My First SEO toys and PPC tools

Now that Analytics is on, I still have not had a chance to really go through it. The documentation front seem a bit thin compared with AdWords. This could means that Analytics is the new wild west or I am just looking at the wrong places. Surely, Google won't make software that is hard to use...

I suppose I have been procrastinating on Analytics. But, I have a good excuse. We just brought the SEO pages live. I purchase/licensed the SEO content for a list of 30 keywords that I want to get good ranking on. It turns out that one keyword I cannot use after getting a better understanding of the system. The key reason is legal, it gets too complicated to do SEO on a competitor's name. So, if you have come up with a good way to do SEO on a legally registered (competitor) name, please let me know!

The structure of the SEO pages are fairly plain. So, I spent some time with the web people to spruce it up so that at least the touch and feel is similar to the overall corporate site. Of course, this process can be taken too far. So, when my VP of engineer told me that it is time to push them out to make the licensed content earn their keeps, I readily agree.

The funny thing about SEO pages is that it could impact the performance of the AdWords program. Based on historic data, when people search on the company name, naturally, we rank highly in both the natural and paid search results. And, there are very few click on the paid search results.

You may think that this is a good thing, why pay when you get the click through natural search. the trouble is that the paid search is optimized to create conversion events. In other words, when you click on the paid search, the route is designed that you either abandon the path or leave your contact info. This is how we generate the bulk of leads.

Now that SEO pages holds the promise of getting really good natural search ranking for the keywords that have been generating the bulk of conversion traffic, this could be a problem when people no longer leave their contact info with us.

Theoretically, I would love to do an empirical A/B test to see if the concern is valid. But realistically, the stake is too high that I have no intention to taking the hit. So, I have been working on these pages to create as many conversion traps as possible.

**
So, here are a few good tools to have

1. To figure out exactly what a viewer would see when doing a google search

http://merjis.com/local_google_search

Why bother: say that you want to see what a viewer in Germany searching for "tiny boots" using a German interface would see. This tool automatically generate the appropriate tags. A very helpful tool if you do regional/local advertising from afar.

2. US location keywords generator

http://5minutesite.com/local_keywords.php

Why bother: you would be surprised the power of regional/targeted ads - at least that I have been told many times over. I have never personally tried it but I am looking for an angle to do it with a regional partner - that is a separate story though. Anyway, the interface is reasonably intuitive. It does not capture ALL permeation in terms of location keywords, but it wold give you a good start. And, that is all that we can ask for in life, a good start.

3. Google match option wrapper

http://www.mikes-marketing-tools.com/cgi-bin/adwrapper.cgi

Why bother: well, if you have not heard my rant, let me just say that it is a worthy effort to include all three match options for every keywords - broad, phrase, and exact. The problem with the level of dedication is that it is painful to do manually. I constructed a macro using Word for the job, but that was still painful. Then, this site comes along and all you need to do is to plug in the broad match keywords and, voila, out comes all three permutations of the match options.

The truth is that I still manually add the match options on a day to day basis when it is just one or two keywords to add since it is a bit of hassle to open up another browser. But, for a mass job, this is definitely the ticket.

4. Statistically predictor on performance

http://splittester.com/index.php

Why bother: well, I think the original premise of the tool is to help you predict how your ad will perform relative to a benchmark ad with minimum data. Think of it as a way to predict the result of the A/B test. I actually have no idea what kind of statistical model they use in the backend. But, I think it is an interesting idea and until I am willing to sit down and figure out my own statistical prediction model, this is a good standby tool.

On a related front, I think the model can be expanded into additional usage than just comparing ad's. For example, why can't I use the same model to compare the number of click through vs conversion rate? I think these would be much more interesting.

5. Where do you rank on Google

http://www.nichebot.com

Why bother: Technically, this site offers mostly keywords related services. Which, to be honest, I have not really looked into. The last option on Google ranking, however, is a great tool. This is especially important if, like me, you are embarking on an SEO project and need to have a third party benchmark to see how things are coming along. My favorite part is that you just dump in all the keywords that you want to have ranked and just let it run in the background. It is not the fasting service. But, for the data provided, I do not mind the wait.


chiefchickenheadless (at) gmail (dot) com sign out
Legend: (at) = "@" (dot) = "."

