Search Engine Watch
SEO News

Go Back   Search Engine Watch Forums > Search Engines & Directories > Google > Google AdWords
FAQ Members List Calendar Forum Search Today's Posts Mark Forums Read

Reply
 
Thread Tools
Old 03-04-2006   #1
tonerman
Member
 
Join Date: Oct 2004
Posts: 179
tonerman has a spectacular aura abouttonerman has a spectacular aura about
Exclamation Stopping Broad Match with Expanded Match Rip-off!

When I discovered a lot of clicks in January and February for search queries that didn’t match the keywords in my Adwords campaigns caused by Google’s poorly performing Broad Match with Expanded Match matching option, I posted a thread many of you have seen called “Up the Creek with Google Broad Match.”

Since then I have seen no Google comments about this serious issue other than those made by the Adwords Representative, who has kindly both responded to my comments and passed on the information posted in that thread to the powers that be at the Googleplex. I also know from friends of mine that the issue has been raised directly to senior managers within Google Adwords. Our own famous “Aussiewebmaster” also may brought it up with Google at the SES NY 06 meeting last week.

Since mismatched impressions and clicks caused by Expanded Match can cost you a lot of money, I wanted to tell forum members how I am protecting myself against this money-sump and get any suggestions SEW members to make it easier and faster.

If you aren’t spending much on your ad campaigns, or you don’t get a lot of clicks, you can probably afford to let Google keep ripping you off if mismatched impressions and clicks are happening. If you spend thousands a month and have a high enough average cost per click to get anybody’s attention, then you need to take action to protect yourself now.

The only successful approach I have is to extract all the queries from Google adwords clicks for the day (or any time period) using my Nettracker log analyzer and looking at them individually. My PPC URLs have identifiers for each PPC source so I can identify all PPC traffic easily. If you don't have anyway to identify your Google PPC traffic you are stuck up the creek!

Log analyzers aren’t easy to read, but Nettracker let’s you export whatever data you are looking at into an Excel spreadsheet. Then I copy the column of initial referrer data from the spreadsheet into a great text editor called Ultraedit.

Using an Ultraedit macro I then get rid of everything in each line of data except the exact query term. I select and copy all the query terms and then go paste just the queries over the same initial referrer column I got them from. Why? So I can determine from the next column of exported log data what the entry page was for any query that doesn’t match my campaigns.

Then I go back to the column of just the query data in Ultraedit. If there are a lot of queries I can click a button in Ultraedit and it will both sort the data and remove all duplicate queries. That makes it faster to read the actual queries, but I won’t know how often it is occurring since I got rid of the duplicates. If I see a click caused by a query I know I don’t want, I usually copy it to another open Ultraedit document to deal with later. When I’ve pulled all the bad queries out I have to go back to the Excel spreadsheet to see what landing page is being shown so I know what campaign I need to modify to stop it.

I can mess with the data a little more and find out how clicks were caused by the mismatch to see what it cost me in $ before I got my hands around its rotten neck! Then I have to post negative keywords to any and all campaigns that are being effected by this baloney and make sure I am not shooting myself in the foot by posting a negative keyword to a campaign that destroys one or more great keywords!

So what kind of money do I save for all this work? Depends on how many bad queries I stop each time I go through this process. In the beginning it was THOUSANDS OF DOLLARS A YEAR – and I don’t just mean a couple. For a while it was 4-6 bad keywords in search terms a day triggering multiple unwanted clicks, now fewer mis-matched queries are effecting me due to my agressive hunting and there are fewer. It wasn’t unusual in the beginning to find 8%-10% of my ad spend for the day going down the drain. See my last few posts in UP THE CREEK for details on my findings.

If you have a lot of model or part numbers, or any other kinds of numbers in your keywords WATCH OUT! Expanded Match will rob you blind! As Cline pointed out in the “Up the Creek” thread it also likes to play with just words. I’ve found plenty of word games.

I don’t sell Olympus anything and I do not have “Olympus” as a keyword in any of my campaigns, but I had to make “Olympus” a negative keyword in all my campaigns to stop clicks for part numbers and the word "Olympus". To rub it in more, the Olympus part numbers didn’t even match ANY of the part numbers in all my campaigns!

