Watch Spending When Using Google Broad Match

NOTE ADDED 2/23/07: This thread also applies to broad matched keywords that have NOT been recently paused or resumed.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Be sure to watch your spending after pausing some ad groups, keywords and campaigns at Google. If you have broad matched keywords anywhere in your account, Google's system may decide to display other ads for the keyword phrases you don't want running.

Here are some reasons why that is not a good thing:

1. The ads displayed will be less targeted and so clicks on them will be less likely to convert, increasing ad spend and decreasing profits.

2. The keywords Google broad matches are likely to be those with the most expensive bids in your account. Instead of paying the CPC on the relevant ads you paused you may end up paying two or three times more for clicks on impressions you intentionally did not want.

If your advertising costs are higher and conversion rates lower for a particular product or ad group and that group has the highest average CPC it is likely that this issue is greatly increasing your spending.

Adding more negative keywords can help block this behavior. If you advertise multiple brands of the same products, consider adding all the other brands as well as any popular brands you do not sell as negative keywords.

Also consider deleting high cost, high impression broad match keywords and phrases. They are the most likely to be broad matched on unrelated searches. If you use a brand name alone (without the product description) consider using only the exact match version.

Remember that if you block Google's system from matching to one keyword phrase it will likely go on to the next highest cost keyword phrase. You can make continuing improvement in ad spending by progressively reducing impressions on your highest cost keywords. This situation also occurs at Yahoo but is easier to control there.

You can change your highest cost keywords to standard match and leave your lower cost similar keyword phrases set to advanced match. The default in Yahoo's new system is advanced so to control costs there you may want to manually set higher cost keyword phrases to standard.

Even if you have other similar keyword phrases set to advanced, changing to standard match can reduce impressions for your ads because your bids may be less competitive. You'll have to experiment to determine the right mix of bids and match type to control costs without reducing traffic and sales.

You may want to set more keywords to standard for keywords with a high cost per conversion and keep those with strong profits set to advanced. Remember that at Yahoo's system too, any changes you make affect other keywords.

PPC accounts are like a balloon...when you squeeze bids one place the "impressions" may move elsewhere. When you pause ads you think you stopped that spending (let the "air" out of the balloon) but the PPC engine may "help you out" and keep that spend going.

[ Message was edited by: flyingrose 02/23/2007 03:19 pm ... Reason: Add First Sentence to Text ]

We are experiencing the same problem. For years, one of the terms we have bid for was a highly specific but infrequently searched for term. Since these terms were so specific to our business but infrequently searched for, our max bid was around $2 per click. Our impressions averaged around 100 per day. Beginning in early January, this term increased from 100 impressions per day to 1,200 impressions per day. Never before has this infrequently searched for keyword received over 150 impressions in a day. This impression surge continued until early February when we pulled this term offline.

Google has changed something on how it implements the broad match/expanded match criteria, which has effected this term. Oddly enough, only one keyword has been affected.

We have attempted to contact Google on several occasions in an attempt to rectify the situation. Google will not reveal number of searches for this keyword (only the number of impressions) or a list of the expanded terms for which this keyword is displayed. Our company has been an AdWords advertiser since it was introduced years ago. We have spent tens of thousands of dollars in advertising with Google and used the 'broad match' option for our campaigns until now.

Google has only offered us 'canned' responses with hypothetical explanations on what may have caused the surge in impressions. Of course, Google profits from the change they made while bleeding our company of advertising dollars.

So, do you suggest that people opt out of broad match and make sure that there's quotes around phrases? Would that be better?

Thank you for posting onlineguy. I have a question regarding your post. You mentioned bidding around $2 per click and a surge in impressions but not if those impressions were generating clicks. Are you paying for impressions on that keyword or are you paying per click but also seeing increased clicks and costs?

I have long suspected that about once per quarter Google would merge keywords that have great CTRs with a more general, higher cost keyword. That totally changes how ads perform.

Keywords that were generating only targeted clicks and strong conversions may suddenly get almost no impressions as their position is outbid by all the new competition OR if your bids are set high suddenly get tons of impressions, possibly far more clicks than normal, and conversion rates drop substantially.

Those who have had either conversion tracking or analytics running for some time may be able to see this by checking historical data.

Then as your ads stop displaying for broad and phrase matched phrases they don't apply to your impressions and clicks gradually drop and your conversions gradually improve.

There is no simple answer to that Bill. All I can say is "it depends" and unfortunately it depends on so many variables there is no easy way to determine what is best.

I do know from tracking I've done in the past that if you do thorough keyword research and have all the searched for variations entered consistently using broad, phrase, and exact match that about half of all sales are generated by the broad matched versions and half by the phrase and exact matched combined.

