Google's Search Volumes Are So Inaccurate

17 years ago

Over time I've come to interpret keyword search-volume results as 1) only a best-guess, no matter where it comes from, and 2) only useful in relation to one another. That is to say, if a tool tells me keyword X has 500,000 searches and keyword Y has 1,000,000, I take that to mean that that particular keyword tool suggests that keyword Y should get twice as many searches as keyword X.

From that, to actually trusting that the numbers are going to translate into actual traffic numbers is a really thin and curvy road... the best you can do is take action going after the keywords that look best and see what results you get, learn, rinse and repeat. Or do PPC with a high bid to really get a fast education on the actual traffic (I do not recommend this).

Also beware of anybody that tries to get you to believe or assume that those numbers are going to translate as real numbers in practice. Anything that sounds like "The keyword tool says 100,000 searches a month, so if we can get even 1% of that we would get 1,000 people visiting our site every month!" or anything like that, just turn around, walk away.

[ 2 ] Thanks

Signature

- Harry Behrens

{{ DiscussionBoard.errors[884608].message }}

Thomas

17 years ago

Isn't another factor to throw into the mix the fact that organic SERPS vary tremendously in CTR, in the same way that PPC ads do?

For example, if the title and/or description in the no. 1 listing says "I'll chop your head off if you click this" and the number 2 listing says "I'll send you 72 virgins if you click this", there's a pretty good chance no. 2 will have a greater CTR.

Tommy.

[ 1 ] Thanks
1 reply

{{ DiscussionBoard.errors[884817].message }}

JustaWizard

17 years ago

I know this probably isn't all that helpful, but don't you think it's a bit presumptuous to assume that you'll get clicked on by 42% of all web surfers just because you're in #1 position?

I understand that's supposed to be the stats, and I understand that the keyword tool may not be 100% accurate, but whether someone clicks on your site depends on the title and description to some extent too and not just blind clicking, yes?

Best,
David

Thanks
1 reply

{{ DiscussionBoard.errors[884876].message }}

smedsribaka

17 years ago

Originally Posted by JustaWizard

I know this probably isn't all that helpful, but don't you think it's a bit presumptuous to assume that you'll get clicked on by 42% of all web surfers just because you're in #1 position?

I understand that's supposed to be the stats, and I understand that the keyword tool may not be 100% accurate, but whether someone clicks on your site depends on the title and description to some extent too and not just blind clicking, yes?

Best,
David

I would also have to agree with this assessment. Plus the fact that Google changes its algorithm twice a week now makes it hard to keep up on what they are really doing (not like we ever really knew in the first place!).

Thanks

Signature

http://www.TheWebSmithGroup.com Helping companies market their business on the Internet to Local Customers or around the World.

{{ DiscussionBoard.errors[885052].message }}

Stephen Crooks

17 years ago

The best technique is to do some adwords testing of exact match for the keyword you want to test. Try and make your ad uninteresting (so people don't click it) and count how many impressions you get after a few days.

1. Make sure your daily budget is large enough for it to show all the time
2. Bid high enough for it to be on page 1 all the time.
3. Use a different adwords account from the normal one to prevent your account from suffering.
4. Make your ad as boring and irrelavant as possible.
5. Switch of content network and search partners as well.

This will give you extremely accurate search figures for your keyword.

Failing that, the Google keyword tool IS as accurate as it gets for keyword research. Your experience could be an anomaly as mine is totally the opposite to yours.

Originally Posted by artlan

So I've been busy lately optimizing for a term that has an EXACT MATCH of 18,100 global monthly search volume. I also took into consideration the local monthly search volume which is listed at 8,100. The numbers are taken from the Google Adwords Keywords Tool.

So after finally climbing to #1 in Google, that keyword brings in anywhere from a whooping 23-32 visitors a day, according to Google Analytics. At best, this represents a search volume of just 960 per month.

So I tried WordTracker and the numbers there indicate that the keyword should bring around 249 visitors a day.

According to traffic distribution by google ranking research, I should be getting 42% of all the clicks for being ranked #1. It seems like, however, I only get around 9%.

