Guys- Please confirm or deny the following: If I bid on the word 'skillet' in AdWords (broad match) then there is no point on bidding on any other phrases that contain 'skillet' (e.g. 'copper skillet', 'iron skillet'). Assuming my bid is high enough for 'skillet', my ads should appear for all of those phrases automatically. The only reason I might want to bid on 'copper skillet' would be if I wanted that combination of words to have a higher (lower?) CPC than 'skillet' in general. Correct? Just trying to confirm my understanding. D
You should always bid on every relevant term on Exact Match. Even if you bid on skillet on Broad Match, you should still bid on [skillet], [skillets], [copper skillet], [copper skillets], etc... If you are using Broad Match, you should also always bid on Phrase Match as well. Run regular search query reports to find out which keywords are generating your clicks, then add them in as exact or negative matches. There are a few reasons to do this... 1) Over time, your traffic will reduce on Broad Match, until you can remove it. At this point, you will know that all of your traffic is relevant, as Broad Match can give some bizarre results. 2) You'll be able to adjust your bids depending on their performance if each is listed explicitly. People who search for a copper skillet may convert better than the ones that search for a skillet. This being the case, you should be bidding more...
Thanks for your help. When determining how much to bid, goes Google use the 'most exact' match it can find, or does it use the highest bid. Example: copper skillet - $1.00 bid "copper skillet" - $0.50 bid [copper skillet] = $0.25 bid So when somebody searches for that term, would Google use the pricing at $0.25 (most direct match) or $1.00 (highest bid) when deciding to show the ad? Wow. So I just did this, and you're right, there's TONS of junk in there. Chinese characters, words that don't relate to what I'm doing, etc. Here's what I don't get though... I have 8000 clicks over the last year on one particular ad group. And when I sort the Search Query Report by clicks, I see: 3017 other unique queries - 7,023 impressions, 3,441 clicks, 49.00% CTR. WTF??? I did get some conversions for those ads, but the percentage is considerably lower than those from some listed keywords. What are all these clicks, and how come google doesn't show them seperately?
The way Google aggregates low-volume search queries is annoying - you're right. There is a solution to this, using Analytics, if you're interested. The more restrictive match type will apply, if you have the exact query on different match types. Google cover it here: http://adwords.google.com/support/bin/answer.py?hl=en-uk&answer=66292
CustardMite- Thanks. Found the analytics solution; will give it a go. I've been looking at things more this morning, and you're absolutely right. Using broad match on simple terms is a bad idea. I have more than two years worth of history data, and I can already tell I've wasted thousands of $$$ on junk. Wish I'd paid more attention to this sooner. With my KW data, I can come up with a pretty good list of phrase match and exact match terms. But I'm trying to decide whether I should get rid of BM completely, or: a) keep general BM terms, but run them at a much lower CPC. The idea being this could help me discover new keywords on an ongoing basis; the downside seems to be that quality score is related in part to overall account CTR, and doing this would hurt that. b) keep BM, but only on 3-4 word terms (e.g. 'large all-clad copper skillet'). The theory here is that Google will match on enough words that it will still be relevant, even if I didn't put the phrase in specifically as a phrase match or an exact match. E.g. the above would match 'large all-clad skillet'. So I'd still get more traffic, but would be able to keep it relevant, which BM on general terms doesn't do. Thoughts appreciated.
I agree with all that was said above and add that broad match not only exposes you to paying for unrelevant traffic, but it can erode your CPC big time (still the most important factor in quality score). Unfortunately, the search term query report only shows you terms that have generated a click at some time. It doesn't show you all the potentially unrelated terms that didn't generate a click yet still drag down your CPA.
There can be value in broad-matching if you can create a good list of negative keywords. If you can't (because Google is too random with its matching), or it becomes impossibly long (e.g. you have to list every city and country outside your target area as negatives when you bid on Hotels) then it can be impractical. Also, you may get a terrible clickthrough rate thanks to keywords with no clicks at all, which you can't identify. This may also make it impractical. I generally don't touch Broad Match, as I can cover all the relevant keywords with Phrase Match, and it's too imprecise...