April 07, 2003
No summary available.
More than 100,000 marketers pay for ads on Google now, which
may make it the biggest online media seller in terms of accounts.
Over the past three weeks, many MarketingSherpa readers have
written in to share their results with Google's new ad option,
so-called Content Targeted campaigns. We started to notice a
discrepancy between the rave reviews Google's PR gave itself, and
the actual data we got from users.
Last week we sent out a survey, interviewed experts and
collected data for this Special Report. Hope you find it useful!
1. Quick backgrounder on Google's New Ad Option
2. Survey Results: Click rates
3. Conversion rates
4. Sherpa's recommendations for your campaigns
5. Watch out: Google Uses Opt-out
6. Research notes & useful links
-> 1. Quick backgrounder on Google's New Ad Option
On February 26th, Google let its AdWords advertisers know that
they would be automatically getting extra ads on a new service,
'Content Targeted Advertising' at no cost until March 12th.
Everyone could try it out, and if they did not like it, cancel
that part of their account. If they did like results, they would
keep on getting these extra ads but at the regular cost (based on
what they paid for clicks in the search ads system).
The concept of putting ads next to related content is not new.
It is called contextual advertising.
Plenty of other companies, including Lee Enterprises, CMP's
TechWeb network, YellowBrix, and even Google's competitor
Overture, offer some form of contextual advertising against
Historically, in general contextually-placed ads have gotten
better click results than ads that were placed in other ways such
as general run-of-site or run-of-network. Few people can speak
to whether these clicks wound up with better conversions or not,
because most marketers were not counting conversions until fairly
Also, much anecdotal evidence in MarketingSherpa Case Studies
reveals that text-based ads are winning over basic graphical ads
online. Broadly speaking, unless your graphical ad sings and
dances to attract attention (i.e. rich media), you are probably
better off with text.
Taken together these two points, contextually served ads work
and text-ads work, would indicate that Google's new Targeted
Content Program would be a success, as long as they could be
sure ads really matched the content (articles, newsgroup
postings, etc.) they were displayed next to.
Susan Wojcicki, Google's Director of Product Management, told us
for the Targeted Content program their tech team adjusted specs
to ignore anything that was boilerplate and/or related to
navigation. Presumably your ad would be displayed only on
pages with actual content relating to your product or service.
The other factor that can make a break a contextual ad campaign's
success is the type of content that ads are placed against.
Generally newsy, how-to, and highly targeted articles on niche
sites tend to get far better ad clicks than newsgroups, bulletin
boards, general interest sites, or stagnant info pages.
Unfortunately Google did not take this factor into consideration
when designing the program. They chose the partner sites for
contextual ads mainly based on traffic (sites had to have more
than 20-million pageviews a month, which very few niche sites do)
and "quality" which seems to mean being G-rated.
However, partner content sites ran the gamut from newsgroups to
How well did ads work? Keep reading.
-> 2. Survey Results: Click rates
Google's Susan Wojcicki admits, "It's true clicks may be lower."
Most MarketingSherpa readers surveyed agreed vehemently. "My
regular AdWords yield a 4-6% click though, the context AdWord
ads yield a .001% click through, said one. "Click rates for the
content-targeted ads are abysmal," reported another.
A handful of responders had equal or higher clicks with content
ads. Elizabeth Karolczak of Oskar Consulting noted, "I got a lot
more clicks but I have really weird keywords. I get .5% on
AdWords and 4.4% on the content ads. My monthly costs went up
Unfortunately Karolczak is in such a niche industry (content
syndication and licensing), that the extra clicks turned out to
be people who were looking for something else entirely. "I
wasn't getting as good traffic as someone who has typed my
keywords into Google." She cancelled the Targeted Content
part of her Google account.
The number of clicks the average AdWords user may be getting from
the Targeted Content program is still fairly low. Wojcicki says
that although numbers vary by advertiser, generally "it's going
to be less than 10% of your clicks."
Google's execs are searching out new content site partners
aggressively, so that number will almost certainly rise. Still,
Wojcicki remarked, "We are only charging for the click, and ads
will not be disabled in any way for the lower click rate."
(Normally, your ad would be cut by Google the minute it slipped
under .5% click rate.)
Why should anyone fuss if click rates are lower? (Plenty
of our surveyed readers are worried about this.) If all you are
only being charged based on clicks, what is the problem?
