2013 PPC resolution: Break up with Google Broad Match
If you haven’t already broken up with Google broad match in 2012, it’s time to make that resolution for 2013! In April 2012, Google released a statement that defined the new matching behaviour for phrase and exact match keywords. A short summary of this change is that for all intents and purposes, ‘pure’ exact match no longer exists.
With this change, Google started to match up phrase and exact match keywords to a more diverse range of search queries. The impact on any account would have been similar to the Impressions vs. CTR graph which shows the behaviour of a paid search account in which impressions increased by nearly 50% with a similar converse drop in click through rate (CTR). All things equal, if clicks remain constant, an increase in impressions would equal a drop in CTR. This has a host of implications for a paid search account, from lower Quality Scores, to higher required average CPC bids.
When bidding on a term such as [Used Cars for Sale] – in exact match, the keyword may appear for the following variations (as an example):
That’s not so bad you think, only 14 unique search queries were matched up to the exact match keyword. Let’s expand our investigation to phrase match on the same keyword “Used Cars for Sale”
A snapshot of the search query report shows us a wider matching to search queries, but for the most part, the keywords are fairly qualified and relevant. The phrase match query mapped to over 2,000 unique search queries.
What about the data that Google is not showing me?
The fact that search query reports for exact match and phrase match are looking a little ‘too clean’ is a blog post unto itself. Why do I say too clean?
Ever since Google rolled out the match type changes, ‘Other search terms’ in the search query reports has grown, significantly. These are search terms that are driving clicks and impressions which Google does not share in the report.
In the screen shot below, you can see that 55% of impressions are being driven by keywords into which there is no visibility.
These search terms are only driving 3.6% of clicks in this particular account, but at a CTR of 0.57% compared to the average 8.82% CTR, it is clear that these keywords are being matched up to search queries that are not relevant to the account.
Serving ads for a higher number of search queries increases competition on the search engine results pages and this in turn leads to higher bidding by the market.
The account used as an example here is not running any broad or modified broad match types, so I cannot show you how ‘broadly’ Google is matching queries to broad match keywords. If you’re running broad match keywords, brace yourself and run some search query reports.
The implications of the match type change are:
- Higher number of impressions on all keywords (from exact match through to broad match)
- A possible drop in CTR owing to increased impressions
- Increased competitiveness in the SERPs as competitors now see more competition on a wider range of terms – based on how Google is allowing ads to be displayed
- An increase in ‘sunk’ cost on ‘other search terms’ into which advertisers have no visibility
- A greater loss of control over keyword baskets and their corresponding matched search queries
What can advertisers do about it?
- Closely review your keyword basket, focusing in on broad match and modified broad keywords
- Run search query reports to see how ‘widely’ these keywords are being matched up to random search queries
- Be aware of how many impressions are being driven by your keyword basket – if this is adversely impacting your CTR metrics, eliminate some high impression broad match keywords
- Review your bidding strategy and bid down significantly on your broad match keywords
- Break up with broad match altogether
Think of the hierarchy like this – if exact match no longer exists as ‘pure exact’ and now queries match to close variations of search terms on both phrase and exact match; and modified broad match captures a wider net of keywords, then, what is broad match capturing and is it driving return on ad spend (ROAS)?