The "Initial Insanity" Factor?

OK, this is something rarely talked about, get inside!

[bigdealioo]

Have you seen a situation where you try to enter a new competitive keyword and to even show on the 1st page you need to bid INSANE CPCs that CANNOT be justified to ANY of the players on that keyword?

You probably have and so have I. So the popular wisdom says - bid the insane amounts until Google gives your ad enough "run" and decides to start bringing your actual CPCs into the realm of sanity.

So I'd like ask you guys. How common is this phenomenon? Does it take a certain amount of impressions for Google to start adjusting your actual CPCs down? Or a certain length of time?

What's been your experience? How have you handled this II phenomenon?

P.S. Hypothetical numbers example. it's impossible to achieve a positive ROI for any advertisers for keyword "pencil" with actual CPCs above $0.30 But when you try to get your nose into bidding on this keyword... a bid of $1 gets you on page 5. You find out you need to bid at least $4 to even start showing on the first page. Ok that's the setup. How you deal with it - is what I'd like to discuss here.

P.P.S. For the sake of discussion my QS on the keyword is Great or Good. So it CANNOT be used to explain the II phenomena.

[bigdealioo]

hmm.. any comment from anyone?

[beesticles]

Yes, I've seen before, usually in the finance sector. Although it's somewhat terrifying, it is effective and necessary to go in high and hard initially. I think in the long run it's better than trying to go from the bottom up. My thoughts are that Google is conservative with its ad serving. It will be reluctant to show ads with little or no performance data over established ones, so you have to give it a significant incentive to do so.

What's important when you start a new AdWords campaign is that you're in it for the long term and are prepared to lose money to begin with while you sort these things out. If you're a 'day trader' you're not going to last long and will probably soon post on WebmasterWorld asking why Google hates you.

[bigdealioo]

beesticles, you seem to know exactly what i'm talking about here! so can you please describe your experience more (without mentioning actual keywords of course)?

How long does it take G to adjust actual CPCs to a sane level for a new ad? In what manner, does the adjustment happen? Is it gradual or sudden? Does G need a certain number of impressions or a certain period of time to base its adjustments on?

thanks!

[beesticles]

So in the finance sector, bids can be at least 10x than you mentioned in your pencil example. Although I've done this lots of times for clients, it's hard to say how many impressions and clicks it takes for Google to bring your CPC down. I suspect this is calculated relative to the keyword itself, so citing numbers here wouldn't be useful anyway. In my experience it takes a few days for the full effect to happen, and it's fairly gradual.

My advice would be:
- only do this on one or two keywords at a time.
- be prepared to haemorrhage cash.
- monitor very carefully throughout the day. As your CTR history improves, you will probably want reduce your bid.
- limit costs by bidding up to position 3, rather than position 1

I do have some lovely charts on this specific process, but I'm afraid they're for my clients only. Perhaps if I speak at SES London then I'll work them into a presentation. Good luck!

[bigdealioo]

great. anyone else?

[tomasvdb]

In my experience, a keyword that doesn't have a good or great quality score from the start, will not substantially drop in price just because it has a good CTR. The main factors in the QS are the 'relevance' of the landing page and the quality of your adcopy.

If you find that for certain keywords your quality score is poor, you might be better off creating a custom landing page on your site than to burn money (and time) trying to get the CPC down.

[bigdealioo]

Tomas, my situation described is advanced and there are several assumptions made. One of them is that my QS on the keyword in question is OK or better. So having a poor QS is not the problem in this case.