PPC Hacking: AdWords and Quality Score Secrets Revealed

Read previous post first for Part 1: http://www.ppchacking.com/2009/02/adwords-quality-score-basics/

…continued…

Now we need to ensure that the content of our landing page will get a good quality score. This quality score will be determined instantly by the adbot who will visit your site as soon as it knows the destination URL - so be careful your landing page has to be 100% ready before you make your advert.

Basically the adbot has an algorithm to determine what the quality of your site is. This will then later on be blended with other factors like the CTR.

This algorithm is different than the typical SEO algorithm Google uses as here incoming links are completely irrelevant: a brand new site has the same chances to get a good quality score than a very old established site. The judgment is done exclusively about the content of the site itself and not the reputation of that site on other sites (incoming links).

What we suggest is the following:

1) Ensure you have the keyword in the main headers:

• Title
• Meta Description
• Meta Keywords
• H1,H2,etc…
• Body

2) Ensure you have a decent amount of pages (Adbot punishes 1 page sites). It’s fine to just put some links to internal pages that are just articles in the footer.

3) Have the following pages:

• Privacy Policy
• Terms And Conditions
• Return Policy
• Contact Us
• About Us

I know this may seems stupid but I guess somewhat they have determined that quality sites always have those pages (possibily because their legal departement asks them to do so - which means they are big enough to be able to afford a legal departement!)

Doesn’t seems to really matter what you write in those pages, just try to write a decent amount of text so that the bot is happy.

If you follow those rules and you bury those pages in the footer even of a single page sales letter you should get a pretty good quality score. I rarely get less than 8 or 7 depending on the keyword if I do all this properly.

Twitter This

Username: Password:

Bookmark It

Giotto De Filippi adwords, adwords bot, new sites, quality score, sales letters

In my opinion quality score is a real pain, however it’s not necessarily a bad thing once you understand it. The main reason is that it’s not completely obvious to trick - so most people won’t manage to trick it - they’ll lose money and quit. That means that whoever manages to understand it will make much more profit than if it was simple and with a lot of competitors.

The first thing to understand is that there are several quality score phases:

Phase 1: The AdBot visits your site, and assigns you a quality score based on:

• The content of your site (Landing Page)
• Your advert (Keyword Relevance)
• The time to load your page
• How much you bid for that keyword - This one is not declared by Google, but we discuss it in this post: http://www.ppchacking.com/2009/02/dont-be-a-miser-with-adwords-or-at-least-dont-show-it/

Phase 2: Actual performance on your advert (CTR)

Let’s start with Phase 1:

What you need to do is:

• Ensure that you have a decent server and the page doesn’t take forever to load (this could be more of an issue with pages that are generated on the fly rather than a slow server by itself - even if I never had any issues with this and I was never penalized by Google because of this even when my pages were dynamic and took a few seconds to load)
• Use the exact keyword you are bidding for twice in your advert, once on line 1 (The title) and once either on line 2 or 3 (the 2 description lines)
• If you do this and you bid the right amount - read: http://www.ppchacking.com/2009/02/dont-be-a-miser-with-adwords-or-at-least-dont-show-it/ - then the initial quality score you’ll get will be about the content of your site - something we’ll see on the next post

In the next post we’ll see the more advanced things related to site content and also Phase 2.

Twitter This

Username: Password:

Bookmark It

Giotto De Filippi adverts writing, adwords, adwords bot, ctr, new sites, quality score

Let’s say that you targeted a specific keyword, but you are not making money on it.

Let’s say it’s “laptop”.

Of course there are MANY things you could do that I’m not going to talk about in this post, like changing advert, position, match type, geolocation, etc…

Let’s talk about finding ALTERNATE keywords.

You probably know that if you bid on [laptop] (exact match) you’ll get searches of people that only search that keyword. So if they search “laptop” they’ll see your ad, if they search “cheap laptop” they won’t.

Let’s assume that you are sure that “laptop” is your niche, but you are losing money.

Maybe you should not target the main keyword, but alternate keywords related to it. What’s the best way to discover those keywords? Simply to let Google do it for you.

You could just bid on the keyword laptop with broad match, however you’ll be still be receiving many clicks for “laptop” that will lose you money.

So what you can do is bid on “laptop” broad match, and then bid on “laptop” exact match as negative.

It’ll look like that in the account:
laptop
-[laptop]

Now you’ll get clicks related to “laptop”, but that are not “laptop”.

Of course you need some web analytics software to track the new keywords you receive.

Twitter This

Username: Password:

Bookmark It

Giotto De Filippi adwords

The quality score is a really complex black box algorithm, there are many things that it takes into account.

