Anyone want to test out my keyword research theory?

[eatyourveggies]

Hi all,

I'm relatively new here but not new to the world of SEO / SEM. Over the years I've loved using SEOmoz and other tools but of course have found certain limitations with respect to how I like to work. That's the case with any tool / service.

So over the years I've put together a keyword research / competitor analysis process that has worked well for me and I'm wondering if it might also work for others.

I've spent the last 15 years of my life as a director of a range of companies, mainly in printing but also in systems development, marketing, etc. I spent a large percentage of my time developing systems and tools to help me with my search engine marketing. I've now sold all my companies and I'm semi-retired, somewhat bored, and would love it if I can assist others with the process I've used over the years.

I'm curious to know whether SEM professionals agree with the way my system ranks search terms from "best" to "worst". If you're interested in testing this process and telling me if you think the resulting list of search terms that I come up with for your website is "spot on", "not bad" or "horrible!", then please read on.

My key motivation here is to educate myself as well as others. I'm not charging for any of this...

If you give me your website URL, your top 5 competitors and your top 5 search terms, I will return to you:

• a complete list of search terms including "niche" and longtail search terms

you can then...

• really easily filter out irrelevant search terms, thus creating a list of negative keywords, ready to import into your Adwords campaign.
• easily group your search terms in "education" and "purchase" search terms so I can analyse these two groups of keywords separately
• see which words are used most often across all the search terms so you can easily create keyword specific Ad Groups in your Adwords campaign.

You can tell me which options you prefer:

• broad match, "phrase match" or [exact match] when getting search engine results
• specify any country you want the SERP results for, or even any city.

I will then:

• do all the keyword research, getting the latest (live) Google SERP results
• combine all competitor metrics (page rank, domain age, juice links, etc) and search term information (search volume, CPC, search term length, etc) together to give you a list of search terms ranked from "best" to "worst"
• do the same for both organic search terms and paid (Adwords) search terms
• do the same for both direct search terms and niche search terms
• take into account "word count" (number of words in a search term) as longtail keywords generally higher-converting search terms

And you can adjust things to change how the keywords are ranked:

• specify "thresholds". eg; you can ignore the really competitive search terms, or ignore the really short search terms
• specify "weightings". eg; you can put a greater emphasis on search term length, or a lesser emphasis on cost per click

As a result of the keyword analysis, it'll also show you who your organic and adwords competitors are based on all keywords, or just your top ranked ("best") keywords.

In that competitor data you can see:

• which paid ads appear at the top, side AND bottom of the results page
• which organic results are shopping, image, video, and local results
• all the metrics for each competitor (page rank, domain age, juice links, etc)

All of this is is handled in a simple web interface that I threw together recently. It's really simple, merely asking for your site and preferences and then an interface to view / sort the results.

Interested?
I'd like to hear from any SEM professionals who want to test this process.

Once I have your basic details, I can get a keyword list together simply (using my internal process / software) and then you need to do some basic sorting, particularly if your search terms are in an industry that I know nothing about. Your input will be required.

From there, give me 24-48 hours and I'll return 2 lists of search terms: "organic", and "Adwords". I'd love to hear your opinion about the relevance of the search term lists. I hope it will also spark some interesting discussion and hopefully help people learn a bit more about keyword / competitor research.

If you're interested, please shoot me a private message letting me know why you'd be a good candidate to test this system. I really do want people who are well versed in search engine marketing. So please include a basic "resume" about who you are. If you have an SEM company and that's your main career focus then I definitely want to hear from you.

Adam

[eatyourveggies]

Well Spyfu is awesome, no doubt about that.. except..

I'm in Australia. Spyfu only covers the US and UK.There are quite a few tools that simply don't cater for Australia.

My software caters for nearly 30 countries, and most major cities in those countries.

Sure, SERP results will change on a city-by-city basis but that's a problem any tool is going to have. If the tool uses the Google API then the SERP results from the API are different (sometimes vastly) from an actual browser searching on Google.

I'm not sure how to overcome the geolocation issue entirely. I don't think it's possible.

