It says on Google: For calculating a keyword-targeted ad's position on a search result page: * The historical CTR of the ad and of the matched keyword on Google. So does this mean, if you have: [widgets] "widgets" widgets in the same Ad Group, and Ad1 is getting a high CTR for [widgets] but very very poor for widgets (as its too broad), it won't effect the historical CTR for the ad, as it's not an exact match? Also, what if you have 2 ads in 2 different Ad Groups which are exactly the same - do they have separate unique CTR history or do they share the same history because they are both identical? - I would guess at they'd have different history as they'd have a different ad number. It also says: The relevance of the keyword and ad to the search query So does this mean if you have a keyword in an Ad Group which is 'Great' 0.01 it has a better chance of being ranked higher than if it had a different Ad and was given poor 0.25, with the same max bid? Because surely the ad in the 0.01 ad group is more relevant in Googles eyes, hence why it's given such a low min bid. I'm just trying to iron a few thoughts out that's all.
Each keyword has a Quality Score, and each advert has a Quality Score, and there's another for the campaign. Your keyword QS won't be affected by broad matches, but your campaign QS might, particularly if broad matches generate the bulk of your impressions. If the keywords are in different Adgroups, phrase and broad match versions wouldn't be taken into account. Whether the exact match results from a different adgroup would be used, or the formula for a zero-impression keyword is an interesting question... Certainly, the advert QS's would be different, in any case. Not sure if I follow your last question, but I'll try to answer it anyway... The QS that you can see is the minimum bid QS, which just determines your minimum bid. But it contains many of the same inputs as the ranking QS, which is used to determine your position for a given bid. Your bid is multiplied by your RQS, and you are ranked against your competitors based on this. You then pay just enough to appear above the next advert down, taking your RQS into account. For example, if you have a bid of £1.50, and an RQS of 2.0 (adjusted bid of £3.00), and the next advert down has a bid of £3.00 and an RQS of 0.8 (adjusted bid of £2.40), your cost per click would be £1.21 (£2.41 to beat the next advert down, divided by your RQS). If you were running two adverts, and the second advert had an RQS of 1.9, then your cost per click for this advert would be £1.27 (£2.41/1.9) In general, the minimum bid quality score is usually taken as a guide to your RQS, but in reality, they could be very different - we have no way to know for certain.
For example, if you have a bid of £1.50, and an RQS of 2.0 (adjusted bid of £3.00), and the next advert down has a bid of £3.00 and an RQS of 0.8 (adjusted bid of £2.40), your cost per click would be £1.21 (£2.41 to beat the next advert down, divided by your RQS). But does knowing this formula actually give you any benefit? As you don't know what the competitor's RQS is, nor do you know what their bid is? "The QS that you can see is the minimum bid QS, which just determines your minimum bid. But it contains many of the same inputs as the ranking QS, which is used to determine your position for a given bid" So in theory, starting keywords off as 'great' will give you a better chance of a being ranked higher than if it was 'poor'? I guess it could be that the Ad that has made the keyword have a 'poor' QS could actually end up having a higher CTR? "In general, the minimum bid quality score is usually taken as a guide to your RQS, but in reality, they could be very different - we have no way to know for certain" Ok thanks, that basically answers my final query. So when first starting out a new Ad Group would you play around with different Ads (and make notes) until you find your Keyword QS to be 'Great' and then take it from there? That's what I've been doing. If they have been poor or ok, I've changed the Ad right away and *usually* later that day it's changed again, so I know whether pausing/deleting that Ad, or adding that new Ad made a positive or negative k/w SQ effect. From that point I've then started the split testing stage.
Regarding the advert working poorly for certain keywords, giving them a Poor QS, this is one of the main arguments for putting keywords (with significant traffic volumes) in their own Adgroups. The advert should then always match the keyword. There are two ways that an advert could damage the keyword's QS - by not containing the keyword (or a derivative), or through a low clickthrough rate. Both are resolved by highly targeted, small Adgroups...