Here is for example something that is very far from being intuitive or obvious:

If you just added a keyword and you bid a very low amount (seems to be correlated with what the cost of being displayed on first page will be) you may end up getting a poor quality score, and of course it’s not going to improve immediately if you then increase the bid later on.

The logic behind it seems to be the fact that if you add a keyword and you bid a low amount (let’s say $0.05) it means that you believe this keyword not to be higly relevant and your thinking is simply that if it’s cheap enough it may actually be profitable, no matter if it’s relevant.

A good trick is to bid a high amount, like $1 or more when you add the keyword. Then a few seconds later check the quality score it gets, and just lower the bid to whatever amount you want.

So if you want to bid $0.05 on a keyword, don’t add the keyword with that bid amount, add it with $1, let AdWords calculate the quality score (usually takes a few seconds) and then lower the bid to the amount you want. If you were to add the keyword with the low bid right away you would end up getting a lower quality score.

Twitter This

Username: Password:

Bookmark It

Giotto De Filippi adwords, adwords bot, quality score

There is an easy way to discover which keywords your competitor is bidding on.

This service called semrush.com allows you to type your competitor’s url and it’ll find the keywords they are purchasing on AdWords.

The way it works is that they have a crawler and a large database of keywords and they just query Google for all those keywords.

So the first difference will be that you’ll not know what your competitors are actually buying, but rather for which search queries they are showing (which actually is more important anyway).

To use it for AdWords you use your competitors URL followed by “(by adwords)”.

So let’s try: google.com (by adwords)

Here we can see all the keywords google.com is showing for:

Click here to try it: www.semrush.com

Twitter This

Username: Password:

Bookmark It

Giotto De Filippi adwords, competitive intelligence

When working on competitive keywords having a good CTR is extremely important. So it would be very useful to be able to tell if your CTR is better than your competitor CTR, and by how much.

Some people may just think that it’s simple and all you have to do is replicate the advert of your competitor (with his display URL of course) for a given keyword to find out the CTR.

In reality it’s not as simple as one may think, most people think that what determines the CTR is the advert they write, however this is very far from the truth. The most important factor that determines the CTR of an advert is the position. So basically a very poor written advert in position #1 is most likely to have a higher CTR compared to a very well written advert in position #10.

So what we really need to measure is the CTR for a given position.

The position depends on many factors, for sure those 3:
- Bid amount
- CTR
- Quality Score

The problem with the way the AdWords interface is done is that what you see as position is actually the average position. This means that if the advert appears once on position #1 and another time on position #10 the average position will be #5.5

The main issue with this is related with the fact that the competitive landscape of AdWords is completely different according to country, language and time of the day. Whereas in some countries there may be 10 competitors, in others there may be zero. Which means that the average position (and so the CTR) will depend more on where the searches that displayed your advert have been performed rather than the quality of the advert itself.

What I’ve found to be an effective way to compare the CTR (note as you probably understood this is NOT about determining the absolute CTR of your competitor, since this is impossible if you don’t manage to show up exactly in the same position and in the same countries but determining if your CTR is higher or lower than your competitor) is to make an AdGroup and put 1 keyword, the one you want to know the CTR of the advert for (for newbies: of course we can speak of CTR only for a given keyword, the keyword will probably be one of the most important factors in determining the CTR of an advert). Then you add 2 adverts, one is yours and the other one is the one of your competitor (with the competitor’s display URL). Now if you search with Google that keyword, in any country you want, and you refresh your search, you’ll see that the 2 adverts are rotating (you may have to refresh a few times). What’s interesting to see is that the position is almost always exactly the same. This may be due to the fact that in order to determine the position to give to an advert Google is not using the CTR of the advert, but rather the CTR of the keyword, which average the CTR of those 2 adverts. So by doing this you guarantee that both adverts will show in the same position, and avoid the problem you would normally have by using 2 AdGroups of having the best beforming advert move up and increase it’s CTR even more and the worse performing advert move down and it’s CTR declining even further.

It you want to do an even more comprehensive CTR analysis I also suggest to try varying the bid amount, so that you can see if the CTR difference is consistent across positions.

Twitter This

Username: Password:

Bookmark It

Giotto De Filippi adverts writing, adwords, ctr, quality score

A lot of AdWords advertisers understand the importance of a quality advert in order to get a good CTR.

CTR of course is very important since it’ll have an impact on how much one will pay per click for a given position.

However as we have seen in a previous post (http://www.ppchacking.com/2009/01/focusing-on-ctr-for-your-adverts-and-forgetting-conversions/) the conversion rate that the advert is getting may be even more important than the CTR.

The right approach to write an effective advert is to understand that that adverts are made of 4 lines, and not just 3. The last line (display URL) is actually as important as the other ones.

