The Secret To Truly Accurate Keyword Research

by Adam Kreitman

Last updated on June 19th, 2018

Keyword research can be a pain in the butt challenge.

There’s no getting around it, however. Keywords are the foundation of your search marketing strategy. Miss the mark on your keywords and your results suffer.

There’s no shortage of keyword research tools to help you come up with keyword ideas…from the free Google Keyword Tool…to Market Samurai…to Wordtracker…to “spy” tools like iSpionage that give you a good idea of what keywords your competitors are bidding on/ranking for.

But these tools aren’t perfect. There are 3 main issues I encounter when using them:

  1. The traffic data they provide can be wildly inaccurate. I once saw an example where the estimates for a keyword in the Google Keyword Tool was off from reality by a factor of 5000 (that’s not off by 5000 searches per month, that’s off by 5000 times the estimated searches per month given in the tool!).
  2. The Google Keyword Tool seems to be “hiding” some very highly relevant and high traffic keywords in its results. So, unless you know exactly what keywords you’re looking for, you may miss a lot of good options that are available to you (there’s a great article about this on SEOMoz by Rand Fishkin).
  3. These tools can’t provide information about a very key piece of the puzzle…conversions. Conversions obviously depend mainly on the strength of the landing page the traffic’s going to, but the intent and mindset of the searcher matters too. Even very close variations of a word may results in very different conversion rates. And that’s information you can’t get from any keyword tool.

These shortcomings bring us to my favorite keyword research tool…Google AdWords.

Let’s look at a few ways to use AdWords for keyword research that addresses the issues above.

Getting accurate traffic data

As mentioned above, the traffic estimates from keyword research tools can be wildly inaccurate. I’ve found this to be especially true when it comes to local keywords.

Often, the keyword tools will tell you no one is typing in a specific keyword + city variation (ie. plumber Arlington) when there are actually hundreds of searches for that keyword (see a case study about this here).

There are a couple of ways to get accurate traffic data using AdWords.

Use Exact match keywords

Exact match means that, in order for you ads to appear, someone needs to type in a search query that exactly matches the keyword in your campaign. In AdWords, you designate Exact match keywords by putting brackets [ ] around the keyword.

If you have a specific keyword (or keywords) that you’re interested in getting traffic data for, use Exact match. In this case you’re simply adding the Exact match keyword to the campaign and waiting to see how many impressions it gets over the course of a few days, weeks or months.

Use Broad Match Modifier

With Broad Match Modifier (BMM) you add a plus sign (+) in front of each keyword. That tells Google that, in order for your ad to appear, the keywords that have the plus sign in front of them MUST appear in the search query.

Those keywords, however, can appear in any order in the search query and can include “misspellings, singular/plural forms, abbreviations and acronyms, and stemming”. The words are just required to be in the search query in some form.

I use BMM a lot for local keyword research. Sure, we know a lot of people are searching for a “big city” keyword “Chicago plumber”, but how many are searching for suburb related keywords like “Naperville plumber” or “Schaumburg plumber” or plumbers in any of the other Chicago suburbs?

To find out, I’ll add BMM keywords like this to the campaign:

+Naperville +plumber
+Schaumburg +plumber
+{suburb name} +plumber

Usually when I do this, I’ll pair a bunch of suburbs with a keyword and add them to the campaign to see which ones get a significant amount of traffic. This gives me a really good idea of which suburbs are getting the most searches and may be worth focusing on more in AdWords and SEO.

Pay attention to “Impression share”

Whether using Exact match or BMM, I’d recommend putting each keyword you’re testing in its own Ad group. This isolates the keyword and lets you see what’s known as “Impression share.”

Impression share is the number of impressions your ad received divided by the estimated number of impressions it was eligible to receive.

Impression share is only shown at the Campaign level or Ad group level…NOT the keyword level. But by putting each keyword in its own Ad group, you effectively get the impression share for that 1 keyword.

Here’s an example of how this worked for a test campaign I set up:

Google Adwords Impression Share

The Ad group only had 1 keyword in it and, as you can see, got a 62.98% impression share. You can also see that the Ad group had 553 total impressions (this is over a 30 day period).

So, doing a bit of math, I can tell that the keyword in this Ad group was searched for 878 times during the month.

Write bad ads.  Yes, you heard me correctly

When you are only interested in the traffic data for a keyword, write bad ads! It’s not the clicks you’re interested in, it’s the impressions. If you write bland, boring ads people will be less likely to click on them which will keep your costs down.

Also, keep seasonality in mind. If you’re gathering traffic data on the keyword “red roses” you’re going to get MUCH different results in February than you would in August.

How to find “hidden” keyword variations

There are two match types that are particularly helpful when you want to uncover keyword variations that may be hard to find in the keyword tools.

The first is Broad match

For Broad match “…your ad may show if a search term contains your keyword terms in any order, and possibly along with other terms. Your ads can also show for singular or plural forms, synonyms, stemmings (such as floor and flooring), related searches, and other relevant variations.”

Usually I don’t use Broad match in AdWords campaigns because you have to be very careful with “related searches and other relevant variations.” Often, Google’s idea of related and relevant can be quite different from your own!

However, for keyword research, Broad match keywords can be very helpful because you’ll be able to see what keyword variations Google is displaying your ads for. This can generate some great ideas you may have missed by using the keyword research tools alone.

When doing this, be sure to keep a close eye on the campaign and add negative keywords as you go to cut down the amount of budget you’re spending on worthless clicks. (A negative keyword prevents your ads from showing up for any search query that contains your negative keywords.)

