in the initial stages of keyword discovery and determining performance, it is not uncommon for us to bump up the bids of the phrase or broad match terms so that they are closer to or even above the exact match terms. This allows us to review our logs for organic search terms coming through on the broad or phrase matches in order to add these as keywords or refine negative keywords.

One thing we’ve seen consistently though is that even though we have an exact match term, google will very frequently serve up our phrase match term, less often the broad match term.

Often (but not always) our phrase match term in these cases has a higher bid. Even so, I would expect Google to serve up our Broad match term if someone searches for that term exactly.

Can someone explain how/when google decides to serve up broad match or phrase match rather than exact match terms when someone does that exact search?

Thanks!

Jennifer
vincevincevince

#:3814788      6:33 am on Dec. 27, 2008 (utc 0)

Broad match terms should be seen as fuzzy targets – even if you wrote ‘blue widgets’ as a broad match term, that’s not an precise hit for blue widgets. A phrase or exact term for ‘blue widgets’ would be considered an un-fuzzy and precise hit for blue widgets and hence more specific and likely to be served given the same bid.

I’m slightly confused by your thread; you seem to be mixing the terms broad match and exact match which are quite different things – could you clarify what the question is?

Do you see:
“blue widgets” shown in preference to [blue widgets]
Or:
“blue widgets” shown in preference to blue widgets (e.g. broad)
rustyzipper

#:3814797      6:55 am on Dec. 27, 2008 (utc 0)

OK, I’ll clarify:

My ad group contains:

[blue widget] ** 0.15 ** url?source=google&kw=blue%20widget_exact
“blue widget” ** 0.11 ** url?source=google&kw=blue%20widget_phrase
blue widget ** 0.07 ** url?source=google&kw=blue%20widget_broad

User searches for: blue widget

Google matches their search to my “blue widget” ad, instead of my [blue widget] ad.

in other words, Google is displaying my phrase match ad, even when someone does an exact search and I have a keyword for that exact term.

Thanks!

Jennifer
beesticles

#:3814828      12:06 pm on Dec. 27, 2008 (utc 0)

Google should always be showing the closest match to the users query, so in your case they should show the exact match. The only times when I’ve seen that this doesn’t happen is when there’s a technical error by Google. I’d advise getting in touch with them and asking them to look into it.
conor

#:3815084      12:05 pm on Dec. 28, 2008 (utc 0)

I have seen similar and am not very comfortable trusting Google to decide which of the multiple match types to display. As a best practice it is not wise to use multiple matches in the same adgroup.

If you want more control on which match type is displayed I would recommend breaking out your match types to separate adgroups and useing negative matching to force the match type you want .eg;

To force exact match in your exact match adgroup

[Blue widget]
-”Blue widget”
-Blue widget

To force phrase match in your phrase match adgroup

“Blue widget”
-[Blue widget]
-Blue widget
rustyzipper

#:3815186      7:26 pm on Dec. 28, 2008 (utc 0)

Google confirmed that this should not be happening. I’ve sent them logfiles. We have about 30,000 keywords across many campaigns & ad groups. I’ve thought about restructuring them as you suggest but it would be a major pain.
Israel

#:3815581      3:00 pm on Dec. 29, 2008 (utc 0)

broad match has me very leary of late.

I’ll give the contrived example of north climate widgets

I can’t give the real broad keyword terms without violating TOS, but it is a viable term people certainly do search on.

I saw I was getting hits from adwords for the single word north

1) Why people search on “north”, I have no idea. No ads show for that single term.

2) Why does adwords show my ad for such an odd query and it was quite often too? Is this my “reward” for a high CTR? Am I accorded an even more expanded broad match?

********

So I’m re-evaluating all my broad matches. If they contain any “common” terms, they’re out. Only supercalifragistic widgets would I dare make broad now.

I still had over 100% ROI in December, but it could have been better. At least I’ve learned something. More to add to the list of “features” that appear to be broken in adwords.

********

It’s a shame that broad match has become so useless because I can’t have an unlimited amount of phrase/exact keywords and it’s impossible to predict the many ways people may “phrase” their queries.

I’m no amateur at this either. 2004-2006 or so, any reasonable term did what it should. Now I’m dealing with an unseen enemy.

********

Anyone else seeing this? Any ideas?

Israel
Israel

#:3815583      3:04 pm on Dec. 29, 2008 (utc 0)

I have seen similar and am not very comfortable trusting Google to decide which of the multiple match types to display. As a best practice it is not wise to use multiple matches in the same adgroup.

Conor,

I’ve seen this stated by experienced folk before.

Why exactly is it undesirable now? It used to be the way to go.

Thanks!

Israel
conor

#:3815587      3:09 pm on Dec. 29, 2008 (utc 0)

Because the don’t seem to get it ‘RIGHT’ ( well not my version of Right anyway ) Most of the time
beesticles

#:3815652      4:39 pm on Dec. 29, 2008 (utc 0)

I don’t agree with you Conor, that seems a like a lot of work for no gain. in my experience Google always selects the correct match type, unless as in Rustyzipper’s case there is a technical problem.
conor

#:3815675      5:05 pm on Dec. 29, 2008 (utc 0)

If you mix match types in the same adgroup I guarantee you that Googles automatic matching will not display the ad you would want all of the time – add differing bids and bidmanagment to the mix and it becomes a mess.

