Google adwords- a brave new world by andrew goodman

Google adwords- a brave new world by andrew goodman

Google AdWords: A Brave New World

(A Pocket Guide to the Road Ahead)

Welcome: Author’s Note

Thanks for stopping by I drastically shrunk down the size of the plan for a complete rewrite of my original Google AdWords Handbook, for three very simple reasons First, that effort largely went into my print book, Winning Results with Google

AdWords (2nd ed.), and that book is now in stores (as of December 10, 2008)!

Woohoo! It took me about 18 months to put it all together

Second, I’m hard at work on a more focused, nimble ebook that aims to solve tough problems for intermediate-level advertisers It’s called “21 Mistakes that Kill Paid Search Profits.” This will be out in January, 2009 Stay tuned

The third reason for this little document is that I wanted a quicker and dirtier

introduction for people looking to catch up with the latest thinking on AdWords – one that could be distributed widely for free So here it is Kind of a “welcome to the big leagues, kid” piece Welcome Enjoy Got a comment? By all means ping me at my personal email address, pagezero@gmail.com

search results on Google Search, or around the web They’re diversifying, of course, and the picture is getting more complicated, but here, I’ll stick mostly to core

concepts

How big again? Google’s revenues in the United States alone are around $10 billion Making the wild assumption that an average cost per click across the whole

AdWords system is 50 cents, that’s 20 billion targeted clicks per year How many ad

“impressions” or views does that add up to? Well, trillions Sounds huge, doesn’t it? But your campaign will be the size it needs to be, and that could be small, medium,

or large Google wants your business, especially if your ads are very targeted

“Averages” are largely irrelevant In your AdWords account and across the whole system, the range is vast between keywords that are valuable to your business, and those that are nearly worthless In the old school “cost per thousand impressions” (CPM) math, it’s not uncommon to see some highly commercial keywords selling for effective CPM rates of $500 or more (or $15-20 per click, or more) on AdWords Just

as it is common knowledge that the “junkiest” ad inventory on the web (personal profile pages on MySpace) is changing hands for next to nothing – pennies per thousand impressions Google’s bias is towards the mid- to higher- priced ads Their logic is: if ads are particularly annoying and untargeted, then why show them at all?

Here’s a factoid to get us in the right frame of mind Google actually prides itself on leaving white space where ads might appear, if the particular search query isn’t very commercial in nature, or if no advertisers are relevant or high quality enough to qualify to show up That means that they maintain a “monetization rate” at least 10% lower than their main competitors, Yahoo and Microsoft, on search results That means that if you look at every search done every day, Google will show no ads (all white space) on about 35% of the search results pages That number for

Microsoft and Yahoo is lower, about 20-25% (The numbers keep shifting, but that’s the Ads Quality principle in action.)

Fig 1 – On search queries that are less commercial in nature, Google often shows no ads,

even if some advertisers would like to appear here

The Google Paradox is that they serve users (search engine users) first, not

advertisers And the result is they make more money on the advertising program, not less, because their relentless focus on speed, quality and relevancy have made them the world’s favorite search engine… by far (In most markets around the world, Google has 65-90% market share of searches.)

Section I Don’t Fail From the Get-Go, Please!

Getting the right kind of start with AdWords is essential to success You’ll lose time and money if you get it wrong In no particular order, but as comprehensively as possible, here are some key tips

Generations of AdWords

AdWords has been revamped and redesigned several times On one hand, we’ve seen ongoing improvements in much of the functionality of how you slice and dice your ads, bids, and campaigns But the really fundamental changes have occurred in how the system treats you Many folks are working on outdated assumptions I call this version of AdWords “2.5” to distinguish it from prior versions Page 7 of an article I co-authored with Mona Elesseily, “Search Engine Smackdown,” covers the generations of AdWords in a nutshell Pages 1 through 6 are useful too We provide

a lightning-quick scorecard summary of how Google AdWords and Yahoo Search Marketing stack up in key areas

http://www.imediaconnection.com/content/14775.asp

[Guess what? Google has changed the rules again  We’re now up to approximately AdWords 2.7, so get your copy of Winning Results with Google AdWords and read Chapter 5 to get the latest take on what it’s all about.]