So watch out! You may put a negative keyword a particular campaign one day, only to find the same unwanted keyword (in my case "Olympus") has grabbed another ad campaign hostage to rip off. In the case of "Olympus" I had to make it a negative keyword in all my campaigns to stop it!

If you have a better, faster, easier way than mine for stopping this highway robbery I would like to hear about it. Any help I can get would be appreciated.

The help I would like the most is for Google to stop this abuse of my time and money. Isn’t it enough that we risk our money with Google managing it with Adwords everyday, regardless of possible click fraud, multiple clicks on the same keywords by lazy surfers, outrageous cost per click for competitive terms, and ROI’s that are getting harder to swallow everyday.

How about it Adwords? Are you going to doing anything to stop this ridiculous drain of our time and money! Tonerman

Last edited by tonerman : 03-04-2006 at 06:10 PM.
tonerman is offline   Reply With Quote
Old 03-06-2006   #2
travisk
 
Posts: n/a
Tonerman, you are getting closer and closer to truly understanding the meaning of the words "broad", "negative" and "responsibility". Keep working on it.
  Reply With Quote
Old 03-06-2006   #3
tonerman
Member
 
Join Date: Oct 2004
Posts: 179
tonerman has a spectacular aura abouttonerman has a spectacular aura about
To dumb to understand your point!

I couldn't decide if you were saying that I had the responsibility to stop broad match with expanded match clicks I didn't want with negative keywords, or if you were referring to Google's responsibilities. My inclination is that you are pointing at me.

If you were referring to my getting closer to understanding the words "Broad", "Negative" and "Responsibility", I agree that in the end I am responsible for how my adwords campaigns perform.

However, having to look at every single adwords click query to catch Expanded Match impressions and clicks that do not match any keywords in any of my ad campaigns isn't reasonable either.

The keywords in an adwords account ought to be meaningful and impressions and clicks for "Olympus P11 photo cartridge" ought not occur when you don't have the words "Olympus" or "P11" anywhere in your campaigns.

Please restate your comment so a dummy like me can clearly understand your point. Thanks. Tonerman

Last edited by tonerman : 03-06-2006 at 05:41 PM.
tonerman is offline   Reply With Quote
Old 03-06-2006   #4
rustybrick
 
rustybrick's Avatar
 
Join Date: Jun 2004
Location: New York, USA
Posts: 2,810
rustybrick has much to be proud ofrustybrick has much to be proud ofrustybrick has much to be proud ofrustybrick has much to be proud ofrustybrick has much to be proud ofrustybrick has much to be proud ofrustybrick has much to be proud ofrustybrick has much to be proud of
Tonerman, your post, IMO, is valuable. It shows how much extra work is needed now, as opposed to X months ago, to get your campaign in tip top shape. Being as detailed as you were, I feel should be commemorated.
rustybrick is offline   Reply With Quote
Old 03-06-2006   #5
tonerman
Member
 
Join Date: Oct 2004
Posts: 179
tonerman has a spectacular aura abouttonerman has a spectacular aura about
Thanks Rusty!

I appreciate your positive feedback Rustybrick! Incidentally, I checked on about $15-$16K worth of Google broad match clicks in August 2004 and didn't see much to alarm me. Haven't spent anytime looking at other older log files but I am going to see what I find around August 2005 based on some of my earlier work and Cline's comments in "Up the Creek". Tonerman

Last edited by tonerman : 03-06-2006 at 08:17 PM.
tonerman is offline   Reply With Quote
Old 03-06-2006   #6
ApogeeWebLLC
dork
 
Join Date: Mar 2005
Location: Maryland
Posts: 47
ApogeeWebLLC is on a distinguished road
!broad

Tonerman, have you gathered enough data to eliminate the need for broad match? With all of the log analysis you have done plus sales tracking, perhaps you could switch to only exact and phrase matches? I suppose it's the lesser of two evils: losing sales from clicks for keywords you're not capturing by dropping broad match or losing cash due to useless clicks from using broad match.
ApogeeWebLLC is offline   Reply With Quote
Old 03-06-2006   #7
cline
Aderit Internet Marketing Consulting: www.Aderit.com
 
Join Date: Jun 2004
Location: Boston, MA
Posts: 146
cline has a spectacular aura aboutcline has a spectacular aura about
Switching to exact match causes too much traffic to be lost. Tonerman is implementing the best solution: lots of negative keywords.
cline is offline   Reply With Quote
Old 03-06-2006   #8
tonerman
Member
 
Join Date: Oct 2004
Posts: 179
tonerman has a spectacular aura abouttonerman has a spectacular aura about
You are exactly right Cline!