Now I understand that Google derives its statistics from searches on the Google search engine as well as the overall search network, but I can't help but notice the tremendous inaccuracy of these numbers.

This deems keyword researching almost useless as you never know even the approximate number of searches for a given keyword.

My questions are as follows:

1) Are there any other tools, free or paid which represent better overall numbers?

2) How do you overcome such inaccurate numbers when you do keyword researching? Do you use a formula to deduct from the original amount display by Google Adwords Keyword Tool?

3) What would you recommend to do/use in order to arrive at a more accurate number?

[ 2 ] Thanks

Signature

Affiliate Marketing Profits 2020 - Step by Step Guides
How to Make $100 a day Online

{{ DiscussionBoard.errors[884933].message }}

Steven Carl Kelly

17 years ago

Ok, this is something I've been preaching for a long, long time. You have to understand how all of these numbers fit together.

1. If you're trying for organic searches in the Google SE itself, then the KW tool numbers are way off for you. You may want to consider checking the SK Tool for more realistic numbers. The KW tool was designed for Adwords advertisers, its volumes represent a much more realistic number for Adwords impressions.

2. For most folks, you'll want to focus on local search volumes. Turn off the global column completely, unless you're targeting languages other than English. You can set the target area using the link just above the box where you enter your keywords. Use it to select the region you are targeting with your product. For me, that's "US English" but I will also add Canada, Australia, and the UK.

3. Remember: the stats include Google itself AND the search network. Again, this is fine for the Adwords advertiser, but isn't so useful if you're doing keyword research for organic search results within Google proper only.

4. To the best of my understanding, the traffic distribution ranking chart referred to is based on CLICKS and not searches. There's a big difference. We also know that the traffic distribution graph cannot be accurate, because it adds up to exactly 100%. First, that would exclude any abandoned searches (do you ever Google something and then not click on a link at all? I do), it excludes clicks on Adwords ads, and it accounts for no clicks that occur deeper than page two in the results (have you ever clicked a link on page 3? Page 4?).

I usually advise people to use exact matching, local search volume, and divide that number by 5 to get a realistic number to start with. Then divide that number by 30 to come up with a more realistic "pure Google SE" only daily search count. Then divide that number by 2 for a decent estimate of daily clicks you can expect if you're in position #1. For position #2, you can expect about about 20% of that more realistic daily search count.

So if we apply the formula I teach to your situation:

- Exact match, check
- 8,100 local search volume
- Divided by 5 = 1620 monthly estimate for Google proper
- Divided by 30 = 54 Google organic searches per day
- Divided by 2 = 27 clicks per day for position #1

So you can expect about 27 clicks a day, on average. You said you're getting 23 - 32, so the average of those would be 27.5 clicks a day -- which is right about the number my formula tells you to expect.

I've been teaching that formula to my clients/students for a while now. I'm going to include that formula in WoMM when it releases.

[ 2 ] Thanks
4 replies

Signature

Read this SURPRISING REPORT Before You Buy ANY WSO! Click Here
FREE REPORT: Split Test Your Landing Pages the Easy Way

{{ DiscussionBoard.errors[884998].message }}

Canada

17 years ago

Originally Posted by Steven Carl Kelly

...
1. If you're trying for organic searches in the Google SE itself, then the KW tool numbers are way off for you. You may want to consider checking the SK Tool for more realistic numbers. The KW tool was designed for Adwords advertisers, its volumes represent a much more realistic number for Adwords impressions.
...

Hey Steven, I found that calculation really interesting -- thanks for sharing. One thing, though. What are you referring to when you say the SK tool. Do you mean the Search based keyword tool?

Thanks
1 reply

Signature

YourNichePLR.com - Fresh new PLR site ala Tiffany Dow's PLR ATM.
50% Off Everything In Your Cart until November! Sign Up and be Notified when New PLR packs are uploaded. Right now we've got acid reflux, kittens, and hair styling packs available.

{{ DiscussionBoard.errors[885917].message }}

Steven Carl Kelly

17 years ago

Originally Posted by Canada

Hey Steven, I found that calculation really interesting -- thanks for sharing. One thing, though. What are you referring to when you say the SK tool. Do you mean the Search based keyword tool?

