What I mean by this is if you're using Wordtracker to research keywords and you use a phrase like wholesale fish.. you'll get results of 98,000 sites, with 75 searches per day.... Now this is all fine and dandy if the person putting that phrase into the search engine types it in using quotations like this "wholesale fish". Only sites that have both words of the phrase will show up and there are about 98k sites.... However, if you just type in wholesale fish, without the quotes, you get 408k sites with those words. Now most people I know don't use or even know about using the quotations to get narrower results....so what good are these keyword tools if you're getting numbers that aren't real world scenario's.... It seems that just doing random searches on Google will give you more accurate results of at least how many sites are showing for a particular keyword phrase.. Any thoughts on this?
No one cares huh? Well here's another scenario...when you use google trends and put in a keyword phrase you'll get results with average search volume and average competition...but if you put that some keyword phrase directly into google..you get 10 million results. How can you call 10 million results average competition? And while a lot of those 10 million sites probably are non existent or are not direct competition, getting to the top 20 still seems next to impossible. So the question stands...how can you find a good keyword niche? And what tools are accurate in telling you relevant search volume and competition for those keywords?
WordTracker's (and other tools) way of determining competition has been heavily criticised by many marketers. The 98,000 competing sites you are talking about, as you already know are merely results of that keyword phrase in Google. These pages merely contain that keyword phrase but are not necessarily trying to optimize for it, and as such, not all of these pages can be termed "competing pages". There is one keyword tool I know that more reliably reflects competition by using advanced operators. It is called Keyword Enchanter. It also tries to address the quotations issues by calculation competition for both the keyword string and phrase. I'm afraid I have to disagree with you here, Christian. Those 408k sites you talk about are also affected by the "non-competing sites syndrome" that I mentioned above. You can only get a more accurate measure if you query only those sites with the particular keywords in their title and anchor text in links pointing to them (two of the most important seo factors). This can be easily achieved by using operators or simply using the site I mentioned above. Not really. It all comes down to how well optimized those in the top 20 are.
There is a reason why you do keyword research like (intitle:"wholesale fish"), the number of websites listed for (wholesale fish) includes websites with the term (fish wholesale) or (wholesale fresh fish). I only want to know what websites include ("wholesale fish") since I can easily outrank the websites with (fish wholesale) for the search phrase (wholesale fish). Having the keywords in the correct order improves your SERPs even if the person doesn't use quotation marks.
It's actually much worse than just quotation marks. Some people, for whatever reason, run up the count. Especially super popular niches, and a lot of consumer electronics. Example(From Overture) 7624 Searches - rio carbon 5gb mp3 player 1211 Searches - rio carbon 153 Searches - rio carbon mp3 player That is blatant poisoning Also, for all tools, it's not an estimation of how many searches get done. It's how many recorded. If you're using overture, http://seoblackhat.com/clicks-by-search-rank.html is a pretty good predictor. Of course, doesn't get around poisoned results. For more poisoned results, I seem to recall casino or poker being really good as an example.
Ok now we're getting somewhere...You guys know your stuff, and that's the kind of info I was hoping someone would present. So if I may ask, if it's not out of line, when you're doing research on some keyword phrases you have an interest in; what's your process and determining factors before you find a good domain name and move forward. I mean specifically what's your process as far as doing manual google searches, keyword tools etc.
For whitehat/grayhat sites, I buy a domain that has all my keywords in the url. That's all that matters. I'm not looking for type in traffic. If it's blackhat, it's just a subdomain with the keywords, on a domain that I do other BH stuff on it. For keywords, I have a full recursive overture scanner. I also have a ask/msn suggest scraper(1 level recursive). I have another piece I use when deciding if a niche is even worth it, that calculates the SE saturation vs. # of searches. It's a pretty nice setup
I have written an article called Advanced Keyword Research at my website www.MarketingHub.info, go take a look in an hours time. The website is currently down since DreamHost (PR8) is having downtime.