This requires registering several domains, that can even be very similar, and have them mirror the same site. Then you just write the same exact advert (the first 3 lines) with the exception of the destination URL.

In the example above, you can see that the advert is exactly the same, except that the second domain contains the word “cure”. This is a website about a cure, and we have 2 versions of the domain, one that just contain the name of the disease, and the other one that starts with cure and then has the same disease name.

To make an example it could be like: www.CureFlu.com vs. www.Flu.com. (I cannot reveal the website of the customer)

Intuitively it would be pretty hard to guess that this can make a difference (except the fact that usually shorter domains are seen as more “authoritative”) but the difference is actually measurable. The short domain has a slightly better CTR (around 10% better) and a better conversion too (30% more).

So the moral of the story is that it may actually be a very good investment to register 10 or more domains that show the same site, and test the CTR and the conversion. In a very competitive area with low margins the domain itself may make the difference between making a profit and making a loss.

Twitter This

Username: Password:

Bookmark It

Giotto De Filippi adverts writing, adwords, conversion rate, ctr

Not everyone knows that it’s actually possible to know the quality score of your competitors.

Sometimes it can be very helpful, for example when you don’t manage to get a very good quality score - maybe because for the niche you are targeting it’s hard to get a good quality score (sales letters, simple comparison sites), it’s hard to decide if you should keep making changes to your site to get a higher quality score, or if you should maybe focus more on the conversion, or just lower your bids.

There is a very easy way to tell what the quality score of your competitors landing page is, all you have to do is create an AdGroup, insert the keyword you want to know the quality score for, then just type exactly the advert INCLUDING the URL of your competitor. Just wait a little and check the quality score you are getting.

It’s a simple but effective trick, basically Google doesn’t check who is the owner of the site when someone buy AdWords for a site so you can just bid on keywords for your competitor, use their URL and advert and discover their quality score.

Twitter This

Username: Password:

Bookmark It

Giotto De Filippi adwords, quality score

Most people pay a lot of attention to the CTR of their adverts, but not that much to the conversion rate of the advert itself (In my experience people seem to measure performance of adverts by CTR and performance of keywords by conversion rate).

In this post I will show an example to illustrate why this practice is wrong and will lead to wasting lots of money.

As you can see from the image above, I tried 3 differents adverts in this AdGroup, trying to improve my CTR and conversion. Since I cannot disclose the advert without risking to disclose the name of the customer I had to hide the advert, just note that just one word was changed in those adverts, and every time I chose a word that was more appealing to a certain segment of the audience for that keyword.

What is interesting to see is that the advert that achieved the highest CTR also achieved the lowest conversion rate. The CTR are all very similar, all almost at 5%, however the worst converting advert has a conversion rate that is less than half the one of the best converting advert.

Now the most important thing to understand is WHY it happens that the conversion rate of an advert is not necessarily proportional with its CTR.

Many people simply assume that you should have the highest possible CTR so as to bring as many people as possible to your landing page. Then from there they believe that every keyword has a certain conversion rate for a given landing page.

However the advert plays a much more important role in my opinion than just bringing as many visitors as possible to the landing page, it’s actually an opportunity to target the right segment for your landing page.

Let’s make an example (since I cannot reveal the real keyword used in the example): assume your are selling laptops, you are bidding on the keyword “laptop” and your landing page is very generic.

You could for example try different segments in your advert:
- latest laptops
- cheapest laptops
- discounted laptops
- light laptops

The important point to understand is that those adjectives are not actual differentiators of the product, like for example saying Sony vs. Acer but they are actually things everyone want in a laptop (everyone want a new, light and cheap laptop, and everyone will be interested in having a discount if possible).

This means that of course if your landing page is about a Vaio laptop and you are buying the keyword “vaio laptop” then the best way to describe it for highest CTR and conversion rate is to include the word “vaio” in your advert, however here I’m talking about a generic keyword like “laptop” or “buy laptop” with a generic landing page.

What happens by segmenting - by the example adjectives I provided - is that you target specific segments that might react differently to your landing page, without you being aware of it until you try.

Imagine for example that the prices on your landing page are very good, then probably “cheapest laptop” will have a higher conversion. It won’t necessarily have a higher CTR (since the most price-sensitive people are not necessarily the largest segment of the people that search the keyword “laptop”) however it’ll convert that segment better.

So the success of this stategy will depend both of the size of that segment and its conversion rate. There is also another element based on the advert position that I’ll discuss in another post.