Yep! Right now, the data feeding that tool makes it a lot closer to accurate for organic search.

Thanks

Signature

Read this SURPRISING REPORT Before You Buy ANY WSO! Click Here
FREE REPORT: Split Test Your Landing Pages the Easy Way

{{ DiscussionBoard.errors[885944].message }}

Travis Van Slooten

17 years ago

Wow...great post! Thanks for sharing this formula. I don't know that I'll go through the trouble of doing all the calculations you've outlined, but I'll definitely do exact matches and local search to get a more accurate figure. How in the world did you ever come to this formula in the first place???

Travis

Originally Posted by Steven Carl Kelly

Ok, this is something I've been preaching for a long, long time. You have to understand how all of these numbers fit together.

1. If you're trying for organic searches in the Google SE itself, then the KW tool numbers are way off for you. You may want to consider checking the SK Tool for more realistic numbers. The KW tool was designed for Adwords advertisers, its volumes represent a much more realistic number for Adwords impressions.

2. For most folks, you'll want to focus on local search volumes. Turn off the global column completely, unless you're targeting languages other than English. You can set the target area using the link just above the box where you enter your keywords. Use it to select the region you are targeting with your product. For me, that's "US English" but I will also add Canada, Australia, and the UK.

3. Remember: the stats include Google itself AND the search network. Again, this is fine for the Adwords advertiser, but isn't so useful if you're doing keyword research for organic search results within Google proper only.

4. To the best of my understanding, the traffic distribution ranking chart referred to is based on CLICKS and not searches. There's a big difference. We also know that the traffic distribution graph cannot be accurate, because it adds up to exactly 100%. First, that would exclude any abandoned searches (do you ever Google something and then not click on a link at all? I do), it excludes clicks on Adwords ads, and it accounts for no clicks that occur deeper than page two in the results (have you ever clicked a link on page 3? Page 4?).

I usually advise people to use exact matching, local search volume, and divide that number by 5 to get a realistic number to start with. Then divide that number by 30 to come up with a more realistic "pure Google SE" only daily search count. Then divide that number by 2 for a decent estimate of daily clicks you can expect if you're in position #1. For position #2, you can expect about about 20% of that more realistic daily search count.

So if we apply the formula I teach to your situation:

- Exact match, check
- 8,100 local search volume
- Divided by 5 = 1620 monthly estimate for Google proper
- Divided by 30 = 54 Google organic searches per day
- Divided by 2 = 27 clicks per day for position #1

So you can expect about 27 clicks a day, on average. You said you're getting 23 - 32, so the average of those would be 27.5 clicks a day -- which is right about the number my formula tells you to expect.

I've been teaching that formula to my clients/students for a while now. I'm going to include that formula in WoMM when it releases.

Thanks
1 reply

Signature

TVS Internet Marketing: Helping small businesses get more visibility & sales online.

{{ DiscussionBoard.errors[885946].message }}

Steven Carl Kelly

17 years ago

Originally Posted by tvanslooten

Wow...great post! Thanks for sharing this formula. I don't know that I'll go through the trouble of doing all the calculations you've outlined, but I'll definitely do exact matches and local search to get a more accurate figure. How in the world did you ever come to this formula in the first place???

Travis

Travis:

It doesn't take a minute to do, and it's even easier if you have a spreadsheet already set up to export the keywords and their volumes. As to where I got the formula... lots of experience, testing, research, and a little bit of extra info from a source that must remain unnamed.

You guys got a free peek into WoMM!

Thanks

Signature

Read this SURPRISING REPORT Before You Buy ANY WSO! Click Here
FREE REPORT: Split Test Your Landing Pages the Easy Way

{{ DiscussionBoard.errors[885955].message }}

Stephen Crooks

17 years ago

I agree with a lot of what you say except the calculations seem way off to me. Once you get the match types and International targeting set correctly then the search numbers are reasonably accurate.

Yes, it was designed for adwords advertisers originally and of course that includes it's search partners as well but as Google's search accounts for around 80% of all search numbers then at worst it's around 20% innacurate.

