|
Ok, I hate to break this to you guys.
I really do.
This is really, well, you just ain't a gonna believe this one, friends.
Here goes, because somebody has to say this. I know half the other posts in this blog are about increasing CTR, increasing CTR, increasing CTR, as kind of the Google holy grail. But one of the things that I end up doing is people come to me with Adwords accounts all the time that are messed up, for me to fix.
Are you ready?
Ok well here it goes....
By the way, this includes major advertising agencies, who pay my salary very handsomely and who therefore will remain nameless, as well as mom and pop smaller web businesses that are trying to get better results...
So for better or worse I get to look at a lot of accounts, which gives me a pretty interesting perspective. One thing I have noticed, is occasionally I find someone that does everything wrong in adwords, but they use the conversion tracking and in spite of having a messy account that makes me cringe in terms of everything I know, the person is still pulling off a really, and I mean REALLY good rate of return on their investment. In some cases they are getting a way lower cost per lead than I would expect. So they do everything wrong, but the end result is right.
I have seen a lot of accounts where the ROI is fine, but Google just disables their keywords because they are just getting too good a deal- or at least that's the way it looks sometimes. I mean if you are in position 50, your CTR will be in the toilet, but you just might get some cheap clicks, and you just might have a pretty amazing ROI too, as long as you don't get disabled. But accounts like this are not very stable, and end up coming to someone like me to improve their CTR and re enable keywords that were once performing well.
But here's an interesting concept----drumroll please....
You don't really need CTR.
Google likes CTR because Google makes more money, and it is more "relevant". But lets say someone is searching for "blue widgets"
There are exceptions, but in general everyone searching on the internet wants everything immediately, for free, but the reality is stuff costs money.
So the ideal CTR ad would be something like "Hey, we've got blue widgets here- get this (benefit) now!
The only problem is that even if you use the word "free" and its variations to filter out people searching for "free blue widgets" as negative keywords, you will still have a certain number of people who are searching for "blue widget" and hoping to get it for free.
Ok, so you have a great ad, that has a high CTR, and you get a lot of people to your site, and you feel smart because Google loves your ad, and you can now get a lower CPC because of it.
But what if you just put in the ad itself- Hey, I've got blue widgets here, and they cost $399. Pay me!
That filters out people who want stuff for free, and eliminates people clicking on your ads who aren't qualified. So now your CTR goes in the toilet, and you have been a very, very bad Google boy. But look, if you had a 3% CTR ad before, and by changing this ad text your CTR goes down to .5%, so your CPC goes up, but your CONVERSION PER VISITOR goes up by 500%, who is the real winner here?
I am not going to go through the math exhaustively, and it will be different for every site, but I just want to point out that sometimes "relevance" involves filtering out visitors you really don't want to pay for, in your ads themselves, and not worrying so much about the CTR.
The bottom line is your cost per conversion, per lead, or per sale, not the amount of traffic, the cost, or your CTR.
The apparent equation Google wants you to do is:
1) really relavant keywords to your product
2) negative keywords that filter out people who really aren't looking for exactly what you have
3) super high CTR ads that make google a lot of money and offer free stuff
4)highly relevant landing pages that are information rich, help the world, support open source and answer peoples search query for free.
5) the advertiser is then supposed to convert this into money somehow
This is what I propose:
1) really relavant keywords to your product
2) negative keywords that filter out people who really aren't looking for exactly what you have
3) honest ads that filter out the freebie seekers and the casual shoppers
4)highly relevant landing pages that sell the darn thing
5) the advertiser converts these real customers into money better and enjoys a profitable website, and spends more time at the beach.
How do I do adwords?
Well, the same old way unfortunately, testing different ads, worrying about my CTR and the relevance of landing pages, lots of negative keywords.
But now I have a hope for something better, and I plan to monkey around with it, based on some accounts that break all the rules and still come out ahead.
This would be an interesting test if you can track conversions down to the AD, which is to test a high CTR ad, against a low CTR ad that filters out people more, and look at the difference in conversion.
The first person that gives me a good case study on this will win something very cool.
Submitted by steveblom on Wed, 2006-04-05 15:37.
adwords click through rate | free stuff
|
So when do you stop testing two ads and declare a winner, in order to optimize your click through rates?
Here's a nifty free little tool that you can plug your numbers into, and it will tell you how much confidence you have in declaring a winner. Adwords has a default rotation, but sometimes decides way too soon which ad is the winner.
Adwords basically uses 1000 impressions, but I am not comfortable with that at all. I have had lots of ads I though were not so hot at 1000 impressions go way up in the click through rate, and some potential winners go way down. Most of the variance is plus or minus 3% CTR in my experience, but that can be HUGE depending on what market or industry you are working with. Here it is.....
