Learn How to Control Your Negative Keywords
Negative keywords help you control when your ads will show. If a search query contains your negative keywords, then you ad will not be displayed for that result.
There are two most common ways of using negatives:
- To stop your ad from showing on a query that never converts
- To make sure the PPC engine shows the correct ad when you have multiple that could be displayed
The first example, stopping your ad from ever showing, is self-explanatory.
The second requires an example. Let’s say I have these ad groups; and each ad group contains only one keyword – the ad group’s name.
- Tennis shoes
- Nike tennis shoes
- Red Nike tennis shoes
A search for ‘red Nike tennis shoes’ could trigger an ad from any of those ad groups.
- The query contains ‘tennis shoes’ so the tennis shoes ad group could show
- The query contains ‘Nike tennis shows’ so the Nike ad group could show
- The query is red Nike tennis shoes; so the red Nike tennis shoes ad group could show
While Google likes to show ‘most restrictive’ match; if one ad has a much higher ad rank for the other – that will often trump most restrictive match.
Diagnosing Ad Serving Problems
The easiest way to diagnose if you have multiple ad groups showing for the same keyword is to:
- Run a search query report
- Download the data and put it into a pivot table
- Sort the pivot table by search terms and number of ad groups per search term
In a perfect world, you will only see one search term per ad group. This means you have no cross over in your terms between ad groups and you have completely controlled the search queries. However, if you see that your search terms are being displayed from 5, 10, or more than 100 different ad groups – then you have ad serving issues.
In this case, you need to use negatives at the ad group level to control your ad serving. If you are are having ad serving issues due to using multiple match types across ad group, then please see this article on organizing by match types.
Ad Group Negatives That Should be Campaign Negatives
Another issue I see is that an account will have hundreds of thousands of negative keywords at the ad group level; however, those negatives are really meant to be global negatives. If you examine your negative keywords by ad group and see a very consistent number; or very similar negatives at the ad group level (that are not being used to control ad serving by match types); then you should move these from the ad group level to the campaign level:
Campaign Negative Keywords
Campaign level negative keywords stop any ads from showing for that keyword in any ad group within the campaign. These are reserved for words that you just never want to show for. If you have a lot of campaigns, I often see another issue arise.
Most people do not do negative campaign level keyword research on a regular basis. Most account managers do negative campaign level research in one of two ways:
- Add negatives at the ad group level instead of the campaign level. This leads to the problem that if you find a word you never want to show for that if you add it to just an ad group; that same query will often show up in another ad group. Making it a campaign negative instead stops that problem from arising.
- Campaign negative research is not done on a regular basis. Often its done when someone thinks about it, or every few months. This caues someone to get out of the habit of controlling their negatives well.
For instance, this is an account that has thousands of negative keywords at the campaign, level. However, in reality there are really only 551 or 229 (you can tell the days they did the research by such exacting numbers):
In the case where you have global negatives you never want to show for; but you don’t want to go through the time consuming task of adding it to every campaign on a regular basis; there is a better solution: Negative keywords lists.
Using Negative Keyword Lists
A negative keyword lists is a list of words that you can save together. You can then apply this list to one or more campaigns.
The real advantage of using a list is that if you find a new word you never want to show for, all you have to do is add it to the negative keyword list and that word will automatically be applied to every campaign using that list.
This is a huge time save.
So, if you are looking through search query data and you see a keyword that you never want to show for – do not add it to the ad group – add it to a negative keyword list. This single change in how you add negatives will save you a lot of time as you won’t have to find that same keyword again in a different ad group to add it again; and you don’t have to add it to each campaign. You add it to a list, and as long as that list is applied to all the appropriate campaigns – then you no longer need to worry about how much time it takes to add negative keywords.
Wrap-Up
Spending time on negative keyword research is essential. Adding negative keywords should:
- Increase your CTR (you aren’t showing for bad impressions)
- Lower your CPA (you won’t pay for non-relevant clicks)
- Control your ad serving (which will help with both CTR and CPA)
- Increase your conversion rates (again, won’t show for non-relevant clicks)
However, most companies have an ad hoc way of approaching negative keyword research. There really is a simple way to do it:
- Create one or more negative keyword lists
- Apply each list to the appropriate campaign
- Examine search query data
- If you see a query you don’t want to show for in just that ad group; make it an ad group negative
- If you see a query you never want to show for, then add it to the appropriate negative keyword lists
This process is pretty simple; but it will save you a lot of time and money by just creating a simple negative keyword review process an sticking to this simple workflow.
Great article, Brad. I faced the problem of search term spills as well (as have many others) and actually I have developed a similar technique described here: http://www.adwiserhq.com/blog/are-you-the-master-of-your-search-terms/
However, as I am sometimes using three match types of the same keyword in one ad group, I am fine with up to a count of three occurrences of that search term per ad group (one per match type).