That tells me that to eliminate broad match would cut sales in half. For advertisers who do a less thorough job than I do on keyword research they'll cut sales by even more than that.

I do delete broad and phrase match on some words such as just the brand name alone (with no other keywords indicating the product) to force Google's system to display the targeted broad match ads I want and not a generic brand related ad.

I also use many negative keywords in the more general ad groups to force Google's system to display the product specific ads I want and not more general brand only ads.

Recently I saw a serious issue where Google chose to broad match ads with the highest bids for a totally different brand and product. THAT is NOT funny and I hope it was not something that is happening frequently.

My client happened to do a search to see if I had paused a brand he decided to stop advertising and saw that the targeted ads that used to display were not there but an ad costing three times more was.

To provide an example of what I mean, it would be like seeing an ad for a Chevy Corvette when someone searched for a Ford truck. Yes, they're both vehicles but you don't want your ads being displayed that way - especially if the Chevy Corvette bid is three times the Ford truck bid!

I've just encountered this issue again in a different account for a different industry. This time I can see that the change in Google's system occurred on Sunday, February 4th.

Historically this keyword phrase consisting of a brand name and the general product description has received less than 30 impressions per day and had a consistent CTR of over 1.5%.

This is easily seen by comparing any individual day prior to 2/4/7 with any day since that date or by comparing statistics for January 1-31 and February 1-21. I have included the statistics for this specific example below.

Clicks went from less than ½ click per day to over 10 per day – an increase of 2400%

Impressions have increased from less than 27 per day to over 1146 per day – an increase of over 4000%

CTR dropped from 1.5% to 0.88%

It is obvious that this keyword phrase has been broad matched to one or more expensive, more general phrases because the CPC jumped from $1.14 in January to $1.74 since February 4 even though the average position has dropped from 4.9 to 6.7

The 13 clicks and $14.79 spent in January generated one sale while 212 clicks in the past 21 days generated no sales and cost $362.68.

Spending went from $14.79 (less than $0.48 per day in January) to $362.68 ($17.27 per day). Spending increased by over 3,590%!!!

Worse yet, where we had one conversion in January for a cost per conversion of $14.79 so far in February conversion counter shows 0 sales and Google Analytics shows one sale for that $362.68 spent. We can hope additional sales come in, but this is clearly not something we want to continue. I’ll post this and then add the fix for this problem in the next post.

Here is a way to diagnose where that huge jump in impressions came from and a plan to correct it:

Go to the Google Keyword Tool. This tool is currently the top left link on the tools tab in Google Adwords accounts.

Type in the keyword phrase with the issue. You'll probably see that it has been broad matched to either unrelated brands in the same industry or very general product terms or both.

Enter all of those unrelated brands as negative keywords either at the campaign or ad group levels. Also enter any general keywords you can use without blocking the ads you DO want to show. Many times you cannot block the general term without blocking your specific phrases. For example:

-widget would block the general search but also your specific brand phrases that include the word widget such as "brandname widget" or "widget company" or "ABC widgets".

The first time I ran into this I searched for every brand name I could find and added negative keywords for every one of them to every campaign. What isn't biting you today is likely to bite you on the next Google match-fest.

I also added negative keywords in every ad group for other products. For example, if you were selling Chevy's you'd add negative keywords to each ad group for every model that ad group is not specifically targeting.

The bad news is if you don't want your specific ads to show up for that general search you'll have to delete every broad matched phrase their tool indicates is matched to that keyword.

Worse news is if you block the ads from showing for that specific ad group they're likely to show up on the next most expensive broad-matched keyword phrase in your account so you have to do this for every brand and every product.

Worst case, advertisers who have limited budgets and/or tight margins may be forced to delete all broad matched keyword phrases and lose half their traffic in order to control costs.

For those using Google Analytics, one way to catch these types of issues and determine which keywords to check on is to click on Marketing Summary under Dashboards and check for jumps in traffic to your Top Five Keywords.

If you fix the top offenders check the report again after 8-9 hours to see if other keyword phrase(s) suddenly show huge jumps. Keep doing that until your account is stable again.

Here is a good way to determine if the visitors this traffic is driving are at least actually looking around when they get to your site:

Go to All Reports, Marketing Optimization, Visitor Segment Performance, Referring Source

Locate google[cpc] under Source [Medium]

Click the "Analysis Options" icon (the circle with the double-upward arrows)

Select Cross Segment Performance, Keyword

Look at the P/Visit (average pages per visit) number.

