Archive for the 'Affiliate Marketing' Category
One of the things that I have been trying to figure out was why I couldn’t get a 10 on any of my Quality Scores (QS). My QS values would always start at 7/10 and sit there. I had one keyword that dropped to 6/10 but the rest just sat there. It was pretty frustrating. But there was good news.
The other day, all of a sudden, one of my keywords jumped to 10/10. Not sure why that one jumped (because the rest are sitting at 7) but I will take all the little favors I can get. The interesting coincidence is that I just happened to be reading Andrew Goodman’s book titled “Winning Results With Google Adwords” and was reading the chapter on “Quality Bidding” at about the same time. Maybe there is a connection there, maybe not but I’m not knocking it.
All of that said, this whole area remains a black art which is exactly what Google wants. There are a bunch of factors that play into effect here and they are put together in some magic formula. And, apparently, the QS is calculated on the fly so that each impression is based on a QS that for that specific impression. Boy, does that make it tough.
Anyway, from what I am reading and what appears to be the case is that there are several factors. These are the ones that appear to be used:
- Relevancy – without a connection between your ad, your keywords and your landing page, Google is not going to give you any sort of QS. What this means is that you need to have your keywords in your ad and on your landing page, i.e., the classic SEO keyword loading issue. Just don’t over do it by trying to be tricky. Just use your keywords in your meta tag for keywords and description as well as in your copy.
- Historical Data – your account needs to have a history of strong CTRs. This is not just the single ad or keyword but your entire account. There also appears to be some information here about “bounces” where the user clicks through and then comes right back which would indicate a lack of user acceptability for the landing page information.
- Predictive Data – this is historical data from other advertisers and their “viability”. This is an algorithm that is obviously cloaked in great secrecy but there is a suspicion that the number of advertisers, their history including number of clicks and their length of use are all parts of the equation.
- Opinion and Editorial Determinations – generally, Google is attempting to automate as much of their AdWords operation as possible. However, there is always the option for human intervention. This can work both ways and is not very common but is a part of the equation.
Anyway, that is what I have been learning and I hope it is useful for you as well. I originally started this post two weeks ago but got pulled away. Since that time, I have gotten a couple more 10’s but they are all in the same Ad Group. Haven’t seen any other 10’s showing up but they will probably come. In the meantime, I am continuing to improve things and am starting to see some serious ROI on the original campaign and a modified version that I just pushed out 5 days ago. I am definitely starting to see things moving in the correct direction. More on that later.
I have been looking into why one of my keywords had a low “Quality Score” (QS). In researching, it seems like the primary factors in QS are relevancy and Click-Through-Rate (CTR). So I took that keyword and another one out of the existing two Ad Groups and created a new one with a very relevant ad and created a new landing page that has as much relevency as I could jam into it. Put all that together last night and submitted the ad etal.
The response I got back from AdWords was interesting. First, it was immediately rated at a QS of 7 (out of 10). Interesting because that was the value applied immediately. However, this morning I got up and found an email from Google AdWords saying that my ad had been disapproved. Oh no, what is wrong with this ad and the keywords, they were very generic and I would have thought completely acceptable.
Based on that, I started looking through the ad and keywords to see if I could see some problem and then started reading the Terms and Conditions, Editorial Policy and all of the other governing documents (which are only of interest to lawyers because they don’t give you any useful information about this sort of situation). The results were that I was completely stumped.
My last ditch effort was to see if maybe the link to the landing page was causing the problem. Sure enough, when I clicked on the link from the ad, it goes to “Page Not Found”. Now, I know why they disapproved it. This is not really something that they want to support. Why you ask? Well, from what I understand, at least when you initially submit your ad (and any time you edit the ad or links for keywords), there is a check done that the domain is valid and that it goes to a valid landing page. This is to eliminate situations where you insert an invalid or inappropriate landing page after the ad is accepted.
Anyway, fixed the link and sure enough, it was approved. In fact, it only took about 20 minutes for the approval to come back. This was a great lesson (another one) in making sure that everything is working correctly before submitting an ad, campaign, keyword, etc. At the same time, it also pays to not freak out when you get a disapproval. There could be something simple. At least its not an account shutdown situation which has been happening a lot recently. I am hoping to stay out of this scenario.
Take care.
I have been working on improving my PPC campaigns (DIYSolarPanels and MakeSolarPanels). In looking at the stats for where the clicks were coming from, the majority of them were coming from outside the United States (I had placed the ads in most of the major English-speaking countries). I was getting good CTRs and decent hop ratios but no conversions. So, I said, what could be wrong here.