That's a very good article...thank you for that. Here's another question if you don't mind... If you decided to buy a new domain name for your project, after doing the research, would you specifically be trying to get a domain with your keyword phrase in the domain name? And would you care if it was a .info or .ws site? The reason I ask is I found a 3 word phrase with around 15000 search/month and the intitle competition is about 10,500 sites. There are domain names available in the format of "my-great-phrase.info". Since I'm not trying to brand the site, does it really matter if it's a .com or something else? Example. New Car Prices. 20000 searches a month. competition 72000 www.new-car-prices.com is taken but new-car-prices.net and others are available. Any reason not to use a .net or .info. What do you think of this ratio of searches to competition?
.infos don't look good in the consumer's eyes(kinda), but the search engines don't care. Not sure about .ws. Stay away from .cn's though. I'm sensing a google purge coming.
Strictly for the purposes of trying to rank high in the SERP's and generate traffic...nevermind branding, just making sure the keywords are in the domain name is very important (forgetting about .com.net.org)...you would agree with this?
I don't think it's very important but it's still nice to have. I would avoid hyphenated domains though, hyphenated domains are worthless when you want to sell the website and I doubt it plays a large role in SEO. The extension won't make much difference if you focus solely on SEO, like I tend to do. One of my websites, www.NarutoWallpaper.biz has the .biz extension but is rank 2 for 'Naruto Wallpaper' and rank 3 for 'Naruto Wallpapers', I get over 1000 uniques per day from Google in under 3 months of being live. Are you finding out the direct competition by using intitle:"keyword keyword" or are you using a different method? With intitle:, 'Naruto Wallpapers' has only 2000 competitors but 50,000 searches per month.
I have been preaching this forever! Perfect case in point. NO MATTER where you live you see real estate agents advertising "WHAT" as a domain name for their business? THEIR NAME!! That's just plain dumb! Research shows that over 80% of homebuyers are looking for a home online. Do you think they are looking in the SEs for "JohnQIdiotRealEstate.com" or "Podunk Homes for sale"? The same goes for online marketing. Another good point made is that of not trying to eat the elephant in one bite, in other words choose smaller niches. I have sites for niches with VERY high type in traffic that I was able to get top rankings in Google with virtually no promotion JUST due to the keywords in the domain. If you want to brand a site that is all well and good I have those too, BUT in the meantime choose domains with the keywords you want to target while you are working on it. Gregg
I'm impressed. Type-in traffic is hard to come by nowadays, unless you buy an already-in-existance domain.
Thanks for the tip on the hyphenated domains... Live-CMS; I was using the intitle method to find the competition and keyword discovery for search volume...like from your article. What do you guys consider too high of competition to have a reasonable shot at raking in the top ten on Google? I've found a few keywords that I like, with 40k searches, but the competition is in the 90k range. Would you avoid that type of competition?
90k in title? yeah, that'd be rather difficult. It all depends on your skill level though. Maybe try a subsection of that. At least try and rank for that, even if your topic is more broad. Remember, for every 40k searches there are for a keyword, there's probably another 100k+ of people adding in random words AFTER the keyword.
Ok, I'll probably stay away from the 90k's then. What do you think about really small niches that have a maybe a thousand searches per month with under a thousand competition...it shouldn't be difficult to rank, but do you feel it's worth the effort for a thousand searches or so? By the way, thanks for putting up with all my questions...I'm really trying to learn some of the tricks to this and you and the other guys have been a huge help.
I can't comment on this too much, but I will tell you one of my niches, which I think beginners could have some pretty decent luck with. Well, I take that back. I'll tell you the stats of the niche. PRODUCT in caps means the product name, in this example. All it uses is Google and overture (http://inventory.overture.com/d/searchinventory/suggestion/) Google Hits: buy PRODUCT = 720,000 Google Hits: "buy PRODUCT" = 37,400 Overture Hits: buy PRODUCT=1,721 Going for strictly long tail results(it's a scraper site), and not ranking worth a crap(non-top 20), $15-20 per day. On occasion, it makes up to $50/day, and some days it makes $2/day. I redirect to an affiliate page(yeah blackhat), which converts 1:125, and there's maybe 3 sales a day, so that equals out to...375ish hits+ per day. Not bad. And if I were to spread it out from there, it would do more. I really really do not believe in the intitle: method of evalutation niches. Maybe if google only searched titles. You can use my results to scale, or reply with your own information formatted like I did above, and I'd be glad to help.