When using this tool, keep in mind that you can't be LESS than 50% certain. If you know nothing at all about which ad is better, you are 50% certain, so if you use the tool and it says you are 60% certain, that is not really a great level of certainty!! My advice is look at the 85% range as an absolute minimum.
Ok, enough already, here is the tool:
http://www.splittester.com
Check it out when you are looking to "pull the plug" on one of your ads and determine a winner.
-To your adwords success,
Submitted by steveblom on Fri, 2005-11-25 16:43.
adwords click through rate | adwords copy writing
Adwords Copywriting and the Taguchi Sequence
For people that are always probing the secrets to Google Adwords, testing different ads that get a better clickthrough rate, so that you can reduce your costs, is an ongoing game. Taguchi gives a new level of this testing. For those not familiar with the Taguchi method, the basic idea is that instead of simple split testing, you can test many variables at once. So instead of testing two web pages, each with a different headline, you can test 40 different versions of the page, like different headlines, different graphics, different forms, including a testimonial or not, shifting the order of copy, whatever you can think of.
What the system does then is uses some complicated math to figure out which factors contribute the most to your desired outcome, and you can then rapidly look at all the factors and make your optimized page. The way marketers have been testing things for years is much different. You make two pages, each with a different headline, and do a test to find the winner. Then you make two different pages where you change something else, and find a winner from that. By the time you have tested all 40 elements, you would have spent a lot of money in testing indeed.
One of my first thoughts when I started reading about this method was how it could be applied to Adwords. If you could test for one keyword, all the combinations that could get your ad the best response and optimize it dynamically, that would be pretty hot. It turns out that Cringely already made this test, and a follow up to it, and the results are pretty hot, and then, not so hot. This is only one experiment, and more experiments should definately be done along this line. If you are a developer who wants to make a Google api to do this, definately drop me a line at steve@Adwordstraining.org and I'll make you rich, famous, and with all the chicks digging you.
At the bottom of this entry I'll give you a lot of interesting links I found so you can learn more about Taguchi and all the software that is available, and if you find something new I'd appreciate it if you drop me a line and let me know.
Now, for Adwords enthusiasts, the obvious question is, can Taguchi help me optimize automatically to get the best Adwords ads our there and destroy my competition? Well, the answer is yes, then no, then yes. Apparantly there are a couple of hidden factors that mess up the Taguchi sequence when you try to apply it to Adwords, although in theory Taguchi should work every time.This is only one test though, and Taguchi works amazingly well for web pages, so I think it has a lot of potential still.
The interview the profit doctors did with the guy who talked about the"Taguchi method" of multi-testing many variables at once, instead of
simple a/b split testing, got me interested in looking to see if there
were any resources that cost less than $150,000 per test.
Actually at the SES convention there was a company called http://www.optimost.com/ , which did tests like this for companies like ask.com
and ebay, for around $30,000 per test, and I am sure they
are using the Taguchi method for this.
Submitted by steveblom on Tue, 2005-04-26 01:06.
adwords click through rate | Adwords Marketing | mastering google adwords
| CTR and the adwords campaign improvement cycle
I can't say that I am 100% on this because after you are managing about 20 Accounts for different
clients it gets harder to do. But if you are doing adwords for yourself you should definately do this:
Submitted by steveblom on Wed, 2005-04-13 02:19.
adwords click through rate | adwords tools | free stuff
|
Adwords CTR Validity Checker tool
When you compare two ads in your adwords account, and one of them is at 2.9% and one is at 1.3% you should kill the second one and keep the first, right? Not necessesarily......
It really depends how many impressions and clicks each one has. Google sometimes decides too early which ads are best and you can easily be misled. I have seen ads I was sure were winners go DOWN a lot after even just 1000 impressions or so, and I have also seen "sleepers" gradually go from around 2% to a whopping 13% after a few thousand impressions. If you kill the ads too early, you miss this. Here is how it works....
Iif you let crappy ads go too long it kills your CTR, so when should you kill these ads and declare a winner? A workable answer is to only kill ads when the difference in CTR between the two ads is statistically relevant. You have to look at how many impressions and clicks each ad has recieved and compare that to the click through rate, and do some math. Well I hate math, sooooo............
Here is the tool that will tell you a pretty good guess:
http://www.vertster.com/adwords-tool/default.asp
What it does is look at the difference between the ads and tell you when it is safe to kill the loser, or if you should wait a while. Very good stuff.
A big adwords guru (who shall remain nameless) has a tool like this but you only get it after you have paid him 197 dollars.
This one is yours free.
|
There wasn't a lot of stuff for Die Hard Adwords fans at the SES conference, but there was one good seminar with some good advice.
