How to Win the AdWords Game II

August 29, 2005

9. Finding your initial keywords

Here is the list of tools you might want to use while brainstorming for the initial keywords:

Google AdWords Keyword Suggestions
https://AdWords.google.com/select/main?cmd=KeywordSandbox

Overture Inventory
http://inventory.overture.com/d/searchinventory/suggestion/

Good Keywords
http://goodkeywords.com/

I’m not going to describe them in detail simply because there are already enough references and descriptions of those services on the net. Besides using those tools, you could also try checking your own past referrer logs. Check which keywords result in conversions and which ones result in higher number of page-views per visitor or more returned visits. If you are selling widgets, then the most obvious thing would be to research the variations and phrases that contain the word widgets and a verb or an adjective that implies a sale or a purchase or narrows it down closer to your target.

Once you get your list of phrases, start with the ones with the most traffic (according to the WordTracker and the Overture Inventory).
If you have more time than money, then proceed from the general terms with low maximum CPC to the specific terms with high maximum CPC as described in Broad matching or exact keyword? - BOTH!

If on the other hand, you need some results as soon as possible and you are not afraid to lose some money then pick a few more-or-less targeted phrases and bid as much as it takes to get into position 2 or 3. Once your ads are in place, give it a few days to see the results and proceed according to your tracking data.

10. People who bid too high are stupid … or not!
When I just started using AdWords, I was really bothered by all those people who seemed to bid way to high and ruin it for everyone else, including themselves. I mean, how could they possibly pay 75 cents per click on the keyword I target, when I couldn’t even pay 40 and produce positive ROI?

Most likely they are just too stupid and are losing money! Well, guess what? Some months and thousands of dollars later I realized something. Not all of them are stupid. On contrary, some of them are really smart. They know exactly what they are doing. They managed to get high conversion ratios and high CTR, so they are not losing money at all.

High conversions allow them to pay more than I could afford for each click, and high CTR lets them occupy top spots for a much lower cost.
The smart ones are not afraid to lose a general portion of a market in return for a dominance of a targeted portion.

Let’s see if we can identify the smart and the stupid for your keyword. Run a search for “widgets” and note all ads that you see.
This is your competition. Now try “free widgets” and note which ads have disappeared. This is the semi-serious part of your competition. They are most likely using broad matching with negative keywords and have enough common sense to block the word “free.” You need to watch out for those guys, study their sites, and note their positions over time; they can copy your ideas from you.

For the next step, do a search for “widgets sdfsdklf” (any set of random characters after your keyword). Those who are still here are either a division of some huge company with millions to spend or some small company that will soon no longer be there. In any case, don’t pay too much attention to them. Those who did disappear are the pros. They are the ones you’ll have to battle with. They are the ones you should copy ideas from. Note the disappeared ads and try to find them in other searches like “buy widgets”, “widgets sales”, etc.
If you find an ad of a pro for another (more specific) keyword, don’t delay and start testing it. Most likely, it has already been tested and it works. Study their ads and their sites. Those are the ones you should try to learn from.

11. If you are getting too much traffic, you are doing something wrong

The same can be said about low CTR, but I’ll just mention traffic here. Let me try to explain myself.

- I’m making too much money, I need to limit my income!

Does that sound like something a sane person would say?

Well, if targeted leads mean income, then limiting your leads is pretty much like voluntarily limiting your income. In other words, if you have to put a daily limit on your production campaigns (I’m not
talking about the ones you use for testing) then you are most definitely doing something wrong. You are either not targeting your traffic well enough or not converting it well enough. Most likely both; unless there are factors like a physical limit to the number of orders you can process.

If you suspect fraudulent clicks by your competitors then track them and report to Google. Do not limit your campaign budget because that will make those low-lives winners and in turn will make you a loser. If you know of any valid and logical reason to limit a campaign budget, please let me know!

12. Do not let your ads get down!
Sometimes, people don’t remember where they saw something interesting. Instead, they remember how they got there.
Hasn’t it even happen to you? You typed a query into your favorite search engine or followed a link off a site you frequent to find something that interested you. You looked at that interesting content / product / service for a bit, and then left that site to return to doing whatever it was you were doing.

Some time later, you decided that now would be the appropriate time to evaluate that offering you saw, but you don’t remember the address of a site or the name of the company. So what do you do? You repeat the steps that took you to that site the first time you
visited. Once you see the search engine results or a page with a list of links, you immediately recognize the one you were looking for.
It’s like recalling where to turn off some street while driving, once you actually got to that turn and everything started to look familiar.
Back to websites.

Think about the state of mind you are in just as you landed on the target page and realized that was it. Are you ready to inhale the content and seriously consider the offering? Do you maybe feel a bit excited about being able to find this resource once again?Do you, for some strange reason, let your guards down and become more receptive to the information you read on that page?

