Bogus Stats Mislead Results on SEM

Ed: Bogus stats mislead on search engine marketing results. The findings imply:

• Paid clicks from search engine pages (versus Long Tail affiliate pages) convert better to advertiser customers.
• Google out-performs Yahoo in their ability to convert customers.

Does Google pay Efficient Frontier to produce this garbage?

26% of Google users search, rather than use the address box, to find a website. This group converts 100%. They are looking for the advertiser's website. Excluding this cluster from the stats, we would conclude:

• No significant difference between Long Tail and pure search conversion. Long Tail affiliates convert users at similar rates to search pages.
  
• No significant difference between Google and Yahoo.

Further, we should ask:

• Advertisers paying Google for 26% of visitors who were already coming - does this make sense?
• Google generates plurality of paid clicks on their pages; not Long Tail pages. Yahoo does a better job serving the Long Tail.
• Both hide low CTR (click through rates) and conversion rates - that statistically may be in the noise range.

Online advertising needs to improve. Let's start with statistical transparency.

Clicks: 52%, 26%, 22%

Conv: 56%, 26%, 18% - Net 56/74 =75% Ratio 75/56 verus 66/54 - T-stat nil

UK Search Syndication Traffic Quality

Our post on search engine syndication network performance in the US showed that in general, pure search converts better than syndicated traffic. Efficient Frontier UK data shows a similar trend, as shown in the table below.

We looked at conversions from UK advertisers in 4 different verticals, and found that, on average, 78% of Google's paid search clicks came from google.co.uk or google.com, while only 54% of Yahoo's paid search clicks came from search.yahoo.co.uk or search.yahoo.com. (Note: this data does not include clicks from the engines' respective content networks) Microsoft Live Search doesn't show ads outside of its own search properties in the UK, so was not included in the table.

Yahoo had a wider discrepancy between the percentage of clicks and percentage of conversions than Google (78 - 82% vs 54 -66%), which suggests that Yahoo's average 662 syndicates covert at a much lower rate than Google's average 228 syndicates. Ask and AOL are Google's two biggest UK syndicates, and account for a similar share of clicks and conversions, showing that they drive high quality clicks in comparison to Yahoo's largest syndicates, which include sites like couponmountain.co.uk and findstuff.co.uk.

Data from our UK Search Engine Performance Report: Q1 2008showed that Google accounted for 85% of UK search engine spending in Q1 2008, compared to Yahoo's 12%. Google's high volume of pure search paid clicks and smaller syndication network relative to Yahoo speaks to the strength of Google's brand in the UK, both as an advertising platform and a destination site.

NEWS: Social media's uphill advertising climb

Social media's uphill advertising climb

Posted by Stefanie Olsen

MOUNTAIN VIEW, Calif.--Reality check: half of the social media start-ups at Tuesday's Under the Radar Conference won't exist next year.

That was the dour prediction of an advertising executive after a day of start-up presentations from a tongue-twisting list of tech companies--including Verismo, Mytopia, Loud3R, Jacked, Sometrics and PutPlace...

Like with Web 1.0 companies, one of the biggest problems social media companies have in the market is that they're technology driven, the executives said. Tech companies often fail to hire media-savvy executives at the top who can sell brand advertising.

Another Web 1.0-like problem: many social media start-ups are marketing the idea that they have tons of data on people's demographics and preferences. The start-ups believe that that data will lure brand managers seeking to reach new customers. Not true, executives said.

"I can't imagine a large agency giving a rat about a small firm's data. Advertisers are looking for aggregation of data...and someone at a very high level to give an overview (of what it means)," said Bedecarre.

One member of the audience, an executive from news site Topix, asked the panelists when a major advertiser would announce plans to spend $20 million sponsoring AOL or Facebook. Bedecarre answered that that kind of deal doesn't translate on Facebook because brands are too used to ads that broadcast to, instead of engage, audiences. They're still not willing to spend many millions of dollars on major Web sites, he said.

"It hasn't happened yet because big advertisers and agencies haven't let go of advertisements. They like to buy lots of ads," he said, referring to commercials.

Ballmer: Software Makers Face Big Challenges

Grant Gross, IDG News Service

Tuesday, June 03, 2008 7:40 PM PDT

Improvements in processing speeds, storage space and wireless broadband will drive a new revolution in information technology, but software makers face several challenges in their efforts to keep up, Microsoft CEO Steve Ballmer said Tuesday.

Software makers need to create better natural interfaces, simplify programming tools and create better search and analytical tools for computer users, Ballmer told a crowd of about 700 people at the American Electronics Association's annual technology for government dinner in Washington, D.C.

Ballmer told the crowd he was optimistic that a new computer revolution would happen in the next 10 years, but he also ticked off a series of challenges for the IT sector.

Computers contain massive amounts of information about their users, but they still don't anticipate their users' needs, Ballmer said. That's where better natural interfaces can help.

"You want your computer to not only understand your words, but start to remember things about you and your intent," he said. "'Get me ready for my trip to Washington, D.C.' My computer knows absolutely everything my secretary does, but my secretary does a better job on that problem today."

Part of the problem is that search technologies haven't changed much in the last five years, Ballmer said.

"Your ability to find and analyze information will go up in orders of magnitude," he said. "Really searching deeply, picking up information and being able to assemble it is still pretty hard to do."

Ballmer also suggested that software development is still too complex, and integrating separate computer systems still too difficult. The complexity of software development is still the "bane" of the IT industry, he said.

Ballmer envisioned a near future where he could be watching television in a hotel room halfway across the world from Microsoft founder Bill Gates, and connect instantly to Gates to point out a pivotal moment in a televised sports event. Ballmer could say, "Bill, did you see Tiger [Woods] make that putt," and his computer would automatically connect him to Gates and they could watch the putt together. If Gates wondered what brand of golf ball Woods was using, Ballmer could capture the image on the screen, search for matching images online and order two dozen balls for the two of them.

When audience members chuckled, Ballmer responded, "We're not that far away."...

Ed: Mr. Ballmer replacing Mr. Gates as the visionary? Not likely.

• Mr. Ballmer continues to focus on PC-centric world. Cloud-computing is the future.
• Artificial Intelligence to infer services for users - Mr. Ballmer has not studied AI with its roots at MIT and Carnegie. 10 years to deliver true intelligence may not be enough. AI is a 30 year old problem that is progressing at glacial pace.
• Yes, integration is a nightmare. Cross browser, platform, domain, and layers ... it's insane. Microsoft has been the legacy that created the problem.