Keyword Research & Optimisation Techniques

Here’ some quotes from a thread discussion keyword research over at SEOers.org that I thought people might find interesting. Ill be posting more about keyword research in the coming weeks.

The thread is focusing on how I go about deciding levels of competition for different keywords and which ones I target first:

When finding out levels of competition i use the allinurl:”keyword”, allinchor:”keyword” and allintitle:”keyword” searches. Doing all these will give you a better idea of what you’re up against. I don’t usually try and calculate the amount of traffic a keyword is going to provide and just go for the ones that i know are going to bring relevant visitors. Remember quality of visitors is the most important factor to think about.

www.seoers.org/BB/general-search-engine-optimisation-discussion/keyword-analysis-best-method/msg57663/#msg57663

Here’ some info on how I come up with the original set of keywords:

I use suggestions from the client, wordtracker, googles keyword tool, observations of the serps and personal intuition to come up with a list which may contain hundreds of keywords. But every keyword on the list will be relevant and a potential converter. I analyse the numbers from wordtracker and use the various google allin:”keyword” searches to get an idea of competition. The number of returned pages for standard google searches means very little. A low amount of returned pages can be just as competitive as a high number, it all depends on the quality of the top sites. I lay all this down on a spreadsheet and once all the data is in place its easy to figure out the easy keywords and the ones that will require more work.

Just an after thought… Once the above part is done Ill match up the keyword(s) with the urls of the site I’m working on, with the tougher keyword(s) assigned to the more prominent pages – golden keyword(s) on the home page! If its a new build then these keywords are implemented in the site from the start. If it’s an existing site then you have to do several things… figure out if they are already ranking somewhere for the keyword and what page (google often gets it wrong so you may have sort that out), do what you can to improve it and so on…

http://www.seoers.org/BB/general-search-engine-optimisation-discussion/keyword-analysis-best-method/msg57909/#msg57909

But one of the most important factors in figuring out which keywords will work is experience, this could be experience in anything from whether client’s payment warrants the time required to achieve what they want or if you already have a good idea what a site could be capable of. There are many factors which just become second nature and are born from an understand what you a, a client and a site can do:

I don’t throw any keyword away due to the levels of competition or the research I’ve done on the SERPs. All of my clients have “golden keywords” which are bound to have high levels of competition. They get moved to the top of my keyword list and become long term strategies. The ones where my allin searches have revealed weaknesses become the short term goals. If you find there are no pages returned for “allinurl” searches then you usually find there are low levels of competition for allintitle and allinanchor. The three figures that you get from that research become a great indication of how well all the competing sites have been optimised.

So my competition research never results in me dropping keywords, just which ones i target first. It’s worth noting that how long it will take to achieve top rankings for keywords should be discussed with the client from the start. If your client isn’t paying for a lot of your time then it’s going to take longer to accomplish positions for their bigger keywords. If a client is niche then you maybe able to achieve what they want with a low amount of time.

The returned pages amount gives you some idea but i quite often get in situations where a site will appear in the top 10 for a search which results in 10s of millions of returned pages but not for 10s of thousands of returned pages for another search. Most searches pull up noisy pages which are loosely related to the keyword and therefore easy to beat.

KEI and all the research in the world doesn’t mean anything until you use the force… (intuition based on SEO experience and instincts).

http://www.seoers.org/BB/general-search-engine-optimisation-discussion/keyword-analysis-best-method/msg57933/#msg57933

This is where me and boogaloodude start discussing what the only really way to figure out the level of competition facing each keyword and what do you actually do with that list of keywords when you’ve done all the research:

The only way to fully understand the levels of competition you’re facing is to check the back links and seo work for all the top returned pages (the deeper you go the more you know!). But what were talking about are trends and reading between the lines of those trends.

When i do the “allin:” searches I’m not bothered about whether they match the standard searches. For example.. if we were doing research for the keyword “online dating” we’d expect to find high numbers for the “allin:” searches and we’d also expect those numbers to drop as we research the longer tailed equivalents. It’s these drops which become the metric to judge the popularity and therefore competition.

If I end up with a list of 800+ keywords then most of the time they can be grouped; whether this be by location or product/service etc. I’ll work through the keyword list and assign each keyword (or it’s group) to a URL/page; then ill optimise that page (and how it is linked) for those keyword(s). If I’m left with keywords which aren’t assignable then ill leave them in the bank and try and add a page for them at a later date.

http://www.seoers.org/BB/general-search-engine-optimisation-discussion/keyword-analysis-best-method/msg58102/#msg58102

With a lot of my clients there are keywords which they want to rank for no matter what so the level of competition is something that Ill make them aware of and from there I’ll tell them how long and how much work I think it will take to accomplish top rankings.

But the client is interested in their big keywords and the idea of long tail never usually occurs to them. So I’ll target the ones with lowest levels of allin searches first (usually the long tail) and work up the list. As the site improves (links etc) the more competitive keywords can be targeted.

I figure that the allinanchor is a good indication of how many links I’m going to need, the allintitle is a good idea of how many pages I’m actually competing against (there will be pages that rank well without a direct title tag match but the title tag has so much seo weight that a direct match will usually beat most of these) and the allinurl gives me a good idea of how many pages are optimised well for the keyword, well its a clue anyway!

All my allin searches have “speech marks” by the way and I find certain websites will keep appearing so Ill check these out as well and check what we said about backlinks etc. And i always you the force!

http://www.seoers.org/BB/general-search-engine-optimisation-discussion/keyword-analysis-best-method/msg58153/#msg58153
So there’s a little insight into how i do my keyword research. How do you do it? Please feel free to check out the whole thread for some great posts by other SEOers users.

(my bolding)