….
One more question:
How do you feel about losing such lead in that state of mind?
Sucks, doesn’t it?

That’s what you are doing when you set a low limit on your campaign budget. The point is:

Once you have found a producing combination of keywords, web copy and ad text
-keep your ad in place no matter what!

Well, assuming you get positive ROI; otherwise, I would not call it a producing combination. I know that I’ve talked about it in the previous section, but this is just too important and is worth mentioning the second time under different angle.

Not everyone uses bookmarks. Besides, people that do bookmark anything and everything they see usually have too much crap in their Favorites folder to find useful information when they need it.

13. Negative keywords

Negative keywords are a really great feature that can help you limit some unwanted traffic. The only problem is that a lot of people try to use negative keywords with broad matching as a substitute for narrow phrases or exact matches. In some cases, doing so might prove more productive. In other cases, broad matching
with negatives might be the only way to go. Still, most of the time such campaigns either cost more than produce or they are not as effective as could be. Let’s assume you target broad match widgets.
I’ll just go over some assumed facts and the most obvious negatives to help you get the feel of it.

If you are actually selling:
-free
-information
-compare*
-picture
-image
-texture
-making
* - that one might be useful if one of your USPs compares yourself with the competition. If your widgets are on the high end:

-cheap
-inexpensive
-discounted

For mass-produced widgets:

-handmade
-custom
-customized
-crafted

The list can go on forever… But!

Even if you came up with hundreds of variations of negative keywords in an ad group, you would still miss thousands more.
Every time you check your server logs and see another word you would like to exclude, you’ve already lost money on that click and you’ve lost CTR on all those who searched with that same to-be-filtered-out word, but did not click. After a while, you’ll get tired of fishing out those negatives and checking Wordtracker
or the Keyword Suggestion Tool to see what else people search for that you need to exclude.

You’ll start thinking that you’ve included enough negatives to cover yourself…
…and that’s when you start losing money.You see, If took a day off while running “not enough” narrow keywords and exact matches -
you’ve lost some potential market. If you took that same day off while running broad matches with “not enough” negatives - you’ve lost money.

Just pause for a bit, think of the way you work, and consider which one of those things is worse. Again, if you are using your referrer logs to find new negatives keep in mind that for each one of those clicks that wasted your money there were many more searches that
simply lowered your CTR. That caused you to lose even more money, since lower CTR means you are paying more for each click.

More importantly, if you are not targeting specific keywords

- you do not want to get visitors that are looking for specifics which do not match your specifics. And negatives cannot protect you here.
Did you get that last line?
OK, just in case, let me rephrase myself. If you are targeting widgets and you know that your widgets have certain attributes, you do not want to attract people who are searching for widgets with a different set of attributes. And this is exactly what happens when you use broad matching, even with negative keywords.

But when you target the exact match [widgets] (even for such general term), you get people who know they want widgets, but they don’t have any specifics in mind. In other words, they are open to suggestions. It is much easier to sell your specific widget to a person that is just looking for widgets, but has no idea what exactly it is he wants.

Negatives offer a great way to limit some obviously unwanted terms, especially for test campaigns with broad matches. Still, you should not rely on them to help you actually target your traffic.
When thinking about your negative keywords, keep one thing in mind:

Time spent on finding negatives would be better spent on finding more exact matches for that same general term.

14. Don’t be afraid to miss a part of the market

This section summarizes the point I was trying to make throughout the report, and here it is:

You need to understand that unless you have unlimited resources, the more market you are trying to cover the thinner you spread. The thinner you spread the easier it is for your competition to run you out of business. Your failure won’t be caused by one competitor, but by many unrelated companies and individuals who promote their own niches; the niches which just happen to be a subset
of the market you are trying to cover. If you are just targeting widgets then you’ll have to spend your money competing with
a guy who concentrates on expensive widgets, a girl who is selling handmade widgets, another girl who is targeting unpainted and pre-assembled widgets and a lot more guys and girls who will eat your lunch and drain your resources a tiny bit at a time.

All that before you actually get to battle with your absolute direct competition. I’ll rephrase what I said in one of the earlier sections of this report:

Do not be afraid to lose a general portion of a market in return for a dominance of a targeted portion. Keep this in mind whenever you are brainstorming on your keywords. Well, this is the end of the general part of my report.I described some ways to increase performance of your advertising campaigns. I talked about conversion ratio, CTR, CPC, ad positions, keywords and some other
important things. I also mentioned and stressed the importance of tracking and testing.

In the following sections, I will talk about the use of proper tracking techniques to improve the way you test. The idea is to reduce the amount of guesswork in every step of the process.

Visit my message board to discuss all these topics or send an e-mail to bcc-books-adwords@aboutrealstuff.com to get more of my ideas.