By the way: Survey respondents reported the following general
click rates across all the Google ads in total:
Under 1% = 25%
1.0-1.5% = 28%
1.6-1.9% = 19.6%
2.0-2.4% = 11%
2.5%+ = 16.4%
Given the breadth of industries and offers represented, this wide
spread makes sense. Anyway, what really matters is: Are you
getting a good enough conversion to make your clicks worth paying
-> 3. Conversion rates
Although Google's Wojcicki told us, "initial data suggests
[conversion rates are] the same," she is about the only person we
could find who expected Content Targeted ads to convert at the
same rate as clicks from Search ads.
We suspect that, psychologically, readers and surfers may be
generally in a different part of the sales cycle than searchers.
Even if they click, they may be less likely to take that next
step and convert to buying something or signing up for something.
Unfortunately, we can not back this with exact numbers for Targeted
Content ad conversions yet because most people are not able to
track them versus regular ads. Although many surveyed
readers suspect conversions may be lower, few knew for sure.
We asked two experts who do have tracking software for their
results so far.
Kevin Lee, CEO DID-IT.COM, who tracked conversions on behalf of
several clients, said, "The results vary by client and industry,
as well as other factors we could not identify. Overall we were
pleasantly surprised by the quality of the contextual traffic,
but do recommend that marketers measure for themselves, because
there were instances where the conversion did drop
John Marshall, CEO ClickTracks, who tracked conversions on behalf
of his own company's B2B ads said, "We can see big differences in
the behavior of Targeted Content visitors. They spend much less
time on the site (33 seconds) than regular AdWords visitors (79
seconds). They can't wait to get out of here."
He adds, "After one month we can say that on our own site, the
Content Targeted advertising repeatedly leads to one specific
action: The user clicking the 'back' button and leaving our
Why? He thinks, "people aren't in decision-making' or 'buying'
Lesson learned: All Google clicks are not necessarily equal. You
should track results to decide if it is worth investing in
Targeted Content ads for yourself. The more niche (and possibly
B2B) your offer is, the less likely it is your offer will match
the profile of the general 20-million+ pageview sites these new
AdWords are running on anyway.
Lee offered this suggestion for MarketingSherpa readers who do not
know how to track Targeted Content Ads separate from regular
"To separate out your content traffic, you need to use your log
files or tag the content visitors when they arrive. For either,
you need to examine the referrer. look at the referrer for
"pagead.googlesyndication.com," and see not only the fact that
the ad was a content ad, but also the original site where the ad
ran. For example, from my wife's site log file (she is a
All the information you need to tag (identify) a visitor as being
from a content ad is right in that referrer."
By the way: Survey respondents who track conversion rates, got
these overall rates for all of their various Google accounts (not
breaking out Content vs. Premium vs. AdWords):
a. Marketers measuring a direct sale as a conversion
(Note: These respondents as a group were 70% B2C):
Under 1% = 43%
1-1.99% = 17%
2-2.99% = 20%
3-4.99% = 9%
5% and over = 11% (mostly under 15%)
b. Marketers measuring a no-cost registration (or sign-up) as a
conversion: (Note: these respondents as a group were 56% B2B)
Under 1% = 14%
1-1.99% = 14%
2-2.99% = 14%
3-4.99% = 17%
5% and over = 41% (mostly under 15%)
-> 4. Sherpa's recommendations for your campaigns
Keep using Google's Content Ads if:
a. You are seeking brand awareness. Especially for B2C
(business to consumer) brands with broad appeal. The current
pricing model means you may be getting a whole lot of
impressions for a fairly low CPM (when you do the math and
count your PPC that way) and at least content targeting does
imply that your ad will show up on a page somewhat related to
- and/or -
b. You can count conversions from these specific ads and your
ROI is good enough for your needs.
- and/or -
c. You are getting so few clicks (and several dozen of our
readers surveyed said they had not gotten any Content clicks
at all yet) that it is not costing you anything, so what the
heck. For now anyway.
Bail on Google's Content Ads for now if:
a. Your goal is to reinforce your brand's essence or
positioning. You cannot control which sites your ads will
turn up on at all beyond the "relevance" of content which may
or may not be perfect. Brand positioning campaigns rely to a
great deal on the media they appear in as reinforcement of
your message. Showing up on any site that happens to have
info on your topic is not good enough.