New advertisers in the “AdWords 1.0-2.0” era always had a feeling of shock and awe when they set up campaigns and found things going immediately wrong with them This current generation of AdWords is no different Some, but not all, new

advertisers, will face high minimum bids on their keywords (whoops, they tweaked

that minimum bid part, but the concept is roughly unchanged) which is very similar

to the old days when keywords that didn’t meet certain qualifications after running for a couple of days (basically, the ads didn’t get clicked often enough by search engine users) were “disabled.” So, the song today is somewhat the same, though the band is wearing newer t-shirts: AdWords is hard at first

Help from Google: Two Sides of the Coin

Many of you are aware that you can get technical support from Google, either by emailing them from within the AdWords interface or by ringing the company up at 1-866-2GOOGLE and entering your customer ID# into the phone system (Your mileage may vary internationally, but Google staff are helpful in most parts of the world.) In fact, you’re very likely to be able to work consistently with the same rep because things are very well coordinated now

Support staff don’t generally work at the main Google campus (the Googleplex in Mountain View, CA) anymore If you’re a national US advertiser, you may well be talking to someone in San Francisco or Irvine Depending on your vertical and Google’s way of assigning help to different kinds of accounts, you may deal with someone in New York, Denver, Dallas, or someplace else Of course, Google has many international offices, as well, and they’re always expanding Finding a helpful ear at Google is always a plus, so insofar as it is possible, try to get to know a rep.

Fig 2 – Google headquarters in Mountain View, CA Brilliant engineers; great cafeteria;

beach volleyball court

Just be aware that there are clear boundaries around the degree to which a support rep can improve your marketing Don’t just expect to hand off your whole campaign strategy to someone at Google, obviously Ain’t gonna happen You need to dig deep and improve everything at your end, and/or with the help of an expert third party

But help *is* there if you need it The problem here is that on the most vexing issues you’ll face, like high minimum bids due to low “Quality Scores,” you might not get very clear answers.

Also, be appreciative on the surface, but somewhat dismissive in your heart, of

“boilerplate” advice Over the years, Google reps have been coached in the “nice sounding” “best practices” suggestions that are best conveyed to advertisers

Pablum-like recommendations are generally easy enough to distinguish from hard core advice Once in awhile you get the good stuff from a Googler Other times, you should feel free to smile, nod, and ignore

What’s With All the “Non-Search” Inventory?

Don’t mind me, but I’m going to talk about issues in the order they seem

appropriate

The core of AdWords is of course “search” ads (the ads that show up next to or above Google Search results, or on partner search sites) This is wonderful, beautiful stuff as advertising goes I won’t waste time defending its wondrous targeting

qualities

From there, you’ll notice “search partner” traffic is available That’s other search engines that show Google AdWords ads that have been syndicated to the other search engine Sometimes it’s an Ask.com search results page Or it might be an Internet Service Provider like Roadrunner that has a search box on their home page

in partnership with Google

Fig 3 – At search.rr.com, the text ads at the top of the page are served by Google.

It’s not worth giving a whole lot of thought to the distinction The amount of search partner traffic you’ll see is relatively low these days, and the performance is

consistently good enough that there is generally no reason to disable this channel specifically

If you go into campaign settings, you can set the checkboxes for “on” or off to

various ad distribution options: search, search partners, and content network.

Fig 4 – The graphic display ads at the top and right are served by DoubleClick

The "sponsored links" on the left are served by Google AdWords

The need-to-know essentials on content targeting traffic are as follows:

• This isn’t search traffic, and people are viewing your ads on pages all over the web;

• That isn’t inherently bad, it just tends to convert to clicks and sales at a lower rate than search traffic, and is less predictable;

• Thus, you must bid lower on it – often 40-70% lower than search

• The default is that this traffic is turned “on” in your account – you must

enable content bidding (separate bids that you can manually enter) in the

campaign settings area, or disable content targeting entirely if you don’t want to have a lot of “high bid” traffic from questionable sources draining your account

• Here’s a wrinkle: Google doesn’t make it intuitive for you to enter separate content bids when setting up a new campaign Go back, and re-check things

in the Campaign Settings area Trust me, it’s worth double-checking

everything at the Campaign Settings level, especially the status of content targeting (now often called “placements”) For making adjustments to

content bids you can also do this at the ad group level or in bulk at the

“Campaign Summary” level (shown below)

Fig 5 – Bidding can be separated for "search" and "content."

• Google uses an automated system to sync up so-called “AdSense publishers” with AdWords advertisers The keywords and ads in your account, along with your bids, are used to determine whether your ad is a good candidate to show on a certain page online This is sometimes known as “contextual” advertising because of the attempt to show users ads in the context of some related interest

• There’s a whole different flavor of content targeting that works on a different model It’s called Placements or now Managed Placements (yikes, Google keeps changing their names for things) What that does is to combine

Google’s old “match on the fly” classic content targeting, with your ability to

restrict that process so this is occurring and showing your ads only on sites you choose It’s cool, but it’s definitely extra work Don’t forget your main goal: to do well on the “search” advertising auction!

