicks on a CPC basis, you can compute an effective CPM (cost per thousand impressions) by looking at the ratio of ad expense to impressions. Across our clients, Google and Microsoft are pulling an effective CPM of around $16, whereas Yahoo pulls near $4.

Your Mileage May Vary

These numbers represent the aggregate experience of our clients, and may not generalize to whole paid search industry.

I’ll mention again that our client base is composed primarily of online retailers.

Online retail is just one portion of the engines’ advertiser mix. Market share results could be quite different among, say, CPG brand advertisers.

We’re just reporting on our corner of the paid search world.

Bashing Yahoo? Nope.

All these results might make it sound like I’m trying to bash Yahoo. I’m not.

Honestly, I’m personally rooting for Yahoo. I like Yahoo as a company. I like their history and I use their services often (Flickr, Groups, Finance). Our Yahoo ad reps and tech contacts are smart, friendly, and helpful. Yahoo contributes a great deal to the open source community – we appreciate that. Yahoo has made huge progress modernizing their PPC platform with their Panama launch late last year. And Yahoo is currently the most popular site on the planet.

I hope Yahoo regains ground in search. Google’s ever-increasing dominance isn’t healthy for the advertising ecosystem at large. I also hope Yahoo’s presence pushes Google (and the whole PPC industry) towards more transparency.

The market will reward the firms who get it right.

May Yahoo and Microsoft start getting it more right soon.

Yahoo

What does Yahoo need to do to regain some share from Google? Yahoo’s problem isn’t click volume, it is click quality. Yahoo needs to provide advertisers with better clicks, by better matching ads to queries and by removing low quality partners from their syndication networks.

Microsoft

What does Microsoft need to do to gain some PPC share? Microsoft’s issue isn’t click quality, it is click volume. Microsoft needs to grow inventory, and do so in a way that does not compromise quality — that is, by getting more people to use MSN.com and Live search, rather than signing up poor syndication partners. They should also shore up their technology, as their API hiccups more than Google’s and Yahoo’s combined.

The Hugely Important Fourth Quarter

For many online retailers, Q4 is it. How a retailer performs in Q4 often determines their year.

Over the past few years, as more sales have shifted online, the holiday sales peak has become more concentrated, occurring faster, later, and more intensely — compared to, say, historic Q4 sales in the catalog industry.

So, over the coming weeks, many of our clients will increase their ad spend dramatically to match the holiday shopping frenzy. As an agency, we’ll be working diligently to make sure the clicks we buy for clients meet their economic targets. And, after the dust settles, it will be interesting to look back in January to see how Yahoo and Microsoft perform during these next weeks, during the hugely important holiday surge.