But otherwise, what I find frustrating with all these tools is they give a bunch of data but don't give any answers. So spyfu shows me historical CPC which is fantastic but it means I'm left to try and use that data to answer my question which is really simple: "which keywords are the most profitable based on ALL data available?"

The Kombat function in Spyfu does what the SEMRush "comparison" tool does - shows me which competitors share keywords. For a seasoned SEM specialist this is great because he / she can interpret all that data and make customized suggestions to their clients. But again, for me... I ran a business. I just needed an answer.

It's about having something that summarizes all the data and presents options to me.

Maybe I'm over-simplifying it, but what I wanted was to be able to give my system my website and 5 keywords, it would return 1,000 keywords that might be suitable, and then it'd automatically collect every key metric for every competitor across all 1,000 keywords and then rank those keywords from best to worst based on all the back-end mathematics.

I created an infographic to help explain the process a bit.

keyword_research_process.jpg

[Jinx14678]

Hi,

Not to be a stick in the mud, but it kind of sounds like spyfu

but with that said good ideas in ranking most important to non, the issues I think you would run into is the geolocation change in serps (someone in dallas, might not get the same results as someone in new york for certain queries.)

Shane

[eatyourveggies]

Hi Sean,

It's software I've written that collects all the metrics... domain age, page rank, home page rank, backlinks, CPC, etc, etc.

I would actually collect domain and page authority as well and for a while even the "keyword difficulty" metric from SEOmoz as I trust their metrics a lot and it's good to see if my results are similar to theirs.

As mentioned to David above, it could all be automated but at the moment there's a bunch of manual tinkering I do after my software collects all the data.

At the same time, full automation wouldn't be possible as the client needs to be involved in filtering out the irrelevant keywords, picking which keywords are "education" or "purchase" keywords, ignoring keywords of less than x searches per month, etc.

However I have developed nice little online modules for these steps that help do this quickly. So you can group the keywords, and filter by various methods, and make mass-changes (like deleting groups of keywords) with one click.

I think if those steps are made really simple and quick then people will be happy to be part of the process. The result will be a more relevant list of search terms in the final ranked list.

If you already have a website for which you're confident you've got an awesome keyword list for, I'd love to know if my system yields similar results.

I hope, if it doesn't yield similar results then people would be interested in breaking down why, I make modifications, and so on. It would be fun (but then again, I have few friends!)

Adam

[eatyourveggies]

Hi David,

It's software I've written. But it's an application that runs on PC / Mac.

The client needs to be involved though. So I need to be given the basics, then my software compiles all the keywords, gets monthly search volumes, and then the client would need to get rid of the irrelevant ones and then I get my software to do all the research, gathering absolutely everything.

It's a massive amount of data normalizations and calculations with the result being the most profitable keywords ranked from 100% to 0%.

The client can get involved again and tinker with some settings to get the list of ranked keywords tailored to their own preferences.

For the parts where the client would need to be involved (initial info, clean up the keyword list, tinker with the results), I've programmed those steps in an online interface so I can have people test this process.

I could have a sign up form for beta testers but really anyone who's an SEM professional who already has a list of keywords for a particular website is someone I'm interested in talking to. I want them to compare what my system / process outputs to what they already have.

So it doesn't really require an online form... just an expression of interest with a 1 minute bio and I'll pick a few people. At the moment I have to trigger each stage of my software to do it's thing, clean up some of the results, etc - it's not fully automated as I never had any intention of making it publicly available. So I want to make sure I'm spending my time doing all this for someone who is REALLY interested in it and can provide some quality feedback.

If you're keen, let me know

Adam

[ske11]

Hi Adam

That is a detailed description, I would also be interested in learning a little more about the tools you are using as there would seem to be several applications working together.

How much of the process involves a person rather than automation?

I like the fact that you plan to split the terms in to education / research and buying terms, this is a valuable recourse for companies.

[David_ODonnell]

Hi Adam,

Sounds interesting, I'm curios to know what tools you're using to compile your data?

Also, you mentioned a web interface, have you considered opening the tool up to limited beta access with user access / sign-up form? (instead of requiring people to pm why they'd be a good candidate / message you back and forth etc.)