Maximum negative matches

Is 40 too many?

[keithh]

I recently experimented using a general keyword with numerous negative matches. I started with around 100 negatives and found the ad did not appear. It appeared when I reduced the number of negatives to 40. Adwords tell me that there is no limit to the number of negative keywords, but this seems to be not true.

Has anyone else experienced this problem?

Keith

[eWhisper]

There isn't a maxium beyong what a campaign can realistically hold to the number of negative KWs you have. I have several ads with over 350 negatives.

If you have a lot, (and I know this sounds simplistic, yet it can happen), make sure your term isn't negatived out (i.e. KW blue widget, -blue), and that your search doesn't contain a negative KW.

[keithh]

I have not negatived any of the keywords. Some of my negatives could possibly be trademarks. Is it possible that the ad could stop displaying because of this?

[eWhisper]

That's a good question. Does G count negative keywords as keywords for the purposes of ads not being shown due to trademarked violations.

Think this is one that AWA needs to answer when he gets back from the G conference.

[Robino]

Try going to the next page of keywords. I think it only displays 30 per page.

[AdWordsAdvisor]

...Does G count negative keywords as keywords for the purposes of ads not being shown due to trademarked violations?

Think this is one that AWA needs to answer...

I've checked with the policy folks who have confirmed my '98% certain' thought: it is not a problem to use trademarked words as negatives - even those which the trademark owner has prohibited others from using.

AWA

[hobbnet]

Why would you want a ton of negative keyword matches? I would think that if you had 350+ negative keywords that using an [exact match] or two for your keywords would be a much simpler and less time consuming process.

But maybe eWhisper can set me straight. ;)

[eWhisper]

But maybe eWhisper can set me straight

:)

Well, here's my attempt. (This post got a lot longer than I expected - but at least it's detailed :) )

I'm going to use a real life example from one of my clients and just change all the kws. (I can make many examples like this as to why broadmatch can be effective - but no one wants to read a 12 page post...).

In this example, the client sells widgets in several states, but it's a physical store and not ordered through the web. This widget is also rare, so people are willing to drive 2-6 hours to buy one - and in some cases will even fly in for one.

It seem that bidding on widget ca, widget california, etc for exact or phrase matches would be the way togo.

However, it turns out for this particular widget, 35% of people search for: city widget feature, or state widget feature (not necessarily in that order, it varies a lot)
(feature can vary, and I've seen about 300+ different features so far that people have used in search strings).

As those states contain thousands of towns multiplied by hundreds of features, we'd be looking at 2-5 million keyword combinations using phrase or exact match. This is not a feasible number to work with.

Because most people who search at G are not identified by regional targeting, they use regional targeting to some degree - but can't rely on it for maximum exposure.

So now we have to figure out how to reach these 35% of people using some keyword combination since we can't use 5 million exact/phrase keywords.

Using broadmatch for widget features, widget state, widget city (for the big ones - you just can't do that for every town of 1000 people), and then using negative kws for every state and the major cities they're not in (plus some others random words based on referral logs) is the way to go.

Yes, they do get some referrals for widget feature city not in their target audience - but these are down to only 10% of their total referrals. While this number may seem big, the cost is covered by the increase in sales by reaching that 35% of the seach market that they can't reach with exact/phrase match.

I am a fan of broadmatch, although, for some companies it's often not worth it. You need to weigh the benefits and negatives of using the system - but in a lot of cases - broadmatch can work very effectively.

[figment88]

Why would you want a ton of negative keyword matches?

To buy clicks at a fraction of what the competition does.

[skibum]

If you've got a client that sells real estate in 100 cities/counties in 15 states, you'd use lots of negative keywords (in the hundreds if not thousands) to knock out all the places where they don't have inventory. Likewise mortgage companies are sometimes not licenced in all states. In order to promote one that operated in only one state and cast a full net w/o getting irrelevant traffic, a geo targeted campaign combined with a nationwide campaign with tons of negatives would would provide the best reach.

[hobbnet]

I guess im partial to exact/phrase matching even if it comes to having thousands of keyword terms that geotargetted campaigns can demand. I feel like I have more control in this situation. But there is obviously a reason in some cases to have a 300+ negative keyword list. Thanks eWhisper ;)

[spaidermen]

How do you put multiple negative kw?
I put something like tennis shoes -adidas -nike in the same row but this does not work, error message is "put negative keyword in a different line"

[sem4u]

"put negative keyword in a different line"

That will be the answer then :)

tennis
-elbow
-court

etc.