15. Tracking macro-actions

I discussed paying less for each click in the section about CTR and CPC; so for now, let’s concentrate on the positive ROI and higher conversion ratio.

The most simplistic definition of return on investment in marketing would be:

- How much money you make minus how much money you spend to make that money.

I will not go over the life-time value of the customer, the increased value of your business based on the size and loyalty of the customer base, and the immediate results of future back-end marketing efforts. I will also leave profit-maximization vs. revenue-maximization discussions out of this report. For now, let’s just assume that a positive ROI is when your net profit covers all your marketing expenses for any given campaign and you are still left we something.Let’s get back to one of the statements I made earlier in this report, which I’ll rephrase as:

You need to have a positive ROI to stay in business, and the higher it is the more you can spend on yet more additional advertising.
How can you make sure you get positive return on your marketing investment? You need to be able to track all visitors that come through the ad and purchase.

After that, you need to calculate your profit on those sales and look at the number of clicks it took to make them. This will give you your maximum cost per click that you can afford to pay. If you start tracking each and every ad individually, you’ll know how much you can spend on each ad and each keyword. This is called tracking macro-actions or tracking conversions (which could be a sale, a
sign-up or whatever you are trying to achieve). Tracking macro-actions is easy, but that’s not the entire purpose of tracking in helping you succeed.

Once you’ve started tracking conversions, you know how much you can pay for each click and everything looks great…
…but what can you do if your competition bids higher? Well, they are either too stupid or too smart. I went over this in the People who bid too high are stupid … or not! section.

The question is: What can you do about it?

Well, you could try to increase your conversion ratio.
Let me clarify something:

Conversion ratio is the percentage of visitors who end up becoming your customers. So if you have one order per 100 visitors, your conversion ratio is 1%. So what does this all mean?

If you convert at 1% and can afford to pay 20 cents per click to maintain a positive ROI, then you could pay 40 cents per click with a conversion ratio of 2%. Your profit would stay the same, but you would be able to pay twice as much for each click.

How can you increase your conversion ratio?

Hmm, good question. Many smart people say that the only sure way to do it is to test.
So how do you test that?
Read on…

16. Tracking micro-actions
In order to test effectively, you also need to track micro-actions, as opposed to macroactions. You see, tracking sales or sign-ups is a good thing, but it’s not good enough. You also need to track page-views, click-streams, exit pages and many more things for visitors brought by each ad separately!

So if you notice that people who click on your ad A and land on the page B seem to visit some page C more often than general traffic - you need to try sending people from that ad A directly to the page C… or maybe adding a more noticeable link to the page C from the page B… or maybe… well, you get the point, don’t you?

What if you are targeting a keyword metal widgets and land visitors on the category page that displays 20 different widgets. And you notice that people who click that particular ad, seem to click on a silver widget at the bottom of your product listing. Well, maybe you should move it to the top? Or maybe you should add a sign “Special” or “New” or “On Sale” next to that item? Maybe you should realize that in people’s minds, “metal widgets” equate with “silver
widgets” and not “steel widgets”? Contrary to what you have believed? You never know until you test.

In order for you to make constructive changes to your site, you need that tracking data. And if you have hundreds of keywords, there is simply no way to get that information without tracking micro-actions on per-ad basis. Let’s take it a step further.
If you notice that people who search for metal widgets end up looking (not necessarily buying) at silver widgets, but people who search for white shiny widgets end up looking at steel widgets - shouldn’t you land them on different pages, with slightly different layouts and product offerings?

Don’t make assumptions about how people act - test and track their behavior. This is the only way to know for sure. Who said you have to send your AdWords traffic to the same pages you use for other
promotions? Who said you can’t create different pages for different ads and different keywords?

Combine your macro-actions and micro-actions tracking and you’ll get another piece of valuable information. You’ll see what people look at as opposed to what they buy! Once you get that data, you can use it to up-sell, cross-sell and do many other wonderful things that increase your average order size, and let you bid more for each
click, while still making profit. And that’s not all!

If you separate your data based on your ads and keywords, you can maximize your conversion for each ad you run. The point is:
You need to know what your visitors do on your site based on where they came from! OK, this covers most of the tracking. If you are not doing what I’ve described in this section - you are missing out. Send a blank e-mail to bcc-books-adwords@aboutrealstuff.com with a subject “Tracking” to receive more tips on different concepts of trackin g, analysis, and various ways to implement it.

17. Tracking continued - testing the page

Let’s back up a bit, and talk about individual pages as opposed to visitor click-streams. In this section of the report, I’ll use a term “attribute” of a page. The attribute can be anything from a headline, to a navigation button to a call-toaction line. It’s just a visual element of the page that can have different value or be presented in a different way.