• You can exclude sites in the content targeting program Hunt and peck until you find how to do “site exclusion” at the campaign level Examples of sites you might want to exclude are social networks like Bebo or MySpace, or sites that are giving you poor performance

• In the Reports tab of Google AdWords lies a complicated reporting tool that can give you pretty much any type of info you might want A Placement Performance report will show you where your ads showed up, and much more The reports can also show you whether a content site or page falls into

a special category, like Parked Domains (a glimpse is in the screen shot below) Google has added a great deal of transparency to their system, taking much of the mystery out of marketing, at least for power users who know how to run reports Don’t buy into scare tactics that vaguely refer to “shady” AdWords clicks That’s so five years ago! Familiarize yourself with the

reporting area in general

Fig 6 – A glimpse at the Placement Performance report

Beyond text ads in the contextual programs, as you’ll see from poking around with the ad creation views of the AdWords interface, there are options to place banner ads and video ads on partner sites Radio and print programs are also being piloted Don’t worry about all this for now Start with the core AdWords stuff that is proven

to work

How Not to Panic on Day One

This is even harder than ever Assuming you get through the onerous setup phase (I’ll get to this in the very next section), you’re going to want to have a strategy in mind or you’ll screw things up, big time

Work backwards: Establishing design, navigation, analytics, goals and

business process

One thing I always advise advertisers to do is to work backwards If your site isn’t well thought out, if you don’t know whether your landing pages will convert, etc – then I suggest you work that out first Sending traffic to see if a really crummy-looking, weak site will still convert for you always produces the same result: you pay for the traffic, and lose money

Similarly, if you haven’t set goals and haven’t figured out how to measure success, you won’t know how the traffic is performing An example: an auto dealership group that wants leads to funnel to the correct salespeople For this to happen, the site needs to be designed to create a lead generation opportunity, persuade people to fill out a form, and funnel that form correctly If a one-dimensional Flash designer sits

in the room with ten company staff and execs who have no knowledge of direct marketing principles and how to incorporate them into a web development process,

you’re likely to design a website that reflects a hodgepodge of priorities but leaves marketing goals and measurement issues on the back burner Get the right people

on the bus (a savvy web design firm), and the wrong people off (all the various execs and salespeople who want to micromanage bells, whistles, and font colors, even though they’ve never taken a single course on programming or design; you need one

or two qualified ecommerce project managers at most)

In another nightmare example of wasted, unaccountable traffic that we were asked

to buy for a client: the underqualified web designer and webmaster for a major beauty treatment chain hosted the website at his home, and funneled all the leads to his own inbox, when they should have been going to both store managers across North America as well as to the executive team and the marketing agency When this webmaster caught wind of a redesign being performed by a third party developer,

he refused all access to the domain and hosting accounts, making transition to a new site at the main company domain nearly impossible

Get the right people on the bus If you don’t, failure shouldn’t come as a surprise

Working backwards also entails installing an analytics or ROI tracking program You can get away with doing this second, but it’s better to know what you plan to do first

Categorization is king: go granular at the planning phase

Another element of strategy is to remember that keywords need to be sync’d up with ads that are related to those words So you’re going to find that you need to logically structure your account into campaigns and ad groups that break things down thematically Doing this in a very specific, highly detailed fashion, is often referred to as going “granular” (that is, fine-grained) with your “organization tree”

or logical plan A lot of little groups of keywords, each with their own ads, would be called a “highly granular” campaign.

How many keywords per group? Anywhere from two or three up to a couple

hundred Any more than that gets unwieldy Actually, anything over fifty is

unwieldy Googlers will often recommend 10-20, but we don’t always need to listen

to what they say, right?

Google is awe-inspiring, if you’re easily awed by massively brainy folks Some of their innovation is scary stuff! “Damn, they’re good,” etc They’re so smart, I think they have to hold back a whole lot of the features they build in the lab – they’re way ahead of us as ordinary advertisers (That’s the opposite of TV show pilots that never go to air because they are so stupid that not even average viewers will be able

to stomach them.)