Passing keywords, amongst others

Yeah, yeah. I know. I had taken a bit of time off. No good excuse. But, then, it is not like this blog is a life and death issue either. Hopefully.

All is well. Life is still crazy in the Silicon Valley startup land. The good news is that our performance for the past year has been so good that we are planning on starting a fund raising process. And, almost to validate our performance, some of the current investors are already jockeying for position with pre-term investment to help us ramp. Well, we will see how it works out.

Google is giving me a bit of heart burn lately. In particular, it has become unruly to manage and I am finding ever more lack of features. For example, I have begun the process of cherry picking the top performing keywords to their own campaigns with higher budget and dynamic keywords to make it perform even better.

That is the theory. But, it has not worked out so far. I am definitely paying more since the bid prices are many times higher than the standard campaign. But, I am not seeing better conversion.

Anyway, it is partly my fault, I am sure.

Oh, the Google Analytics account is finally turned on. I am glad. Of course, I still need to hook it up with Google AdWords to get the full details. I just logged into the Analytics account the first time about one hour ago and I am sure that it will take quite a few days for me to get up to speed and figure out what I am looking at.

This should be a good challenge.

**
So, I have begun to apply more pressure on the PR firm. They seem competent enough and there are certainly enough carrots that they have been dangling in front of me.

I have been pretty "easy" on them in the past month or so since it was the Thanksgiving/Christmas season. For one, I was not that motivated to work too hard. The other thing is that the PR firm has been saying the the reporters and calendars were closed for the year. May or may not be true but I was not pushing them too hard.

Now, it is a new year and I have made it fairly clear that I want to see results. They have had their play time already and it is time to earn their keeps.

It would be interesting to see how much can they deliver. The verdict is out but I want to give them Q1 to really perform and if Q1 is no good, then Q2 would be their chance for redemption.

Will keep you posted.

**
Here are some other cool undocumented Google tags. With them, I can tell the following thing when a viewer clicks on my adWord ad:

1. If the click comes from Search vs Content network.
2. What is the keyword match (if search network) or the site (if content network)
3. Which ad is triggered.

{ifsearch:keyword}{ifcontent:site}={keyword}{placement}&advert={creative}

Taken in their individual components:

{ifsearch:keyword}

If the click is from search network, pass the text "keyword"

{ifcontent:site}

If the click is from content network, pass the text "site"

{keyword}

Pass the actual keyword match.

*This is important* the parameter passes the matched keywords instead of actual key words. For example, with broad match, a search for "big shoes" can match the keyword "large shoes". In this instance, the passed keyword is the matched "large shoes" instead of actual "big shoes."

One way to address this issue is to scrub it against the web log to find the delta and add to the keywords list.

{placement}

This gives the position of the ad if the ad appears in the Content network site (I think, but I stopped using Content network a while back so did not bother to find out the details.)

{creative}

This passes a numeric string which is the ID for the specific ad. Using the ID number, you can match it with the actual ad under AdWords reports. (A rather clumsy process for users who want to optimize the content, but that is another issue.)

**
I am not much of a decorator and I was pretty resistant to the idea of doing office decoration. Mainly because I am too lazy to get involved. But, what must happen, must.

Everyone was assigned a task in terms of to-do and to-get. Then, we had a pretty miserable time in agreeing on a common decor vision. To be politically correct, it was an issue of styles and tastes. But, of course, *I* have the better taste and *fortunately* my style prevailed.

On the other hand, prevailing on that front does incur the issue that I had to manage the execution which was the whole reason that I was not keen on the idea to begin with. Not exactly a "winner's curse", but it felt like it.

So, the office is now decorated and I am curious to see how long we will keep it up. Good thing that the new plants are plastic. (Yes, I am cynical.)


chiefchickenheadless (at) gmail (dot) com sign out
Legend: (at) = "@" (dot) = "."