Israel – we still get great results from broadmatch only campaigns where we have a really tight campaign and adgroup level negative lists and use exact negative match on broad terms. eg;

-[north]

Best Way to See AdWords Search QueryNew Search Query Report lists mostly “other queries”
BigAdventure

#:3415362     10:56 pm on Aug. 6, 2007 (utc 0)

What’s the best tool to see every search phrase that was entered associated with an AdWords campaign that cost you a click charge? My traffic is manageable enough that I would like to see them all.

Google’s new Search Query Report shows a few and lists the majority as “other queries” What use is that?
Rehan

#:3415673     10:52 am on Aug. 7, 2007 (utc 0)

Analyze your server log files. You’ll get some information in the referer details there that is not listed in the Search Query report.
BigAdventure

#:3415817     2:14 pm on Aug. 7, 2007 (utc 0)

Thanks Rehan Is there a way to get more detail?

I have a client who is trying to run a campaign that is rather unusual with negative keywords.

Say they sell “Britney Spears Apples” but they recently got a legal spanking for using the term “Britney Spears” They want to come up for searches of “Britney Spears Apples” without actually buying that term. Instead they want to buy “Apples” and try to negative out all the other terms “green apples” “washington apples” etc.

I know that Google would allow them to buy a trademarked term…just can’t use it in their ad text. But they don’t care about Google. They don’t want to take the chance of being sued again.

So the question becomes how can I get good detail on every search phrase so I can start to negative out everything but “Britney Spears Apples”

Thanks
inasisi

#:3415832     2:28 pm on Aug. 7, 2007 (utc 0)

Check out the new Search Query Performance report More details here -
http://adwords.blogspot.com/2007/07/how-to-use-new-search-query-performance.html#links
onlineleben

#:3415862     2:48 pm on Aug. 7, 2007 (utc 0)

BigAdventure,
with a little PHP you can solve this quite easily: Call an include file from within your landing page. Lets call it tracker.inc
tracker.inc writes to a text file with date/time, ip-address, and referer into a logfile (testlog.txt).
Don’t forget to set write permissions for the log file.
Code for tracker.inc is here:

$timestamp=date(”Y/m/j H:i:s”);
$addr=getenv(”REMOTE_ADDR”);

$ref=getenv(”HTTP_REFERER”);
$myfile=”/testlog.txt”;

$fd=fopen($myfile,”a”);

$success=fwrite($fd, $timestamp.”�”.$addr.”�”.$ref .”\n”);
fclose($fd);

The textfile you can import into MS Excel and select all lines beginning with http:// www .google .tld/search?hl=en
Do a Text to columns conversion by using & as delimiter.
Next you get a column where the data beginns with q= and the original query.

I know, it is a little complicated, but it helps me in finding the real search phrases.
BigAdventure

#:3416829     2:11 pm on Aug. 8, 2007 (utc 0)

Thanks online But PHP is probably beyond my skill set. Doesn’t your whole page have to be in PHP to use that?

I really need something though. I just did a search query report for one campaign, and out of 32 searches, 23 are the so-called “other unique queries. Since the others were all my narrow search phrases, I would assume those “other unique queries” are probably the garbage that I need to negative keyword.
netmeg

#:3416851     2:28 pm on Aug. 8, 2007 (utc 0)

Doesn’t your whole page have to be in PHP to use that?

No, it doesn’t.

Copy it just like this into your .html page:

<%
$timestamp=date(”Y/m/j H:i:s”);
$addr=getenv(”REMOTE_ADDR”);

$ref=getenv(”HTTP_REFERER”);
$myfile=”/testlog.txt”;

$fd=fopen($myfile,”a”);

$success=fwrite($fd, $timestamp.”�”.$addr.”�”.$ref .”\n”);
fclose($fd);
%>

If you have a lot of traffic, that file could get pretty big, though.
onlineleben

#:3416859     2:34 pm on Aug. 8, 2007 (utc 0)

If you have a lot of traffic, that file could get pretty big, though

for my main site I download the file daily, forthe smaller sites on a weekly or monthly basis. After downloading, I delete them on the server and the next visitor creates the new file.
onlineleben

#:3416865     2:37 pm on Aug. 8, 2007 (utc 0)

Doesn’t your whole page have to be in PHP to use that?