Even if they didn't count searches that didn't lead to clicks then that is irrelevant anyway, why would they count searches that didn't lead to clicks?

I agree with you about the search based keyword tool and that is rapidly becoming my favourite keyword tool as I discover more about it plus the "insights for search" tool is also amazing for researching across various regions.

I suggest people run their own tests using the method I descibed in my previous post and then compare search numbers with google's keyword tool and you will be suprised at just how accurate the tool is.

Originally Posted by Steven Carl Kelly

Ok, this is something I've been preaching for a long, long time. You have to understand how all of these numbers fit together.

1. If you're trying for organic searches in the Google SE itself, then the KW tool numbers are way off for you. You may want to consider checking the SK Tool for more realistic numbers. The KW tool was designed for Adwords advertisers, its volumes represent a much more realistic number for Adwords impressions.

2. For most folks, you'll want to focus on local search volumes. Turn off the global column completely, unless you're targeting languages other than English. You can set the target area using the link just above the box where you enter your keywords. Use it to select the region you are targeting with your product. For me, that's "US English" but I will also add Canada, Australia, and the UK.

3. Remember: the stats include Google itself AND the search network. Again, this is fine for the Adwords advertiser, but isn't so useful if you're doing keyword research for organic search results within Google proper only.

4. To the best of my understanding, the traffic distribution ranking chart referred to is based on CLICKS and not searches. There's a big difference. We also know that the traffic distribution graph cannot be accurate, because it adds up to exactly 100%. First, that would exclude any abandoned searches (do you ever Google something and then not click on a link at all? I do), it excludes clicks on Adwords ads, and it accounts for no clicks that occur deeper than page two in the results (have you ever clicked a link on page 3? Page 4?).

I usually advise people to use exact matching, local search volume, and divide that number by 5 to get a realistic number to start with. Then divide that number by 30 to come up with a more realistic "pure Google SE" only daily search count. Then divide that number by 2 for a decent estimate of daily clicks you can expect if you're in position #1. For position #2, you can expect about about 20% of that more realistic daily search count.

So if we apply the formula I teach to your situation:

- Exact match, check
- 8,100 local search volume
- Divided by 5 = 1620 monthly estimate for Google proper
- Divided by 30 = 54 Google organic searches per day
- Divided by 2 = 27 clicks per day for position #1

So you can expect about 27 clicks a day, on average. You said you're getting 23 - 32, so the average of those would be 27.5 clicks a day -- which is right about the number my formula tells you to expect.

I've been teaching that formula to my clients/students for a while now. I'm going to include that formula in WoMM when it releases.

Thanks
2 replies

Signature

Affiliate Marketing Profits 2020 - Step by Step Guides
How to Make $100 a day Online

{{ DiscussionBoard.errors[886602].message }}

Steven Carl Kelly

17 years ago

Originally Posted by Steve Crooks

I agree with a lot of what you say except the calculations seem way off to me. Once you get the match types and International targeting set correctly then the search numbers are reasonably accurate.

Much respect you, but we'll have to agree to disagree, Steve. I've developed this simple formula over many, many years of intense tracking and analysis. I decided I wanted to come up with a reliable way to use the KW tool to give numbers I could count on, and after stacking everything up and breaking it back down again, this formula has been the one that has consistently delivered.

Regardless of the "why's" involved, this formula has worked for me and my clients and the students from my seminars. Naturally, everyone's experience is unique and there are many factors involved that we haven't even discussed that affect CTR.

No disrespect intended to your opinion. I can only address my own experiences and my own research and findings. If your experience is different but it works for you, then that's fantastic! Then we're both happy, and that's really cool.

Thanks

Signature

Read this SURPRISING REPORT Before You Buy ANY WSO! Click Here
FREE REPORT: Split Test Your Landing Pages the Easy Way

{{ DiscussionBoard.errors[887417].message }}

Steven Carl Kelly 17 years ago

Originally Posted by Steve Crooks

Even if they didn't count searches that didn't lead to clicks then that is irrelevant anyway, why would they count searches that didn't lead to clicks?

Completely and totally relevant to the OP's statements!

First, if you're talking about the traffic distribution chart, then clearly it proves the chart to be inaccurate. Well, we know for a fact it is inaccurate because it isn't mathematically or logically possible to be distributed as they indicate.

Second, if you're talking about Adwords then it counts as an impression. That's why Google provides the internal preview tool that allows you to search all day long without it counting against the search volume numbers or the impressions for your Adwords ads (which reduces CTR, which increases your cost).

Third, it is extremely relevant and important to organic search results and the expected CTR provided by first page ranking for targeted keywords when doing organic SEO.
- Thanks
Signature

Read this SURPRISING REPORT Before You Buy ANY WSO! Click Here
FREE REPORT: Split Test Your Landing Pages the Easy Way
{{ DiscussionBoard.errors[887449].message }}

sarafina

17 years ago

Originally Posted by Steven Carl Kelly

Ok, this is something I've been preaching for a long, long time. You have to understand how all of these numbers fit together.

1. If you're trying for organic searches in the Google SE itself, then the KW tool numbers are way off for you. You may want to consider checking the SK Tool for more realistic numbers. The KW tool was designed for Adwords advertisers, its volumes represent a much more realistic number for Adwords impressions.

2. For most folks, you'll want to focus on local search volumes. Turn off the global column completely, unless you're targeting languages other than English. You can set the target area using the link just above the box where you enter your keywords. Use it to select the region you are targeting with your product. For me, that's "US English" but I will also add Canada, Australia, and the UK.

3. Remember: the stats include Google itself AND the search network. Again, this is fine for the Adwords advertiser, but isn't so useful if you're doing keyword research for organic search results within Google proper only.

4. To the best of my understanding, the traffic distribution ranking chart referred to is based on CLICKS and not searches. There's a big difference. We also know that the traffic distribution graph cannot be accurate, because it adds up to exactly 100%. First, that would exclude any abandoned searches (do you ever Google something and then not click on a link at all? I do), it excludes clicks on Adwords ads, and it accounts for no clicks that occur deeper than page two in the results (have you ever clicked a link on page 3? Page 4?).

I usually advise people to use exact matching, local search volume, and divide that number by 5 to get a realistic number to start with. Then divide that number by 30 to come up with a more realistic "pure Google SE" only daily search count. Then divide that number by 2 for a decent estimate of daily clicks you can expect if you're in position #1. For position #2, you can expect about about 20% of that more realistic daily search count.

So if we apply the formula I teach to your situation:

- Exact match, check
- 8,100 local search volume
- Divided by 5 = 1620 monthly estimate for Google proper
- Divided by 30 = 54 Google organic searches per day
- Divided by 2 = 27 clicks per day for position #1

So you can expect about 27 clicks a day, on average. You said you're getting 23 - 32, so the average of those would be 27.5 clicks a day -- which is right about the number my formula tells you to expect.

I've been teaching that formula to my clients/students for a while now. I'm going to include that formula in WoMM when it releases.

How do they derive the data for local searches. Is it based on IP address or when people use the local version of google? ie. google.ca or google.co.uk

Thanks

{{ DiscussionBoard.errors[888245].message }}

mikey1090

17 years ago

I have noticed discrepancies. Makes it hard when I'm thinking of launching my own product in a niche with 20k searches/month if the niche isn't actually that popular.

Thanks

Signature

.

{{ DiscussionBoard.errors[885546].message }}

artlan

17 years ago

To all those replied and in particular Steven Carl Kelly, thank you for such an insightful thread, I definitely learned a whole lot.

The SK tool that is mentioned a couple of times in this thread, can someone provide me a URL?

Thanks

Thanks
1 reply

{{ DiscussionBoard.errors[887555].message }}

Steven Carl Kelly

17 years ago

Originally Posted by artlan

To all those replied and in particular Steven Carl Kelly, thank you for such an insightful thread, I definitely learned a whole lot.

The SK tool that is mentioned a couple of times in this thread, can someone provide me a URL?

You're welcome. And:

Search-based keyword tool

Thanks
1 reply

Signature

Read this SURPRISING REPORT Before You Buy ANY WSO! Click Here
FREE REPORT: Split Test Your Landing Pages the Easy Way

{{ DiscussionBoard.errors[887559].message }}

thecableguy 17 years ago

JMO but I think Google counts the combined search and content networks while other keyword tools gives only the search results.