One of the first thoughts would be that the Landing Pages weren’t working well and they were sending lots of poorly qualified traffic to the sales page. However, when looking at some comparisons, they looked to be in good shape. Plus, one of them was one that has provided good conversions previously. So that was probably not the issue.
So, where do I go next? I thought to myself, is there a potential that the impressions and clicks from outside the US were not providing the potential conversions that I was looking for. From my stats, it was looking like about 60% of my traffic was outside the US and I didn’t really want to lose that traffic but if it wasn’t converting, then I didn’t want to be paying for it either.
So, I cut off everything except the United States and Canada. Now, this is where things get a little strange.
After I made the change, my number of impressions per day has been dropping (down by 1/3 and continuing to drop) but more importantly, the CTR dropped from ~3.2% to 0.22%. It has recovered a little bit today but still not up where it was. So what caused such a drastic drop in CTR?
A little research has led me to a very interesting little bit of analysis. What I have found is that I have a definite correlation between position and CTR. Here are the results that I found over that few day period:
Graph of Average Position versus Click-Through-Rate
So, what this is telling me is that there is a strong correlation between “Average Position” and “Click-Through-Rate” but the sensitivity is a lot higher than I expected. With a change in position from 4.4 to 5.9, the CTR has gone from 5% to 0.2%. That’s huge but a very, very critical piece of information. I am going to be adjusting the numbers so that I get my position higher and see where that takes me.
Good clicks to all!
Okay, so I should know that I am going to learn some lessons along the way but some of them I would like to avoid. I have had a couple of big ones over the last 10 days that I would just as soon as not had.
First, I found out that my landing pages were still sitting in test mode which meant that when folks clicked through to go to the sales pages, they were getting a test feedback rather than getting there. Doesn’t make for great sales that way.
After I got that fixed, everything was cruising along just fine with good CPC, CTRs and decent hops. However, I wasn’t getting any sales. Isn’t that the point of this whole thing? So, I got in touch with my mentor and described the problem. he looked in his ClickBank records and didn’t find any hops from me. Surprise, surprise!!! Why is this.
Well, after going back and forth a couple of times, we figured out that my links were not pointing at the right place. I was going to a sales page that has one of the lowest gravity factors in this niche. Oh well, it only goes to show that you really need to test and be very careful about making sure that the details are correct.
Now that I have everything fixed up, my two Landing Pages (LPs) (DIYSolarPanels and MakeSolarPanels) are working fine and its just a matter of time before we see what the real results should be.
Okay, I am back in the saddle! I had to take a few weeks off in order to deal with some other projects, customers and issues. During that time, I shutdown my campaigns because I knew that I needed to re-write my landing pages and I wasn’t going to get the time. So, now I have the time.
I have re-written my landing page or provide more information and less sales “stuff” and I can see where that goes. I started by turning on one Ad Group (DIYSolarPanels) and determining that the ads were still drawing impressions and clicks and they were. However, I made a big booboo. I have code in my landing page that tracks where the click came from and where they went. In that code, I have a flag that sets the code to either production or test mode. Unfortunately, I had a very sickening feeling yesterday when I found that the code was set to test mode. I know that this flag was set since I turned things back on but I am wondering how long the flag had been set prior to turning it off. This would definitely tell me why I wasn’t getting any hops to the sales page. Oh well, it’s a mistake that would be easy to make and there isn’t anything that I can do about it. Just have to move on but learn the lesson to check things before releasing them for operation.
One other note is that the process of testing things is really slow when you are only committing $10 per day to the operation. At this rate, it takes about 4 days to tell whether or not there is a potential buyer out there. Obviously, if the hop & conversion rates were higher, it wouldn’t be a problem. However, I think I am running in the right area so its a matter of waiting around to see what happens.
One thing that I have learned is that Google plays games with their bids (CPC). I am not sure how their algorithm works but it obviously changes the “Estimated First Page Bid” based on my performance. I have been told before that you can get lower rates as you show performance (and that does work) but I wouldn’t ahve thought that the “Estimated First Page Bid” would change that quickly. Oh well, part of the learning process.
That’s it for now. Going to keep plugging away. I will be upping my budget tomorrow because I am starting a new Ad Group (MakeSolarPanels) and I want to see how that one does. Will let you know.
Comments Off on Finally got a 10! Modified