- and/or -
b. You can not measure conversions or awareness created by these
ads, and you are spending what you consider to be a
significant amount on them.
Plus, here are three more MarketingSherpa recommendations to
improve your Google results of all kinds:
Recommendation #1: Send clicks to a specialized page whenever
75% of marketers tracking conversions who sent clicks to
specialized pages got a 1% or higher conversion rate. Only 52%
of marketers tracking conversions who sent clicks to a general
page such as their home page got a 1% of higher conversion rate.
That is a big difference. Pay attention to it!
Recommendation #2: Check your Google campaign management
screen frequently, at least a couple of times a week, and tweak
your campaign accordingly.
Over and over again, as we looked at respondents with the highest
conversion rates, one thing was blatantly clear. As a group
33.5% checked campaigns daily, another 28% checked at least twice
As a group, marketers with the lowest conversion rates, checked
their campaigns far less frequently (43% checked only two) three
times per month, monthly or even less frequently than that.
Recommendation #3: Track your conversion rates.
We were sad to learn that 37.8% of survey respondents were not
tracking conversions from Google clicks whatsoever. Another
33.4% were tracking but only from all their Google clicks
combined irregardless of search term or Google ad type.
19.3% of marketers said their goal for Google campaigns is
"awareness" instead of conversions. They want people to click
over and check out their sites but not to take any immediate
action such as buying or signing up for a newsletter.
If you are in that pool of marketers, you should not consider
yourself "off the hook" when it comes to conversion tracking.
Instead, you should track clickstreams, visit length, and return
visits, to determine whether those clicks really looked over your
site and will remember it, or if they hit the back button and
At the very least, this data will arm you to defend your
advertising budget the next time management questions it.
-> 5. Watch out: Google Uses Opt-out
Some marketers we talked to were unaware that Google was using
opt-out for the new advertising, and that they would be charged
for all clicks received after the two-week trial ended on March
12th unless they took action.
Google's Wojcicki said, "We sent them a newsletter, and it's
obvious every time an advertiser goes to the online reports."
The problem is most people don't read every word of an email
newsletter and as our survey showed, 27% of Google advertisers
responding checked their campaign results less than fortnightly.
Marketer Bob Schnyder of ECD Inc who checks his campaigns twice
daily and spends thousands each month on search marketing said,
"What they don't realize as big as they are is they need to tell
you many times more. To be honest I got something about it, and
I looked internally when I log on to AdWords, but I didn't see
the little button to stop it."
Michael Herman, Manager of Internet R&D for ChristianityToday.com
said, "I assumed it was opt-in - that was my first assumption.
They didn't do a good job of letting even us geeks know."
Andrew Goodman, author of '21 Ways to Maximize Your ROI from
Google AdWords Select' said, "The opt-out is not evident on the
main campaign summary page. It's not that easy to notice."
Our take on this: Google may be a beloved brand and a great place
to advertise, but charging clients for a new service without
proactively asking first is not best practices in doing business.
That said, Goodman and others pointed out that at least with
Google you have the option to stop the Content ads. Overture,
which many praise for better campaign reports than Google, offers
no such option.
-> 6. Research notes & useful links
We collected much of the data in this Special Report by running a
survey to a selection of MarketingSherpa newsletter readers who
tend to be professional marketers working for mid-size
corporations in the US, with a 60% bias towards B2B.
Although we sent the survey link to approximately 38,000 people,
we asked that only marketers who had conducted a paid Google
campaign in 2003 respond. 406 actually took the survey.
Admitting our mistakes. Two of the questions in the initial
survey were set up incorrectly so answers were not always
accurate. Luckily this bug was caught quickly and corrected.
We did not include data from the "broken" answers in our results
data because we did not want to skew results.
Last word: It is early days yet in the Targeted Content front,
so no results can be considered complete or official. These are
only indicators. Also, results from various industries and
offers can vary so widely that no single set of data will every
work for everyone.
That is why individual campaign measurement is so important.
We will return with further surveys in the future. Get yourself
in shape now so you have useful results to share.
Thanks, and best of luck with your campaigns.
Kevin Lee's company http://www.did-it.com
John Marshall's click results page:
Andrew Goodman's '21 Ways to Maximize Your ROI from Google
AdWords Select' report: