Huge differences in the amount of returned pages

[inertia]

I've been doing some keyword research for an air conditioning rental client of ours. My usual process is to head to wordtracker, collect a big list of keywords and head to google to check out the competition for each of those keywords. One of the things ill check is the number of returned pages... When searching for air conditioning rental i get 884,000 returned pages. I figured this was a bit low so i got on the msn and asked a few other people to check out the results they see. The first guy had 16,000,000 and the second guy had 4,000,000. They are both based within 30 miles of me.

What im confused about is the huge difference between these results? All results were worldwide and non of us were logged into google. Everyone in my office gets the same result as me. Can anyone shed any light on this as it is making the level of competition very hard to calculate (if any UK people want to check what results they get and let me know ill be more than gratefull!)

Peace out...

[kgun]

Originally Posted by inertia

When searching for air conditioning rental i get 884,000 returned pages. I figured this was a bit low so i got on the msn and asked a few other people to check out the results they see. The first guy had 16,000,000 and the second guy had 4,000,000. They are both based within 30 miles of me.

My results

1. "air conditioning rental" --> 11 900 hits on Google.
2. air conditioning rental --> 16 600 000 hits on Google.

Side note, a time since I learned about Word Tracker, but do you know all of the functionality? I got 6 good audios about WordTracker from this Onsite search engine workshops taught by industry leaders Robin Nobles and John Alexander site. Here Internet Marketing Center - Learn How to Make Money Online is another good marketing resource site.

[inertia]

Been using wordtracker for a while so i think i know it quite well. But i will have a listen regardless. Cheers kgun.

Can i ask, were those results from .com or .co.uk?

[kgun]

Originally Posted by inertia

Can i ask, were those results from .com or .co.uk?

.com but .co.uk give identical number of hits on both, while there is a difference in the exact search


co.uk -- "air conditioning rental" 35 600 hits

[wige]

A few things to note with the number of returned pages reported by Google. First, it is only an estimate, and is very rarely within 10% of the actual number of matching pages. In fact, if you wait a few minutes and do the search again the number can change by several percentage points. This is even the case when searching with the "site:" parameter on a large site. Second, Google will filter out pages that seem to be similar in content to one another, as well as filtering out pages that have limited content. This can cause the number to vary from the number reported by other search engines.

As an example, in a search for "air conditioner rental", Google says there are 30,600 results. However, only 391 pages actually show up in the SERPs. Subsequent searches only returned 345 and 215 documents, respectively (could not get the filter disabled at all, maybe a bug?).

[kgun]

Originally Posted by wige

In fact, if you wait a few minutes and do the search again the number can change by several percentage points.

I know, but that did not happen in this case. I got identical results on the free (without " ... ") search two times within minutes.

May be I will get another result If I log out and try later today or tomorrow.

[inertia]

I get the same result every time - 884,000 - and everyone in the office gets the same. I've just checked the result using a proxy server and it was 14,300,000. When i checked all the keywords using the proxy they were all a lot higher than the standard results from my IP.

I've always used the returned results number as a benchmark for how competitive a keyword is going to be. I realise that this is not a perfect system and some human discretion has to be used to properly evaluate a keywords competitiveness. The main reason im puzzled is the huge differences between results. Different data centres will give different results and i agree that they will fluctuate all the time but from 884,000 to 14,300,000!!!!??? Thats a big difference from what is supposed to be the same search. This could potentially effect what keywords i choose to target.

[kgun]

I see the following possibilities:

1. Internal filters at your company. I doubt that.
2. Your company always get results from a local or special data center. Have you used domain and IP tools to check the Ip you are on? Have you asked people in the same building / street, but on another IP to do the same search?
3. Geo location problems. How far out from your location do you have to go to get more "natural" results?

[inertia]

I asked a couple of friends who work about 20 miles north and their results seemed more natural. The next stage is to ask people in the same office block i suppose. They will be on a different IP because we go through our own router. I'll do that!

The factors that i was considering were more along the lines of personalised search. The phrase "air conditioning online" does get searched for quite a lot from here (due to it being our clients biggest phrase). I was also thinking that it may have something to do with the ranking software which we run???

[cgrantski]

If you jump to the last page of results, you'll get the much smaller real number of hits. I'm wondering if those might agree a lot better than Google's first wild guesstimate.

[incrediblehelp]

Why are you worried about the number of results returned for query? Surely you are not putting any sort of stock in this number right?

[Tom O'Leary]

I'm in Bellingham, Washington and got a result of 16,900,00 for:

air conditioning rental

[subsystems]

I think Icrediblehelp is right. Although it might seem helpful to know how many competing pages you will be going up against for a given keyword, I would tend to put more weight in seeing if say the to 40 listing all have high PR. It would be a better indicator of competition and the difficulty of reaching good rankings.

[Mijas Real Estate]

I've just searched for air conditioning rental and @ Google 16,600,000 & @ Google 16,400,000 so not much difference between them & thats from the location of Mijas, Costa del Sol, Spain
I hope that helps

[tomcatuk]

Thought I'd take a peek (I'm in London).

.com - 16700000

.co.uk - 17000000

The only reason I can think of is the datacentre Google is running your query differing to the ones your 30 mile away friends are getting.

[dbramley]

Thought I might as well take a look myself!

Based in Hartlepool, UK:

.com - 17,000,000
.co.uk - 1,800,000

[known]

I'm in northern Michigan.

"air conditioning rental" rendered 42,400 results

air conditioning rental rendered 15,400,00 results

Hope this helps.

Cheryl C. Cigan
Known.com Web Page Optimization - Known Web Page Products are content-rich bridges of information between your potential customer and your business

[deepsand]

With respect to which Google data center serves a request, a factor more important than geographical proximity is which ISP is being used, as such will determine at what point in the cloud the request is injected. Next door neighbors, with different ISPs, can witness different results.

In fact, even from the same location, such can occur. I once ran repeated traceroutes from a client's office, over a PayTech T-1 link, a Verizon T-1 link, & a Verizon DSL link, consistently reaching different hostservers.

[astro]

Results 1 - 10 of about 17,000,000 for air conditioning rental. (0.17 seconds)

Beat ya all!!!!!!

Results 1 - 10 of about 43,400 for "air conditioning rental". (0.29 seconds)

Not so spectacular! Just for the record I am in Greece. Both results returned from Google .co.uk

I have managed to escape the big brother syndrome on our little island as Google have a little quirk in Greece, I have not come across elsewhere. No matter which Google URL you enter in the address bar (.com .co.uk, .au etc.) you will be re-routed to google.gr automatically, which as a search engine for accurate and relevent searches is close to useless as the database thus results are manipulated here. (did I really say that!)

Another interesting fact is I find my websites feature higher in some countries than others. In Australia for example my two main sites are one and two under one search term. Yet this is not repeated in other countries. Does the criteria for listing results change according to country? If so can you target particular countries. Scandanavians have recently discovered our little island and can now fly direct here. Just a thought.

/astro

[Webnauts]

Google US returns 15,000,000 air conditioning rental - Google Search. But still I do not understand what do you worry about? You must concentrate on search volumes and not on number of found sites/pages in the SE results.

[Terry Van Horne]

Inertia. I'm not sure why you are doing this but if it's for competition evaluation then that is a weaker search to really show competition. Dan thies taught me these and they were real eye openers:
allintitle and allinlink give real indications of the true competition.

There several reasons why different results show up some not mentioned are personalized and universal/blended search. Which are likely also affected by most of the reasons given here for discrepancies.

A lot of it is a function of location/geography

[deepsand]

There are a host of nifty advanced operators available on Google.

Google Search Operators - Google Guide

[inertia]

I check a lot of things when doing my keyword research, including the Adwords tool, SEO quake for the first 3 or 4 results pages, the number of returned results with speech marks and the kei on wordtracker. By checking all of these i am able to roughly figure out what it's going to take for me to get a client into top positions. The allintitle and allinlink searches are (foolishly) something which i have never done and i will add them to my list. What puzzled me most was the HUGE difference in returned results. Ive checked again this morning and its at 18,000,000. (Am i winning??!)Im aware of personalisation, geo-location etc but fluctuations like this seem beyond the affect of these aspects.

Many thanks for the help people..

[thepoacher]

I got 18,000,000 this morning too, inertia - I'm down in Hampshire.

[Webnauts]

Originally Posted by inertia

I check a lot of things when doing my keyword research, including the Adwords tool, SEO quake for the first 3 or 4 results pages, the number of returned results with speech marks and the kei on wordtracker.