I got killed trying to quickly create an effective all exact and phrase match campaign. I worked my butt off trying for two weeks, but you need a gazillion phrase and exact match keywords to get the same amount of traffic broad match will give you. It is a shame that expanded match is so poorly implemented now because it worked perfectly as the original matching option.
tonerman is offline   Reply With Quote
Old 03-06-2006   #9
ApogeeWebLLC
dork
 
Join Date: Mar 2005
Location: Maryland
Posts: 47
ApogeeWebLLC is on a distinguished road
Did you try phrase match even on single keywords? For example, instead of a keyword list like:

1250w
minolta 1250w
minolta 1250w toner
minolta 1250w cartridge
minolta 1250w cartridges
minolta 1250w toner cartridge
minolta 1250w toner cartridges

Use exact and phrase even just on 1250w:

[1250w]
"1250w" <-- this is perhaps the most useful thing to try
[minolta 1250w]
"minolta 1250w"
"1250w minolta"
[minolta 1250w toner]
"minolta 1250w toner"
[minolta 1250w cartridge]
"minolta 1250w cartridge"
[minolta 1250w cartridges]
"minolta 1250w cartridges"
[minolta 1250w toner cartridge]
"minolta 1250w toner cartridge"
"minolta 1250w toner cartridges"
[minolta 1250w toner cartridges]
"cartridges for minolta 1250w"
"cartridges for minolta printer 1250w"
"cartridge for minolta 1250w"
"cartridge for minolta printer 1250w"
"1250w toner"
"toner 1250w"
"1250w cartridge"
"cartridge 1250w"
"1250w cartridges"
"cartridges 1250w"
...

I realize you'd need an awful lot of permutations to capture most of the long tail. However, how is this more work than adding negatives? Since you have a year or more of log data, why not parse through it to find positive combinations rather than negative keywords? With heavy use of phrase matching, this might be a better solution. You'll never keep up with negatives. I know. For one of my clients, I have a negative keyword list of ~500 keywords. I've been switching to more exact and phrase matches as new negatives crop up everyday. It's simply not economical.
ApogeeWebLLC is offline   Reply With Quote
Old 03-06-2006   #10
tonerman
Member
 
Join Date: Oct 2004
Posts: 179
tonerman has a spectacular aura abouttonerman has a spectacular aura about
Phrase Match verus Broad Match with Negative Keywords

Your point is well taken. As a matter of fact for the term "1250w" and all of its permutations with other keywords, I have 1186 active keyword phrases for the 1250w ad campaign alone.

You are also right that I have over 4 years of log data. I pulled all the queries from clicks that resulted in orders for all my brands, models, and products. Then I separated them into brand, model number, part number, and type part. I sorted them all within each brand, eliminated all the duplicate queries and then replaced each word in every query where appropriate with the words "model, P#, brand, etc. I left the pure english like it was - for instance "drum for" was changed to "type for."

Each manufacturer has their own lingo on their product boxes. For one brand it is a "module", for another it is a "cartridge". Then there are part number variations: "1250 w" or "1250" or "1250w".

Knowing the structure of every query made for 4 years and then designing a template that let me do a search and replace for the place holders I created for "Brand", "Type", "Part Number" etc resulted in the end to 1186 phrases.

That was the amount of work done for one product! We sell hundreds of products in hundreds of ad campaigns. There simply isn't enough time to put in the amount of time and effort needed for all of our products and ad campaigns to switch to phrase match overnight.

In my case I had lost my temper with Broad Match one day and flipped all my keywords and campaigns to Phrase Match without thinking this decision through. My traffic dropped 50% overnight and after two weeks of that disaster I reverted everything that hadn't been converted in to phrase and exact match back into Broad Match. Our sales had taken a beating for two weeks and we couldn't absorb another bad week. The cure was worse than the disease!

