This post was made for interns exactly 1 month ago, but I’ve decided to open it up to the public. The instructions regarding the BCMI tool obviously don’t apply, but you should understand that concept from previous videos. For now it’s pretty safe to say that keyword research is slightly more keyword agnostic then it used to be, so using the PA and the SEOMoz Keyword Difficulty Tool is not a bad idea. Also this is based on expired domains but it could just as well be for new domains (you just don’t have to worry about matching niches).
Before you read this post, I suggest you take a quick refresher over these 3 posts:
Keyword Research Part 1
Keyword Research Parts 2 and 3
How my Competition Tools Judges Competition
Ok now that you’ve refreshed your memory, login to your account on the tool. This week you will see a couple new links in the left menu – CMI and BCMI.
CMI (Competition Metric Index)
This tool tests a single query and shows you the APA of the top 10 (scroll to the bottom). It’s good to use for a quick single check. I also recommend you double check some of the BCMI results here if they look too good to be true. It does suffer a lot from proxy issues (below), so you may have to click Go! A few times before you get a result.
BCMI (Bulk Competition Metric Index)
This is the tool you will be using a lot. You are limited to 3 tasks a day, and 800 rows per task. You upload a single column csv with a keyword (no brackets or other chars), and it will process the list and spit out the results in another spreadsheet (similar to the DA Checker). This may take awhile because of the same proxy issue, and occasionally you may see an error in the results. If so, just put the errors into your next task.
A quick note about Google and Proxies
Google hates it when proxies scrape it. I have 100+ private proxies being rotated, a sophisticated quarantine system and de-captcha system, but I’m still having trouble. The issue is that Google bans entire subnets of proxies from proxy providers, so even if you are throttling things well, Google will ban you because someone else on your proxy service is abusing them.
How to find keywords for your domains
I highly suggest you find keywords that are loosely related to the former topic of your site. The easiest way to know the topic is to check the anchor text of the domains linking to the site. Sometimes you won’t be able to get specific anchors, so what you could do is do a related:domain.com search on some of the domains linking to it.
I say loosely related and I mean it. If your expired domain is about a church in some city, go ahead and tackle anything Christian or Catholic, or anything related to that city (and just give it a religious edge). If your site was formerly the domain of a mayoral candidate in California and has lots of links from California newspapers, you could target something related to california (perhaps jobs related). If your site is about mushrooms, make it about healthy snacks or recipes. Don’t get hung up on your keyword being exactly related to the former topic of your expired domain, but also don’t stretch too far the other way and turn your mushroom site into an auto insurance site.
So once you have a few different niches you think your site could accommodate, simply plug everything into GAKT (try both the Keyword and the Ad Group Ideas). Export a ton of results, combine them into a single spreadsheet and run the macro that turns the sheet into a calculator. Focus on the domains that are $20/mo+. Take up to 800 of those and shove them through the BCMI tool and you will have completed your first pass of targeting. Grab the domains that have all 10 results with an APA <60, the majority of which are <50 to look at closer for pass 2.
Your 2nd pass of targeting will be to use SEOQuake to check the # of pages indexed in the SERPs that passed round 1. Just like I explained here. Only target keywords for domains that have a site in the top 10 with <30 pages indexed, or a generic articles site in the top 10.
You may have some trouble finding keywords, and if you are, I’d like to encourage you to try this option. Do everything exactly the same way, but instead of basing it on exact match in GAKT, based it on phrase match. Try to find domains that will make >$40/mo on phrase match, but on exact match have an LMS <600. To do this you should check both exact and phase match in GAKT when you search so you can make the comparison. Your CMI results and your contextual CPC amount should be based on the root exact match search (the one that has an LMS <600). The idea is to earn on the longer tail, which should be very simple to rank on, since no-one is targeting even the short tail version of the keyword due to low LMS numbers.