I added at another thread some additional tips for competition keywords research: How to get competitors keywords

[subhzash]

Search volume is important. Number of results are not at all important. I suggest use Google adwords tool and take the high search count keywords. Thats it. Dont bother about number of results



Regards
Subhzash

[Terry Van Horne]

Originally Posted by subhzash

Search volume is important. Number of results are not at all important. I suggest use Google adwords tool and take the high search count keywords. Thats it. Dont bother about number of results

Exceptional tool for Adwords! Ummm.. search volume doesn't give you anything to determine a budget and doesn't mean anything if you aren't in the top 10 or 20 ranking. IMO, Search volume and other KW analysis can be used to determine/refine your competition analysis. Knowing the volume

Competition analysis using the intitle and inlink research tells you:
1. Real Project Difficulty: exteremely important for quoting fees and Top position ETA (IME- they always ask
2. To some degree the amount of content needed ie more difficulty=more content or a Ton o' links
3. Link development needs (are the competitors using link text aggressively? ->More content)

I further refine this by " "ing the phrase which tells you the number of exact matches in title and linktext -true indicators of SEO competition the rest were lucky or..... mediocre SEOs. I've seen some results that you would think would be tough judging by the #'s of results...but... SEO... not so much.

[inertia]

I agree. Although this is still one of the areas where i find it difficult to fine tune my predictions and therefore price. For example, i have a potential client who has given me half a dozen keywords that he wants to be top for (i wont go too deep into the particulars). Now, ive had a look at the serps, ive done some research and i reckon it can be done quite easily with a few months work. But! That is me estimating for myself. How do people price up that kind of third party SEO? Especially when we all know that you cant make guarantees!

[Terry Van Horne]

I do ala Carte (it's the chef in me) pricing. For me it is important to know the difficulty because like a mechanic I work on an hourly rate, which varies in price depending on the service I'm providing. When I do the RFP work for SeoPros I do the research above and then estimate a number of hrs. based on my experience of roughly 15 mins/submission with a limited review of resource. Content I access the current content for SEO suitability. For copy I use a $300/page and make a suggestion for how many new pages are needed. I also include a site review accessing the privacy, TOS, Search and Company Overview (what do they do it and how they do it on the site, not some nonsense mission statement) and include a rough guestimate of hrs. to upgrade any site architectual or navigation/indexing problems.

[inertia]

Do you mean bespoke (that's the carpenter in me) pricing? So you judge each client on their own merits? I was trying to push for this sort of pricing method at my old place but they sold set packages which i thought was ridiculous. I'd have a guy selling insurance on the same package as a guy selling... tool steel! It made no sense.

Just another quick note. I am finding lots of results where the returned pages is higher for UK only searches than world wide searches? Something else that makes no sense?!

[Webnauts]

Originally Posted by inertia

Just another quick note. I am finding lots of results where the returned pages is higher for UK only searches than world wide searches? Something else that makes no sense?!

Just a question: Are you seeing those different results when being logged in Google and when not?

[inertia]

Im logged out of my google and always do when im doing any research.

I've noticed this and commented on it a few times over the last year. World wide results are often LOWER than UK only??? Which is impossible and suggests that there are data centre problems. Is it just me or has google been getting stranger and stranger over the last 12 months?

I've just checked a few of the keywords where this phenomenon was happening and it is no longer. This must be temporary data centre errors?

[Webnauts]

It is obvious that it is Google. So don't waste your time trying to figure out what Google is doing wrong if it does not affect your work.

[sryk222]

hmmm..interesting! google is becoming very popular these days! the days for yahoo is counted! lol

[Terry Van Horne]

Originally Posted by inertia

Do you mean bespoke (that's the carpenter in me) pricing? So you judge each client on their own merits?

For some of my services I charge $500/hr. for others... I'll do it for free if I like your story.Every website is unique! even those within the same industry and niche are unique in
1.the status/quality of the content
2.promotion of the site to date
3.skills required to execute the plan

How can you have a package solution if you don't know what the solution is for? You can set rates as in $/Hr/task (which are based on input costs) that gives you a formula but... the formula doesn't work without assessment. Any SEO who tries to sell you without an extensive assessment is not an SEO they are a web designer who thinks they do SEO or as I call them... poser/wannabe SEO.

[inertia]

Originally Posted by Terry Van Horne

For some of my services I charge $500/hr. for others... I'll do it for free if I like your story.Every website is unique! even those within the same industry and niche are unique in
1.the status/quality of the content
2.promotion of the site to date
3.skills required to execute the plan

How can you have a package solution if you don't know what the solution is for? You can set rates as in $/Hr/task (which are based on input costs) that gives you a formula but... the formula doesn't work without assessment. Any SEO who tries to sell you without an extensive assessment is not an SEO they are a web designer who thinks they do SEO or as I call them... poser/wannabe SEO.

I agree. Each site has different requirements, level of competition, prior achievements etc... You cant possibly make standardised packages which apply to every site (except consultation services where its a definite period of your time). I have sometimes found that when estimating what and how much work needs to be done on a site i occasionally get it wrong and what appeared to be an easy job turns out to be much harder (and vice versa).