How many impressions do you allow before making changes?

[Nick Brighton]

So I'm running this Adwords campaign and it's very early days. But I wanted to know how many impressions you would typically let a keyword/ad group get before deciding to make changes?

I'm looking at some keywords with around 250 impressions and a low CTR. I get the feeling that 250 imps is not enough to start messing around with the keywords/ads etc.

Is there any hard and fast rule, or something which you've found to work well when optimizing your campaigns?

Cheers,

Nick

[CBMoneymachine]

No hard rule here mate but you are right that 250 impressions is too small to start tweaking things. If you start getting up to 1000 impressions on keywords and no clicks then maybe start looking at some other factors such as position etc, make sure that you ad groups and ad texts are really relevant to each other. If you keep getting poor CTR then you may want to focus your ad groups even more so you can try some keyword Insertion in your ads. Also check the search terms through the adwords interface to make sure that your ads are actually appearing for relevant terms - you might have overlooked some other meanings of your keywords which you would want to try and prevent your ads appearing for by adding related terms as negatives. Keep an eye on it and set yourself a limit. In reality you really want to have enough cash for a large enough sample of data to begin making informed optimizations. Sorry I keep jumping from one thing to another here but read about standard deviation & you should be able to find out within reason if your ads could just be in an expected trend or if they are genuinely doomed in their current state - you'll need a bit more data to do this. I personally think with enough data any campaign can be optimized & made productive. At the moment though I would just keep an eye on spend & your broader keywords to make sure that they are not killing cash (which happens quite regularly at the beginning of a campaign)

[CBMoneymachine]

Having said that, I am assuming that your ad group/keywords has a pretty decent estimated amount of traffic. If not then you will need to tweek things earlier.

[tryinhere]

Nick, yes each to their own in regards what they do, if your write down and start to create a rule book you want your on field players / keys to work by that will give you a start point that can be adjusted as you move forward. Do not look for perfect out of the gate, just small improvements.

In regards your question i am guessing you are referring to the display network ? and if so the answers would be different to the search network, as both would have differing measures.

In both cases i would rate impressions low on the pecking order of importance, the key number here is your ctr over and above impressions as impressions itself along with the clicks only facilitate the numbers needed for ctr to be calculated .

coming back to ctr there is no set rule, but what you want to look for is a bench mark, from that marker and combined with a qualifier period (as set in your rule book / say 100 impressions search / 1000 impressions display ) you can as umpire say at this qualifier point does my key meet the bench mark ctr required ? if not blow the whistle and show that key the red card and send it of the playing field.

[Nick Brighton]

Originally Posted by tryinhere

In regards your question i am guessing you are referring to the display network ? and if so the answers would be different to the search network, as both would have differing measures.

I was talking about search, not display... but I get your drift.

In both cases i would rate impressions low on the pecking order of importance, the key number here is your ctr over and above impressions as impressions itself along with the clicks only facilitate the numbers needed for ctr to be calculated .

For sure... that's kinda what I meant. Basically, I was referring to impressions as a way to measure the volume and final judgement on the CTR.

Obviously a keyword with 10% CTR that has only 100 impressions doesn't mean anything.

However, a keyword that has 2% CTR and has had 1500 impressions would suggest that there's enough data to know that the CTR is consistent.

So yeah, at what point do you look at the number of impressions to judge whether you need to work on the CTR?

(that should have been the right way to word it!)


P.S - Maybe it's just intuition and personal choice, I suppose.

[tryinhere]

Originally Posted by Nick Brighton

I was talking about search, not display... but I get your drift.

P.S - Maybe it's just intuition and personal choice, I suppose.

Yes you will need to write your own rule book, for me i will cull any key with no clicks well before or at 100 impressions (search) very rare more than that as a key with less than a 1% ctr rate does not turn my happy pills on, for others it may be OK, and i will start to cull at numbers very low, 250 impressions in search tells me most of what i would need to know for sure, and would have been gone well before that if it did not meet my qualifier.

I think that comes with time so set your rule book qualifier higher and as you become more comfortable with your calls slowly lower the cull / qualifier number as you become better at reading the market.

[Lucid]

Turn things around. Don't go by a certain number of impressions to make a decision. Go by a certain number of clicks. I go with at least 20 clicks for each ad. Note that you have to have two ads and test one against the other. You don't just pick a click rate out of a hat. It therefore doesn't matter if it's the search or content network.

[Fraggler]

Are you split testing your ads? If not, start doing it now! 250 views is stacks to get a decent idea.

Try and beat the winner. When you get a CTR you are happy with you can reduce the impressions of the new ad so you don't lower your potential too much.

[Nick Brighton]

yeah, split testing ads as we speak guys - good points though, especially about just monitoring clicks, that kinda makes sense.