The second is Broad Match Modifier

Broad Match Modifier is also a good match types for finding good keyword variations. While, as mentioned above, I use it a lot for local keywords, it can be used to find variations of non-local keywords as well.

Let’s say you sell saltwater fish tanks and want to discover some popular variations that people may be typing into Google.

The BMM keyword ‘+saltwater +fish + tank’ may turn up alternatives like:

saltwater fish tank kits
saltwater fish tank fish
marine fish tank (close variant)
set up a saltwater fish tank
and others.

Whether using Broad match or BMM, you can see the actual search queries people typed in by using the Search Query report in AdWords. You can run that report off of the Keyword tab in AdWords by selecting “Keyword details” as shown in the screenshot below.

And when mining for good keyword variations, you actually want people to click on your ads (unlike when we were just interested in traffic data) because you only get search query data when people click on your ads.  Google doesn’t share search query data for impressions, only clicks.

How to get conversion data

Since conversion is mostly about your landing page, there’s no way a keyword research tool can give you estimates on the conversion rates for keywords.

However, different keywords convert differently. And if the singular vs the plural version of a keyword makes a big difference in the likelihood of a prospect to buy, wouldn’t that be nice to know before you embark on a potentially long, costly SEO campaign?

Testing conversions in AdWords is pretty straightforward. Simply add the keyword(s) you want to test, (using Exact match) and send the traffic to your landing page(s). Make sure you have conversion tracking in place and measure the conversion rates to see which (if any) keyword produces best conversion rate.

A Few Key Considerations

Google recently made some changes to the way they handle Exact match (and Phrase match) keywords that are important to note. The default setting in a campaign is that your ads for Exact match keywords may be triggered if someone types in a search query that is a misspelling, plural version or “close variant” of your keyword.

When using AdWords for keyword research, however, I’d highly recommend changing your campaign settings so that doesn’t happen. You want to precisely control what search queries trigger your ads. So be sure to select the “”Do not include close variants” option as shown below.

Also, set up a separate “Test” campaign for all your keyword research testing in AdWords. The overall performance of your campaign can be negatively affected by poor performing keywords and Ad groups (which is likely to be the case with these research campaigns). You don’t want that to happen so it’s best to keep your tests in their own campaign.

Choosing the right keywords can make or break your search marketing success. While keyword research tools are very helpful, they have their limitations.

By adding Google AdWords to the mix, you can uncover the real story about potential keywords and base your campaigns around the most accurate keyword research data available.



Get updates on new articles, webinars and other opportunities:

Adam Kreitman

Adam Kreitman coaches business owners on how to make their websites more compelling to their prospects.. and to Google. He owns Words That Click, a firm specializing in Conversion Optimization and managing Google AdWords campaigns for small businesses.Follow him on


Comment Policy

Please join the conversation! We like long and thoughtful communication.
Abrupt comments and gibberish will not be approved. Please, only use your real name, not your business name or keywords. We rarely allow links in your comment.
Finally, please use your favorite personal social media profile for the website field.


Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

  1. Andrew Dunoon says:
    April 8, 2014 at 9:12 pm

    Great post, thank you. If I have a daily budget and reach my limit, will this mean that the number of impressions will stop recording showing less then there actually was?

    • April 9, 2014 at 9:23 am

      Hi Andrew-

      Even if you reach your daily budget limit, AdWords will report all the impressions your ads actually recieved. However, there may be a lot more impressions you could have received if you had a higher budget. To figure out how many total impressions are available for a keyword, you can look at the Impression Share data.

      For example, if you got 5,000 impressions for a keyword and Google shows an Impression Share of 50%, you’d know there are actually 10,000 Impressions available for that keyword.

      Hope that makes sense!

  2. November 10, 2013 at 12:31 am

    I totally agree with you Adam. Keyword research is such a pain in the ass especially if you don’t know where to start. Thanks for your guidance though.

  3. Todd says:
    October 28, 2012 at 12:10 pm

    Great article here, Adam! I’ve been using those same tools you mentioned in the beginning and am looking for other research ideas to add to the repertoire. Thanks for the tips!

  4. October 26, 2012 at 6:01 pm

    This post took me by surprise. Write “bad ads,” but really write striking ads that just beg to be clicked. Interesting experiment.

  5. Dan says:
    October 26, 2012 at 6:01 am

    Thanks for sharing!

    I tried creating a bland ad once for traffic purposes, it backfired and gave me a much higher CTR than ad i created for click throughs. Any ideas on how to create even blander ads?

    I was also not aware of impression share so thanks for opening my eyes to that!

    • October 26, 2012 at 10:33 am

      @Dan – I’ve shared my secrets for writing compelling ads here before though I’m not sure I’m willing to share my secrets for writing crummy ones! 😉

      Seriously, what I’ll do is just write plain, generic ads that really don’t say anything interesting, contain no benefits, and don’t have a call-to-action. Look at the competition’s ads and just try to blend in and be even more boring than their boring ads.


      • Dan says:
        October 29, 2012 at 3:41 am

        Thanks for the reply Adam, i shall try making it even more boring than the last time and hope that it doesn’t back fire again!

  6. October 23, 2012 at 11:40 pm

    Great post! Some very creative thinking here with the use of bad ads for the sole purpose of keyword discovery. I also was also not aware of Impression share at the adgroup level which is good to know.

    • October 24, 2012 at 1:33 pm

      Thanks, Guillermo. That Ad group level impression share data is very helpeful!


Show Me My Heatmap

Looking for key #Webanalytics tools? Try @CrazyEgg for the best way to track your #digitalmarketing efforts in a simple to read #data form

Brad Chuck


What makes people leave your website?