If you have the luxury of time you can put some very good phrase match campaigns together. However, the best phrase match campaign in the world is unlikely to do as well as a good broad match campaign with the right negative keywords.

Both of these approaches cost time and money. However, the return in higher traffic and conversions from Broad Match will over-shadow phrase match everytime.

That doesn't change one negative thing I and many others have said about how poor Google's Broad Match with Expanded Match option works today - it's just my experience in trying to deal with it and I wanted to share it with others.

If Travisk meant in his comment about my beginning to understand the meaning of "Broad", "Negative" and "Responsibility" that when you use Broad Match that you have the responsibility to determine what negative keywords you need to have in your campaign I agree with him. Anyone using Broad Match as originally implemented understood that creating any needed negative keywords to prevent unwanted impressions and clicks was the advertiser's responsibility - not Google's.

However, when Google opens up matching to include keywords that aren't within miles of being relevant or related to your broad match keywords (for instance "Olympus P11 Photo Cartridge" when you do not have a single campaign with the keywords "Olympus" "P11" and "Photo Cartridge" in any broad keyword anywhere). then I think it is Google who needs to undertstand the meaning of the words "Broad", "Negative" and "Responsibility".

Last edited by tonerman : 03-07-2006 at 12:11 AM.
tonerman is offline   Reply With Quote
Old 03-07-2006   #11
cline
Aderit Internet Marketing Consulting: www.Aderit.com
 
Join Date: Jun 2004
Location: Boston, MA
Posts: 146
cline has a spectacular aura aboutcline has a spectacular aura about
tonerman, have you noticed any improvement in your overall account performance since you have detected the broadmatch problems and installed albeit imperfect solutions?

It would occur to me that you are more aware and sophisticated in this regard than your competitors, and after a few months this targeting advantage should be giving you a better CPA than what your competitors achieve, allowing you to bid higher and get more traffic, in turn allowing you to simultaneously increase both gross revenues and gross margin.
cline is offline   Reply With Quote
Old 03-07-2006   #12
tonerman
Member
 
Join Date: Oct 2004
Posts: 179
tonerman has a spectacular aura abouttonerman has a spectacular aura about
Good suggestion for improving my results.

I've implemented Google Analytics (although it isn't perfect for my site yet), and I have my own unique and highly sophisticated means of accurately tracking all initial referrer PPC and organic queries and conversions.

I have been so head down and locked getting both my traffic and conversions back on track from my angry decision a few weeks ago to flip all my campaigns to phrase match from broad match (without considering the consequences, or having created new keyword phrases in all the different query structures needed) that I haven't had the opportunity to take advantage of whatever overall improvement in performance I've experienced from tightening things up.

I'll take a look at some of my more critical campaigns and keywords and try your suggestion of raising the bids and seeing how it effects my ROI. Thanks for the suggestion. I appreciate it. Tonerman
tonerman is offline   Reply With Quote
Old 03-07-2006   #13
cline
Aderit Internet Marketing Consulting: www.Aderit.com
 
Join Date: Jun 2004
Location: Boston, MA
Posts: 146
cline has a spectacular aura aboutcline has a spectacular aura about
Tonerman, I would not act too quickly. When I observe -- based on conversion data -- that I'm getting a targeting advantage over my Adwords competitors, I slowly raise bids. Rapid bid increases tend to cause bid wars, which hurt everyone. Slower ones let the other bidders evaluate their profitability and step aside, as their results warrant.
cline is offline   Reply With Quote
Old 03-07-2006   #14
tonerman
Member
 
Join Date: Oct 2004
Posts: 179
tonerman has a spectacular aura abouttonerman has a spectacular aura about
Got it!

Quote:
Originally Posted by cline
Tonerman, I would not act too quickly. When I observe -- based on conversion data -- that I'm getting a targeting advantage over my Adwords competitors, I slowly raise bids. Rapid bid increases tend to cause bid wars, which hurt everyone. Slower ones let the other bidders evaluate their profitability and step aside, as their results warrant.

Thanks - I will act accordingly! Tonerman
tonerman is offline   Reply With Quote
Old 03-07-2006   #15
tonerman
Member
 
Join Date: Oct 2004
Posts: 179
tonerman has a spectacular aura abouttonerman has a spectacular aura about
Angry Perfect example of why you must check Google PPC queries everyday!