Although that number can vary depending on your site design and product mix you want to see at least several pages. If your average is 1-2 that could indicate either the traffic is not relevant to what you're offering or possibly click fraud.

Dear Google Adwords: I would like to convey this message:

When someone makes a brand specific search they are seeking that particular brand. Showing them less targeted ads cuts CTRs in half, can greatly increase spending as in this example, and be devastating to conversion rates and an advertiser's cost per conversion.

It is unwise to broad match to other brand names and general keywords. A sophisticated advertiser could research and add all other brand names as negative keywords if they can diagnose the issue and know how to do that. It is likely the majority of advertisers do not have those skills.

Matching to general keywords is even worse. An advertiser cannot eliminate these generic searches with negative keywords because they would also block the specific phrases for which they do want their ads appearing. This forces advertisers to either delete broad matched phrases altogether, cutting their traffic and spending with you in half or pay more for advertising than is profitable.

Bad broad matching is worse than no broad matching. It is bad for the advertiser, bad for the Internet user, and bad for Google's revenue.

P.S. We could sure use an account level negative keyword function - especially if we're going to have to add every brand we don't advertise (usually because we don't sell them).

[ Message was edited by: flyingrose 02/22/2007 01:28 pm ... Reason: Add P.S. ]

Google has implemented major changes and there are reports about all kinds of issues. It would be wise to monitor the threads here plus the Official Google blog and the SEO Round Table Google related blog for issues and announcements.

One other tip that may head off this issue in your accounts. If you're checking recommended budgets (which I recommend you do regularly) and a campaign suddenly wants a much larger budget consider that a potential red flag that some of your keywords may have just been matched to much broader, high traffic phrase(s).

I'm deleting more and more broad matched keywords. What I look for is climbing impressions, high average CPC, poor CTRs, and cost per conversions that are too high.

Check for your ads triggered by keywords with the highest bids being displayed for REALLY general searches. Here are my strategies for advertisers who want to keep advertising costs tightly controlled when dealing with this issue:

1. Consider deleting broad matched keyword phrases that have CTRs lower than about .7 and definitely those under .4 to .5

2. Add all brands you do not sell as negative keywords.

3. Add negative keywords for the other brands you do sell in campaigns/ad groups for all other brands.

4. If you're running ad groups for general product lines plus specific products, add the specific products you're targeting to the general campaigns/ad groups.

5. To keep Google from showing your highest cost ads for general searches, create ad groups for really general words that apply to your products and set low bids and budgets on them.

[Fantasy Wish Request: I wish Google and Yahoo would coordinate their schedules to stagger the emergencies they're causing so I can keep up with them and maybe get more sleep too.]

Hi FlyingRose,

I am sorry for responding so late. We have been swamped with Adwords and other issues.

To answer your question about my clicks increasing, the answer is yes. Our impressions increased, and therefore our clicks increased.

We had our budget set fairly high so our keywords would be shown throughout the day. However, this was based on the fact that we knew what impressions to expect for the previous few years. Then, something changed with Google.

In the past, we never really used many tracking methods. We only used our Web stats package, which is very limited. This week we will be implementing a new tracking system.

We are now testing out Google Analytics and others. We really want to be able to find out what keyword people used when they clicked on our Ad. Does anyone know of a package that can do this? It doesn’t look like Google Analytics will do this?

Thanks all!

I understand being swamped. I still haven't been able to resolve the issues Yahoo's migrations are causing and every day yet another account gets forced into that challenge and now this.

The Google issue is an even more serious crisis than the Yahoo situation because of the potential it has for enormous unpredictable spending jumps. My clients and every small business I've ever helped with their PPC accounts is at extreme risk right now.

I don't have enough hours in the day to protect them all unless I opt every one of them out of the Google search network and pause or delete every broad matched keyword to protect them.

Since doing that may cut their traffic and sale revenue in half that choice is difficult to sell as a preventive action. If I don't do that and they get hit with a spending spike they'll be a lot more unhappy than if I do it.

FlyingRose,

What tracking system do you use?

As mentioned, we want to track the "broad keyword" that people search for. For example, if we bid on the term, 'widget', we want to know if the person who clicked on our Ad searched for 'big blue widget' or 'small red widget'.

Do you know of any systems that will allow us to do this?

Thanks again!

I just saw your latest question. I don't know of any current way to determine what the actual search was and tie it to any particular Google broad match displayed ad. I have asked Mike Levin to check into what would have to be done to do that.

You can use his program discussed in the My Long Tail thread to find the keywords used to access your site.

I also use the Monitus Keyword Monkey feature in accounts that use his Yahoo Store connector tool.