No, it doesn’t.
Copy it just like this into your .html page:

Provided that PHP is running on your server and setup in such a way, that files ending with .htm or .html can execute it.
netmeg

#:3416889     3:02 pm on Aug. 8, 2007 (utc 0)

Opps, yea, I forgot that part, it’s such a default for us I forget other people aren’t necessarily set up that way.
smallcompany

#:3418072     5:07 pm on Aug. 9, 2007 (utc 0)

A Search Query Performance report does what you want. It will show you exact keywords which triggered clicks onto your ads. One fact is that many of these queries will fall under “other”. I guess that for various reasons, they cannot pull every single keyword. I am hopping to see that improving over the time.
peer_esv

#:3418866     1:48 pm on Aug. 10, 2007 (utc 0)

If you can set it up then do your own web server log analysis to get this data. We find that it is much more precise than google’s Search Query Report.
Zealot

#:3418870     1:57 pm on Aug. 10, 2007 (utc 0)

Why not use Google Analytics? You can extract the search term from the referrer with a simple regex and place it the user defined field. Provided the tight integration with adwords you’ll be able to drill down on campaigns, adgroups, keywords and even the exact search phrases for each keyword. If you opt to import adwords cost data and enable e-commerce then you can even analyse roi down to the exact search phrase.
onlineleben

#:3418873     2:02 pm on Aug. 10, 2007 (utc 0)

more precise than google’s Search Query Report

Yeah, the SQR needs a lot of improvement
do your own web server log analysis to get this data

This is the preferred way to do things, but sometimes, depending on the analytics software you use, you don’t get everything you really want or you get even too much.
I for my part use analog as an analysis tool. Although it is a little bit old fashioned when you want to configure it via a text file, it gives me the possibility to even analyze a single page over a longer period. Just exclude all pages and include the page you want to analyze.
In case you have a landing page that only receives ppc traffic and is not linked from within your site and is not listed anywhere in the SERPs, you could go this way.
I prefer my little script from above (in an enhanced version) as it also gives me the possibility to track which ad the click was coming from.

There are no other options. Sure, you can sell, give or throw the business on someone else and make it their problem, but that doesn’t change the math. If(totalcashin<totalbizcashout)=You are now officially a consultant. No exceptions.

No matter what kind of business you have, you absolutely, positively must have a revenue strategy. No revvie, no survivee.

In the digital world not everyone, particularly those in Silicon Valley, seem to understand that. There seem to be two kinds of startup companies. The first understands this concept and knows going in exactly what they are going to sell, to who, what they hope to sell it for, their hoped for margins and just as importantly, who will sell it and how. When they model their business, they model the least expensive way to get into and stay in business with the focus on reaching profitability within months rather than years. They aren’t modeling in an exit strategy. They recognize that any company that is profitable gives them choices. They can meximize their ownership percentages. They can pay themselves an amount commensurate with their profitability. They can go public. Or they can sell. Profits provide choice.

The 2nd kind of startup relies purely on financial models to justify their future revenue streams. They create a company hoping to generate enough volume in whatever it is they hope to sell around, traffic, users, whatever. The goal is to then find a way to monetize all the volume or to execute on an exit strategy. They spend a ton of time playing with spreadsheet models. They are experts at plugging in CPMs, pageviews, unique users, sell through rates. They also know how to list “comps”. Companies that they hope to emulate and if they have only some subset of their success will more than generate enough revenue to fulfill their exit strategy. They are projection warriors. Perfectly reasonable, right ? Right, if you understand the reasoning of the people who fund these types of startups.

Venture Capitalists (VCs) tend to fall in love with concepts. They have the same problem that NBA GMs and Owners often do in evaluating players. We both see our last success in the promise of our next opportunity.

Entrepreneurs who start companies with the required, expected and motivating stars in their eyes usually forget the VC model for funding these types of companies. What’s the Venture Capital funding model ? Fund 10, hope the 1 or 2 winners more than makes up for the 8 or 9 losers. That’s right. Most VCs expect to have at least an 80pct failure rate. Which in turn means that 9 out of 10 of the entrepreneurs behind these types of “make it up on volume’ companies will end up as “consultants” .

On the flipside, if you talk to any company I have ever invested in, the only thing I care about are profitable sales. What are you selling? How hard are you working at selling? What are your revenues ? Why are you paying yourselves a salary rather than a commission ? What unique initiatives are you working on to generate sales TODAY.

When I invest in companies, I expect 100pct of them to be successful and grow and QUICKLY be profitable. I may not hit many homeruns, but I sure hit a lot of singles and doubles and rarely strike out.

Whats the point of all this and how does it apply to success and motivation ?

If you are an entrepreneur and looking at starting a company, its VERY easy to put off the hard part. Which is generating sales for your company and making a profit. Believe it or not, its far easier to go out and raise enough money so you are “in the game”. You can raise the money, start the company and take your chances. If you run out of money, you can raise more. Until you can’t. At which point you enter the world of consulting, having learned from your experience.

On the flip side, if you want to start and grow a business that you retain control of, put money in the bank from and can make a long term commitment to, then always remember that sales should be the first thing you focus on when you wake up in the morning. Profitable Sales to happy customers is the best path to making money. If you go to bed at night thinking about how to sell more and how to make your customers happy. You probably are in a good place. If you go to bed and wake up thinking about how to raise money to stay in business, you might as well get the new business cards and think about what your new consulting blog is going to look like