Once you got the click-streams and structural layouts in order, it is time to start working on the attributes of your pages.
From this point, I’ll talk in the context of a single page (mostly about a product offering page), but the concept can be applied to many groups of pages, to sales pitches (or parts of them) or even to your ads. I’ll talk about optimizing your ads later on in the Ad copy section. For now, your goal is to increase conversion ratio, so you can beat your AdWords competition and still make profit.

What you say and how you say it affects your conversion, so it’s only natural to test and track your attributes (elements of your pages).
Let me give you a few examples of questions that attribute testing and tracking can resolve:

Which sign on your “cart” button converts better?
● Add To Cart
● Checkout
● Buy Now
● Buy This Item
● Purchase

Which sign describing the stock status of an item converts better?
● This item is available
● This item is available
● This item is in stock
● This item is in stock!
● This item is currently in stock
● This item is currently in stock!

Which way of showing the actual price converts better?
● Price: $12.95
● Your Price: $12.95
● Our price: $12.95
● Best Price: $12.95
● Your Best Price: $12.95
● Our Best Price: $12.95

Which way of showing the “regular” price converts better?
● Regular Price: $19.95
● Everyday Price: $19.95
● Regular Price: $19.95
● Everyday Price: $19.95
● Not showing the “regular” price at all

Those are just 4 out of hundreds of possible attributes. I just described content, so don’t forget about colors, fonts, alignments and other visual variations. And guess what? You need to know which ones increase your conversion ratio! Guess what else? You need to know which ones increase your conversion ratio based on the source of traffic, per any given ad! Guess what else besides that? You need to know which combinations of those attributes produce higher conversion ratios! For example, What if using “Add To Cart” sign on the action button together with a price description “Your Best Price” works better than some other combination of those two attributes?
For that particular ad with a keyword “metal widgets”?
But not for “white widgets”? How would you know that without testing and tracking?

If you don’t have such data then you are missing out on a huge opportunity to rip your competition apart and leave them wondering about what happened! On the other hand, what if your competition is testing and tracking their attributes? What if one correctly matched attribute increases their conversion ratio by just 0.02%? What if they test 100 different attributes and tune each one to add 0.02% on average to their ratio? That’s an extra 2%!
In such case, if you convert at 1% - they convert at 3%. Which means they can easily pay 60 cents per click where you can only pay 20.

Not a good situation, is it? But do you know what’s the worst part of all? You’ll have no idea about their conversion ratio and the price their are paying per each click. You’ll just keep wondering how can they possibly be in business, running those ads month after month.
Well, enough about your worst nightmare (called reality). Here is the real problem … and a solution. How much time would it take you to test 100 attributes? Probably a lot. And the worst part lies with the famous 80/20 rule (or a 20/80 rule, depends on how you look at it). 20% of your attributes will affect 80% of the result.
In other words, you will waste your time testing 80 attributes out of that hundred. The only practical (even though not a full-proof) solution is to test in parallel. I’m not going to get into much detail in the this report, but you should get some idea on what to do and how to do it. You need to split your traffic into different groups, each with a different set of values for your attributes, automatically rotated on random for each new visitor, but kept consistent for all returning visitors.

And you need to be able to track the micro and macro-actions for the values of those attributes, so you can see which ones are more effective. Don’t think about the attributes too much, just keep adding variations and check the results. The idea is not to predict which version of which attribute will perform better or which attribute is more important. No matter how hard you try, you’ll fail to guess correctly most of the time.

Instead, create variations for all attributes you can think of, and then collect the results. Your test data will quickly show what you should concentrate on. Once you have gathered some results and know your 20% (the significant attributes) - then start working on those and forget the other 80%. You can use this approach to test pretty much anything you want, but for now, just concentrate on the page attributes. OK, by now, I hope you agree with me that proper tracking is vital to your success, especially in the PPC marketing.
Again, as with the previous section:

Send a blank e-mail to bcc-books-adwords@aboutrealstuff.com with a subject “Tracking” to receive more tips on different concepts of tracking and various ways to implement it.

P. S.
I hope you enjoyed reading this report. It was not a step-by-step guide with exact numbers, charts and case-studies. I just described my findings during months and thousands of dollars spent advertising on Google AdWords. Even though no one can always be right at everything and I do not claim to be the best at the AdWords game, I still hope this report will help you avoid some mistakes. Send an e-mail to bcc-books-adwords@aboutrealstuff.com if you would like me to publish more of my thoughts and detailed experiences about AdWords (in particular), PPC (in general), tracking, testing, front-end or back-end marketing, conversions, or sales Also, drop by my discussion board where you can ask any questions, post suggestions, and get the latest updates on this report.

Good luck!
Konstantin Goudkov,
bcc@genericgifts.com
www.GenericGifts.com