I did my usual check on Google PPC queries today and found two clicks from the same user for the term "micr ink for hp 832c".

I don't have 832c anywhere in my campaigns. PPC Cost $7.06 (5% of my ad spend so far today. Of course the landing page didn't have this product. We sell laser TONER cartridges - the HP 832c is an INKJET printer. It is a DESKJET.

I have DESKJET and INKJET as a negative keyword in all my campaigns. The 832c is a Deskjet. Expanded Match can't connect the dots. This mismatch came in out of the bluesky troday.

What in H--- can we Forum members do to stop this kind of unbelieveable rip-off? If Google shut down this kind of revenue stream across all Adwords accounts their stock would probably go down like a rocket. Not much incentive for Google to stop this abuse is there?
tonerman is offline   Reply With Quote
Old 03-07-2006   #16
cline
Aderit Internet Marketing Consulting: www.Aderit.com
 
Join Date: Jun 2004
Location: Boston, MA
Posts: 146
cline has a spectacular aura aboutcline has a spectacular aura about
Sure, the $7 is a waste, but at least you can figure out how to stop it. Your competitors probably did not find this. You've now gained a competitive advantage.

I manage *a lot* of Adwords. While I do consider this to be a big problem, it's a problem that in the overall mix doesn't happen much. When it does happen, it can be horrible, but that's infrequent. Whatever Adwords does with expanded broadmatch will hardly be a blip on the revenue chart -- for Adwords.

Google itself will probably address this problem when they realize how bad it is making their search results look. People come to Google for relevant search results. This harms relevancy. No Yahoo Search Marketing editor would allow this kind of error to be made. Google will figure out that they have to leave a bit of money on the table to protect their reputation.
cline is offline   Reply With Quote
Old 03-07-2006   #17
tonerman
Member
 
Join Date: Oct 2004
Posts: 179
tonerman has a spectacular aura abouttonerman has a spectacular aura about
MAybe you are right Cline! Perhaps Expanded Match is a blessing

Hadn't thought about this being a competitive advantage much. Maybe I should keep my mouth shut and take advantage of my early detection and quick shutdown of mis-matched queries and hope Expanded Match screw-ups accelerate as fast as possible!
tonerman is offline   Reply With Quote
Old 03-08-2006   #18
tonerman
Member
 
Join Date: Oct 2004
Posts: 179
tonerman has a spectacular aura abouttonerman has a spectacular aura about
Uh Oh! I've been thinking Cline.......

I've been mulling some of your comments and insights as well as my experience and I'm considering the advantages for me in Broad Match as a primary matching option.

It takes a year and a day to construct Phrase Match and Exact Match keywords and keyword phrases and I've got a lot of campaigns. Take the term "1250w". I replied further up this thread that I have 1186 phrase match keywords in that campaign right now. Suppose I dumped them all and went with three Broad words: "1250w toner". "1250W ink" and "1250w cartridge". I haven't thought about all the negative keywords yet for this post, but that's no big deal - I can gin those up pretty quickly.

What would be the pros and cons of this strategy Cline? Sure, Expanded Match will ding me a few times, but I think I know from prior data where it is going to screw up and start popping impressions right off the bat, so I would nix those words with negatives from the start). I'd have to stay on top of the campaign with my query checking routine, but the increase in traffic and conversions with this strategy for an important campaign might make it worthwhile.

I await your thoughts - mine are usually half-baked!

Last edited by tonerman : 03-08-2006 at 12:57 AM.
tonerman is offline   Reply With Quote
Old 03-08-2006   #19
cline
Aderit Internet Marketing Consulting: www.Aderit.com
 
Join Date: Jun 2004
Location: Boston, MA
Posts: 146
cline has a spectacular aura aboutcline has a spectacular aura about
Your keyword strategy depends on your situation. If you are spending excessive (on an ROI basis) time doing keyword research and managing your campaigns, it may make sense to do that. However, subtle differences in search term phrasing can create large ROI differences, and these differences can inform bidding strategy.
cline is offline   Reply With Quote
Reply


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off