I’m not really supposed to share their beta tests with you, but imagine this: what if Google allowed you to instruct AdWords to take your “lumpy” (less fine-grained) account build and automatically make it more fine-grained for you, leaving you with only a little ad copy cleanup to finish up? Don’t quote me on this, but they’re

working on this as we speak

So, until the granularator is released to the public (granularator being my imaginary name for that tool), the biggest mistake you can make is to write a very small

number of ads and to begin dumping huge numbers of keywords into the account in only one or two ad groups It’s not effective Don’t do it this way

You can’t test haphazardly from Day 1

In fact, in the new world of AdWords the bar is higher than ever because you’re being evaluated for something mysterious Google calls “Quality.” In AdWords 1.0 back in the day, quality was measured on only one variable (other than the amount you bid on a keyword): clickthrough rate or how often your ads were clicked Now, quality is a whole formula

One of the elements of so-called Quality-Based Bidding is that new accounts don’t have any data yet, so Google’s system guesses as to whether you’re putting relevant keywords into the system They don’t share the exact formula, but they might be looking at whether your keywords seem to match up well with your ads and landing page There are a number of other things that are assessed through a combination of automated and editorial review The end result is that even on Day 1 you can begin

to establish a history that may set the tone for the next 6-8 weeks of your

campaigns Beyond setting things up carefully, you should also be worried that initial data – clickthrough rate and potentially user behavior on your website – is being watched closely and might quickly tip your Quality Scores into Poor territory

So as sad as it sounds, you can’t go throwing keywords into the system “to see what sticks” anymore Instead, you need to focus on carefully thought out keywords – the crème de la crème of your potential keyword list, so to speak Start with these, and meticulously written ads and well-chosen, user-friendly landing pages Then let that good history establish itself (including a sound regime for ad testing, discussed extensively in Chapter 8 of Winning Results.) Only from that point should you focus

on “keyword expansion” to increase your overall volume

You’re Under Evaluation: CTR and Other “Quality” Elements in

Quality-Based Bidding (2008 Edition)

CTR = Clickthrough Rate, or the number of times your ad is clicked divided by the total number of times it appears on a user’s screen

Let’s get this out of the way up front: getting high clickthrough rates on your

keywords (and the contribution to that from the ads you write, and the way that you match ads to keywords in your account) has been a key feature of AdWords all along It’s nearly as strong now as it ever was

The old AdWords (from February 2002 through August 2005) ranked ads on the page based on a simple calculation:

[Your Bid on a KeyWord] * [Your Historical Clickthrough Rate] = AdRank

(That is, Max Bid multiplied by CTR = Ad Rank which determines your ad’s rank or position on the page.)

The higher the ad rank among advertisers vying for top ad placement on the page, the higher that advertiser’s ad will appear in rank order of ad positions 1 through

10 (10 or more.) If you’re ranking well down the list among competing advertisers, your ad won’t even appear on the first page of search results

So, while how much you’re willing to pay for a keyword partially determines your ad’s visibility, that other element – CTR, which is sort of a proxy for ad “relevance,” –

is important, too In fact, being diligent and careful in campaign organization, and testing ads, was always the secret to “economizing” on AdWords, because this would maximize your CTR’s

A few quirks of that (former) method of calculating ad positions were as follows

First, CTR’s tend to be much higher in the top two or three ad positions, and

generally speaking gradually decline as the ad falls down the page To avoid undue

ad position effects to prevent anomalous behavior (like the “lock-in” strategies of bidding to #1 position to rack up high CTR’s to “seal off” your advantage over

competitors some of us promoted back in 2002), Google accounts for ad position in their calculations

Second, how do we define “historical” CTR? Google always grappled with how much history to look at, and how much to disclose A particular problem was: the majority

of keywords in the universe of keywords searched by users were low-volume,

providing limited historical data It might take half a year for your account to show much of anything Not a very intelligent system

Google adopted a number of awkward workarounds to stop people from taking advantage of holes in the system One of these was putting keywords “in trial” or “on hold” as the system evaluated them This was always a murky way of doing things

Furthermore, Google had a 0.5% CTR for keywords (normalized for ad position)

“cutoff” point, below which your ads would simply be disabled While this indicated their strong preference for relevant ads and keywords, volatility in CTR’s could knock out perfectly good keywords, causing advertisers major headaches in getting them reactivated

Through it all, the importance strong CTR’s remained a consistent principle in the ad auction Along with that emerged a nascent idea which would evolve further in 2005 and beyond: the importance of establishing a strong account history

The “Big Think” – Quality-Based Bidding Versions 1 and 2 (AdWords 2.5

and 2.6, I Call Them)

Google had to solve a whole bunch of quirky issues that arose in earlier versions of the auction They also seem to have decided that they had a number of larger

problems to solve: unscrupulous advertisers taking advantage of the system to run massive campaigns at the expense of Google’s resources; inconsistent editorial policy decisions that could be an embarrassment for the company if customer

support reps gave conflicting advice; and the increasing complexities of a few “don’t

be evil” policies that stretched beyond keywords and ads to the advertiser’s website [An example of a rule enacted to signal Google’s commitment to the user experience was: no pop-up ads on landing pages Predictably, this spawned the invention of

“not quite pop-ups” (pop-ins, coded differently) that got around the literal rules but probably just stiffened Google’s resolve to devise solutions that wouldn’t require them to get into little sparring matches over every element of their policies.]