Just to make the whole post very clear, let’s use a simple metaphor. If you are looking for employees to hire thru an ad in the newspaper, if you want to receive as many calls as possible (CTR) you’ll just mention that you are hiring, but most of the callers will not be suitable for the job, so you won’t hire many people (low conversion rate). The more you make your offer specific, the less calls you’ll receive (CTR) but the more those calls will be qualified and the more chances you have of hiring the caller (conversion rate). You may get less calls in the second scenario, but very qualified people that you might want to hire would never call for a generic ad, so if you want to hire qualified people (segment) you should specify it in the ad and you’ll be able to hire more overall (more conversions).

So the best thing to do is to try all these possible segments to see which one work better for a given keyword/landing page combination.

Twitter This

Username: Password:

Bookmark It

Giotto De Filippi adverts writing, adwords, conversion rate, ctr

This is an issue that I’m sure is having many people struggling with.

When you add an advert in AdWords, Google will send its AdWords bot to crawl your site and determine it’s quality score according to several factors (Time to load the page, presence of certain elements like privacy policy, terms and conditions, number of pages, presence of certain keywords on the pages, etc…)

On that basis you’ll get a quality score number comprised between 1 and 10. The higher the quality score the lower you’ll have to pay for a click in a given spot. This is just a first quality score assesment based on the crawling of your site and the relevancy of the advert to the keyword.

After that Google will adjust your quality score also based on the CTR you are getting.

However if you get a very poor quality score (like 1) and they put your minimum bid at $5 or $10 it’s basically impossible to improve the CTR since you are not showing up at all so you don’t get any impressions/clicks.

What I found is that the quality score algorithm Google uses is FAR from perfect, actually seems it has very serious flaws, which CAN be exploited by a savvy AdWords player.

So once you write your first advert ever for your brand new site you could just happen to get a very low quality score on the first crawl, and if this happens you may be in serious trouble, keep reading to understand why.

Please note that this is not just typical with sales letter sites, but it can actually happen on larger sites (even if for more developed sites there is another easier solution I’ll describe later in this post).

What Google will tell you is that you should improve your quality score by making changes to the site. (Assuming that when you click on the details of the quality score it tells you that the landing page is not relevant, which is the hardest problem to solve).

In my experience this won’t work at all, no matter what changes you make to your site your quality score won’t improve.

I did a lot of research on this issue, and CONSISTENTLY get the following outcome:

If this is your first advert ever for your brand new site and you get a very poor quality score, basically that domain is trashed.

Means that no matter which changes you make to the landing page, or no matter if you create a new AdWords account, you’ll be given again a very poor quality score. What I then do is I register a new domain (after making changes to the site), I don’t even bother to create a new AdGroup, I just go in the Ad Variations Tab, create an IDENTICAL Ad Variation except for the Destination URL, delete the previous advert with the “punished” domain and wait a few minutes. Your quality score will immediately be reassessed and if the corrections you made to your landing page are right you should get a good quality score. If not you have to keep trying with new domains.

The key thing to remember from this is that there is no point to make changes to the landing page if it got a very poor quality score in first place. No matter what changes you do Google will remember that domain, even if you create a new AdWords account.

So what I do now is I never experiment the landing page with the real domain I want to use, I first experiment with some domains I register just for that purpose, and only when I’m happy with the quality score I get I make a mirror copy of the site on my valuable domain, since I need to be 100% sure that the first AdWords bot crawl ever for a domain must lead to a good quality score.

Now how can those flaws be exploited? Very simple…

While I was struggling with those issues, I decided to call the AdWords customer support and ask them to do a manual review of the site. They are aware that their algorithm has flaws (even if they don’t admit it) so they offer this recourse, of manually reviewing the site.

However they will manually change your quality score only if they determine that your site deserves it, means that if it’s just a sales letter site, or it’s something odd like a miracle cure, a get quick rich site, etc… there is no way they are going to manually improve the quality score. They would probably do it if your site is a good portal, a quality online store with many products, etc… but if you have just a sales letter, or you have no much content you won’t get the quality score changed.

However a well done sales letter can convert very well, so those who are able to master this will find themselves with much less competition, since all the amateurs players will have poor quality score and only the professionals will be in the game.

So basically Google as per their policy says that a sales letter site SHOULD get a poor quality score, and that’s what you get if you ask for manual review. However manual review happens if you ask for it, so it’s better not to ask, register several domains and experiment with your landing page until you get a good quality score, then mirror the site with your final domain,  add it to the Ad Variations and remove the old domain. Remember that once a brand new domain gets a poor quality score it’s for good, so experiment first with worthless domains that you just registered!

In a future post I’ll explain what are the things to try to improve the quality score of your landing page, so you know what to experiment with.

Twitter This

Username: Password:

Bookmark It

Giotto De Filippi adwords, adwords bot, new sites, quality score, sales letters adwords bot, sales letter quality score