And that prompted me to add the ad group name to my pivot table, so I could see if in fact a count of 2 or 3 actually means two or three different ad groups in which the search term appears or the same ad group for several match types, in which case no action was necessary, as the search term was matched with the same ads, regardless of match type.
Nice post 🙂 glad to see someone else is using this type of an analysis.
There’s an easy way to work around that – just don’t include match type in the pivot table. If you don’t; then the pivot table will treat the match types as the same since the formatting isn’t in the row with the keywords. The exception being modified broad match – so you could just do a quick search/replace and remove all the + signs.
If you’re segmenting the ad groups by match type, then this does become a bit more complex as you will naturally have the same keyword in multiple ad groups and will have to look for keyword/match type combos that appear more than once.
For small accounts, stopping the bleeding isn’t too heard. For large ones, we’ll usually accept 2-3 ad groups showing for ‘unusual queries’ and also add a filter that only show queries with over 100 impressions or some number based upon the traffic.
It’s when a large account starts showing in 5, 10, or in some cases more than 100 different ad groups that I think something really needs to be done to fix the issue.
Actually, when I saw counts of 140 ad groups for the same search term in your post’s pictures I thought “wow, I don’t envy the guy who has to add the negatives there” :).
In my case, due to (I like to think) pretty tight matches and thorough organization before the start of the campaigns (or before adding new ad groups), I think I got a maximum of 5-6 search terms that got to maybe two ad groups instead of one. Once maybe three ad groups. So for me having the match type in the pivot table (thanks for the exclusion tip) did not mean too much work.
But 100 ad groups getting their ads triggered by the same search terms can only mean poor setup AND deep pockets, IMO.
That account is for a company that has about 5 million unique negative keywords (most campaigns have at least 500k), 20 million active keywords, and multiple accounts linked together to house it all.
While its a decent account, the reason I was working on it is to help them clean up the ad serving mess between all these accounts.
So, it is a bit extreme of an example, but I’ve seen more than a dozen accounts in the past 3-4 months that have at least 100 terms pulling from at least 5+ ad groups. Hence the inspiration to finally write about controlling negatives more.
So it’s in the “huge” rather than “big” category :). I assume that at this size there’s bound to be some serious overlap, as nobody inserts such a high number of keywords manually.
In my case I wrote the article after seeing the wrong ad in the Ad Preview Tool and trying to find a way to make sure, without checking on all my keywords, if I get ads from different ad groups for certain keywords. Plus some complaints from some other advertisers regarding the “wrong ad showing for my keyword”.
I never thought of using a pivot table for an ad serving audit – great idea!
Since Brad mentioned the account’s large size in the comments … does anyone have a methodology for plowing through a high volume of search queries?
I work in a B to C vertical and a search term report just for the past 7 days is 11K search terms … when I filter this report to remove exact match and exact close variant and remove any search queries that didn’t get clicks* or conversions, I still have almost 6000 unique search queries to review.
Most of the search queries are fine from an intent standpoint so I don’t have a justification at this point for turning off BMM – plus I’m still in discovery mode to a certain extent and trying to scale volume since these are ROI positive campaigns.
Let me know if I should move this comment/question to a forum.
*I know search terms that are getting impressions but not clicks are going to mess with my CTR and ultimately QS and CPC – I inherited these accounts recently and I’m focusing right now on stopping the bleeding.
@caradebeer In that case, what I’d probably do is just create a few segments that you can easily filter on and make decisions:
More than x clicks and x conversions > add as exact match.
More than x clicks and 0 conversions (or CPA above X) > is keyword showing from correct ad group? If yes, make it a negative. If no, then put it in the correct ad group. Don’t add all the negatives at once. Put them in a file to see if there are commonalities (such as there are really only 4 words that are root words and you just don’t want to show for any of them) and make those negatives.
That’s not going to be perfect, but it should help you to make some decisions at scale while you get things under control and then revisit the negatives.
Thanks so much – that method is going to work for me. It actually only took me an hour or two to go through the filtered keywords manually but it’s not something I can count on having the time to do. Appreciate the response!
Hi Brad, I like the idea of using negative keyword listx but lists cannot be moved to multiple accounts. So for now, I just have home-made negative lists that I paste from account to account via Adwords Editor. It is strange that the feature do not allow negative keyword lists at the MCC level
@trenoult I’d really enjoy negative keywords lists at the MCC level. They could make them as a template that can be applied.
Of course, I’d also like account level negatives as well for search (might not apply to display). I find that almost every account has some words that just don’t make sense for the account at all.