Alan
- Thanks
{{ DiscussionBoard.errors[888042].message }}

nofearman

17 years ago

Originally Posted by artlan

So I've been busy lately optimizing for a term that has an EXACT MATCH of 18,100 global monthly search volume. I also took into consideration the local monthly search volume which is listed at 8,100. The numbers are taken from the Google Adwords Keywords Tool.

So after finally climbing to #1 in Google, that keyword brings in anywhere from a whooping 23-32 visitors a day, according to Google Analytics. At best, this represents a search volume of just 960 per month.

So I tried WordTracker and the numbers there indicate that the keyword should bring around 249 visitors a day.

According to traffic distribution by google ranking research, I should be getting 42% of all the clicks for being ranked #1. It seems like, however, I only get around 9%.

Now I understand that Google derives its statistics from searches on the Google search engine as well as the overall search network, but I can't help but notice the tremendous inaccuracy of these numbers.

This deems keyword researching almost useless as you never know even the approximate number of searches for a given keyword.

My questions are as follows:

1) Are there any other tools, free or paid which represent better overall numbers?

2) How do you overcome such inaccurate numbers when you do keyword researching? Do you use a formula to deduct from the original amount display by Google Adwords Keyword Tool?

3) What would you recommend to do/use in order to arrive at a more accurate number?

If you 2 line desfription (and/or heading) in search results is irrelevant or NOT enticing to web surfers, NO ONE is going to click on your site link listing no matter what position you are in.

Thanks
1 reply

{{ DiscussionBoard.errors[888690].message }}

Izesta 17 years ago

I posted this as a new thread today, but wish I had just responded in this thread.

This is related, so I hope no one minds me posting my comment/question here.
_____________________

I am a bit perplexed with Google Local Search and how to measure/track search queries and results.

I hope somebody will chime in with thoughts/expert guidance that's so commonplace here.

I did my own Local listing earlier this year. My Dashboard in my Google LBC shows 1816 impressions over the prior 30 day period. So, 1816 searchers would have "seen" my biz listing as a local search result during that 30 days. Only 127 have clicked through to my site. Doesn't seem like a lot, but I get a good share of calls and email, so I'm okay with it. I do list at #1, 2 or 3 for several different keyword searches - keywords I chose when I did the listing.

Here's what's weird: In my LBC, Google shows the "Top 10 Search Queries" - or, what searchers typed in when they saw my business listing as a match in the Local listings. Not one of these 10 Search queries match my keywords. And not one of them is a local search term. I realize we can now enter in one word and get Local Search results in some cases. But the queries in my LBC do not result in Local Searches.

If I type in what Google gives as my Top 10 Queries, I do NOT get local listings at all. Just organic. With one exception. Google shows my first name as one of the top 10 Queries. When I type in my first name, I do get my business in a 10 pack listing (IP address tracked of course).

When I look at the words used in Google's Top search queries, it's nothing more than a hodge podge of words taken from the info in my business listing.

I would really like to know what my "true" Top Search Queries are. This is information Google brags about being able to provide.

None of this gives me a warm and cozy feeling about Google's numbers and analytics for Local Search or anything.

To summarize: What's really happening in Google Local Search does not seem to mesh up in any way with Google's analytics. I'm sure Google wants Local Search to appear successful, but to provide false query data in my LBC that I can easily verify is a tad foolish. I thought Google Local was FAR from the infancy stage in terms of this basic data - (or what I believed was basic data). I wonder whether to believe the 1816 impressions.

Has anyone else looked at this closely or have any feedback?

Hope this made sense. Thanks for any advice/comments.

Izesta
- Thanks
{{ DiscussionBoard.errors[1004550].message }}

Google's Search Volumes Are So Inaccurate

Trending Topics

Tribute to Mark...

Building a Children's Educational Ecommerce Ecosystem Instead of Just a Shopify Store

What if we banned Google?

Google Adding Ads to AI?

Is sleep overrated ??