It must have become evident to Google engineers that a shifting list of rules

sounding not dissimilar to a grade-school teacher’s “don’t chew gum” stipulation was not suited to the shifting terrain of evaluating user responses on more general principles Saying “don’t chew gum” only means that some wiseacre will hold a stick

of gum in the air with their right hand, while moving his jaws vigorously, until some new rule is made Google wanted to come up with a system of laws, so to speak, that was not based on childish gum-chewing edicts but rather, a complex

multidimensional “scales of justice” that would give “good” advertisers the green light, while slowing or stopping the progress of “evil” ones

Google managers also wanted to use advanced economic techniques to squeeze even more profit out of the existing advertiser base

There is a certain perverse elegance to the way that a bond rating agency or an actuary in an insurance company can roll a complex reality into a simple, defining number This simplicity – in part because it can make the appearance of subjectivity, and the apparently “negotiable” elements of the process, vanish into a black box – must have been attractive to many Googlers, faced as they were with growing

complexity and a growing need to justify often unpopular decisions to a messy constellation of actors (Google stockholders, search engine users, nice advertisers, naughty advertisers, government regulators, and so forth)

So the team working on AdWords ranking methods thought and thought, and

devised a system that would solve a laundry list of problems in one go

To quote WKRP in Cincinnati anchorman Les Nessman, in his parody of a typical

glad-handing politician, Google’s so-called quality-based bidding system would have

to be “at once simple, yet complex; firm, yet flexible; and above all, fair to every American.” Aha So this is why I must sometimes feel that “Poor” Quality Scores are unfair… I’m Canadian We can be such pests Don’t even ask the French what they think

Beyond that, quality-based bidding had to be something you could express in

computer code so that it could take an accurate reading of a complex reality, and do

so in an efficient and consistent manner There are too many data points for

humans interacting with websites and coding their assessments in Moreover,

overrides are available to editorial staff both to correct overly harsh Quality Scores and to punish obvious violators of policies who have somehow managed to score too high by tricking the system How, exactly, these manual human interventions are integrated into the process is 99% opaque, so therefore in a form of translucence you’ll only consider beautiful if you’re on some weird drugs we can’t seem to buy up here in Toronto

From the boilerplate responses to your questions to AdWords support reps on these issues, you’ll see what I mean They won’t give you the formula, or info as to

whether a policy specialist looked at your account and turned some knobs, or

anything like that

The good thing is, Google makes extensive information available on how Quality Score works

http://adwords.google.com/support/bin/answer.py?hl=en&answer=49174

This document changes frequently At present, they bullet-point eight factors

without discriminating among them, describing the mathematical relationships among them, or giving you much inkling about the weightings In reality, landing page quality – one of the factors – is treated differently from relevancy factors like CTR, and in itself is a complex, ever-shifting, and opaque algorithm By “differently,”

I mean that only a minority of advertisers will have problems, and landing page issues generally don’t factor into Quality Score enough to derail a strongly relevant campaign

Note in particular that the Quality Scores used to determine ad rank on the content network don’t spill over to affect your campaigns on the search network and Google Search.

Following the first release (I dub it AdWords 2.5) in August 2005, there were

actually two Quality Scores responsible for two different processes in an AdWords account: (1) keyword Quality Score for ad ranking purposes; (2) Quality Score for keyword status (minimum bid) That was rather confusing to advertisers The latest version is even less clear, and will no doubt continue to change My concern is that advertisers will run around trying madly to optimize a lot of detail in their accounts

I want to assure you that strongly-built, conventional, categorized and tested accounts today run very smoothly in AdWords as they always did, and the advice provided in 2002 and 2005 in my publications (for example), and that

well-provided in the 2008 edition of Winning Results With Google AdWords, is really all you need to do well You should not attempt to predict every hiccup in the

algorithm

Just take Google being transparent today about the fact that the historical CTR on the display URL’s in an ad group is included in the calculation of Quality Score For years we’ve been testing display URLs for our clients, as part of A/B and

multivariate ad tests In a seven-variable multivariate (partial factorial or Taguchi)

ad test, display URL is one of the variables Page Zero recommends testing Best practices pursued by the best basic-level advertisers still work And advanced

tactics like testing your ads very thoroughly, including variations on display URL’s, still work

If you thought hard about the implications of Google including some of these factors

in their algorithm, it might be pretty eye-opening What if eBay.com as a display URL was starting to lose favor with search engine users? What if that fact alone caused