The strategy is based on this interesting concept: A good portion of Search Queries are actually unique,meaning they have never been typed in before. The stats given were upwards of 50% of the search queries at Google on any givenday have NEVER BEEN TYPED IN before.
I have seen this type of research in a couple other places before, but not sure where I saw it. If anyone knows, send me the link and I will be forever grateful. For veteran SEOs this research might be unbelieveable, but if you track how you yourself are doing power searches you will see that you end up trying a lot of different versions and combinations when you are trying to hunt something down.
There was some other research done by another company though that had totally different findings. There findings were that most people use combinations of only 2 words for the most part, and rarely use even 3 terms. And never any complex phrase searching, always just2 or 3 broad terms.
How can two different companies get different results like this? I guess it depends where you look. One source of data comes from stuff like the ask jeeves peek behind the keyhole. Here's a bunch of places to check out just in case you aren't familar with this sort of thing:
Be aware that some of these things will show adult content:
Ask Jeeves Peek Through The Keyhole
http://www.askjeeves.com/docs/peek/
Shows you some of the search queries at ask jeeves
MetaCrawler MetaSpy
http://www.metaspy.com/
same deal here
Kanoodle Search Spy
http://www.kanoodle.com/spy/
Yahoo Buzz Index
http://buzz.yahoo.com/
This may get more interesting when they get their new search algorithm fully implemented.
Ok now this is one source of data, the other one was done by a company that monitored server logs of a lot of websites and additionally had a piece of spyware (that people volunteered to use for the experiment) installed on their computers and it logged everything people typed in and searched for.
It was also an interesting experiment because a "searching session" could be monitored so that you could track everything people would do in one sitdown with their favorite search engine. They found in some cases that people would type in two words, then when they didn't find what they wanted type in the SAME two words again, kind of like hoping the Engine would guess what they were looking for and somehow show different results!!
Sound like someone you know?
Anyway I thought this was pretty funny. So what does all this have to do with the strategy to improve your results in Google Adwords?
In a nutshell, the strategy is to go ahead and use BROAD MATCH on some terms and go ahead and get some slightly less targeted traffic for awhile. Could be a week, or a month depending on your budget, but the longer the better. Then analyze your server logs to see how people found you. People type in the darndest things sometimes, and looking through the logs or your analyzer software can give you some new ideas and new directions for your keyword lists. Then, when you are ready, add these keywords where appropriate and use them to expand your keyword lists and target better. Ideally, you should see your clickthrough rates improve because you are using some better keywords.
I usually do this in the opposite direction- I look through the server logs to find keywords that I don't want people to find me on, and I add them to my negative keyword list in my Adwords account so that I am notpaying for useless clicks. But I do a lot of "nichy" products and campaigns where ROI is paramount, and less but more highly focused traffic is perfectly fine.
I am going to look through my campaigns and clients to see if I can find a match for trying out this strategy.
We will see how it goes.
Send your comments and see ya next time
Submitted by steveblom on Mon, 2005-04-04 16:33.
adwords click through rate
Ok lets say you have some decent ads now that are pulling a decent CTR (Click through rate)
Lets say that the ad groups are organized correctly so you are able to expand and are not shooting yourself in the foot.
The next step, once you have some decent ads that are getting at least 1% CTR or more, is to start RUTHLESSLY cutting your Cost Per Click. I have been amazed at how much you can start cutting your costs once you have that good CTR.
Just reduce your CPC (cost per click) by 10 cents each day until you are down around position 4,5, or 6 and your clicks are a little smaller, but not that much smaller. I am in position 4 or 5 in a lot of places, paying maybe 25-50% of what my competition is all the time.
This strategy isn't for everyone- If you have a site that really performs well, or you have deep pockets, by all means go ahead and spend the extra money and be number 1- You will get lots more clicks. You can still reduce your costs to be number 1 or 2 by testing better ads and reducing the price on a gradient. And you can still get the benefit of reducing your costs to be up there by 20% or so. And you will be amazed how many times you can lower your CPC and it doesn't affect your position at all!! You have to have established, high performing ads to do this though.
But I am very ROI driven, so I have found better results by simply dropping the price, dropping the position, and increasing my daily campaign budget. I have also found that being number 1 is great for the ego, but the traffic just isn't as targeted. The serious searchers go a little deeper into the first page of search results because they are hunting something down. If my ad is just a little better then the top 3 or 4, they will click on my ad, and they are the searchers I am looking for. A lot of general searchers look at the first 2 or 3 Organic listings and maybe click on the top ad or top organic spot, and off they go, but they often have a more general interest or are doing research. Whatever the reason, I have found that I am getting a better conversion rate by actually being lower on the page than being number 1.
In super competitive industries this strategy won't work, but there are lots of traffic bargains out there still, and lots of niches where this works just great.
|
| | | | | |
