So this is how Google keeps increasing revenue

Testing new ad - Same CTR, one word difference, 4x cpc

[Soze]

I haven't changed ads in months, and my cpc hasn't changed in months. In fact, even when I first created the ads, the cpc was the same.

Just yesterday I wanted to run some test and swapped out non-essential word X from the ad with non-essential word Y from the ad.

Low and behold, the cpc on the new ad was 2-4 times higher on every new ad. Even if they had the same or better CTR. Keywords are all still 10/10.

When I check ad variations, the old ad does as it did before, but the new one costs too much. That's too bad, because some of them got even more clicks.

So what's the lesson here? Don't bother trying to improve your ads if the old ones are doing fine. 100% of the time the cpc ridiculously. What a joke.

[avalon37]

I have experienced the exact same thing recently; disturbing new change to say the least.

[poster_boy]

I might give it more than 24 hours to shake-out. In my experience, higher CPCs are to be expected from newer ads (and I doubt that Google's algo could identify how 'non-essential' the change was) but, after enough data is collected, the CPCs usually return to logical levels based on key performance indicators.

[avalon37]

Usually yes poster_boy, but I have new ad versions live for almost 10 days now things are still "screwy". When I paused my "new" ad variations things/positions went back to normal.

[dpam]

I wonder what would happen if your bid didn't have that much headroom? It seems Google's algo is now tuned to 'take what's there'. I personally wouldn't generally have a bid at 4x my CPC (although Google clearly now does reward and encourage that with top positions and other actions).

[Soze]

Well, now it seems that some of the other ads are costing a bit more now too, even though they are back on their original ads. It's like adding new ads to the group hurt the quality of it. I'll be deleting them all asap. And it's not a coincidence that this all started to happen right after I created these new ads.

dpam, that's not an option. Need the bid sufficiently high to keep the spot.

[Soze]

It's been several days now, and all new ads are terribly expensive and can vary wildly in price from day to day.

New ad .08 yesterday, .58 today. Old adgroup still at .04 every day.

[poster_boy]

New ad .08 yesterday, .58 today.

Wow, that is volatile... but, on how many clicks per day are these CPCs based? Also, are you splitting traffic evenly (campaign setting = "rotate") between your new and old ads?

[JS_Harris]

a 10/10 ad doesn't seem to cost any more/less than a 5/10 ad anymore.

Also, some terms that generate adsense clicks worth 5 cents or less 100% of the time still cost 1.50 or more to advertise on.

If you completely create a word, complete typo - non existent word, you are still needing to pay a lot to advertise for that typo even if nobody else is paying for it.

And another thing I don't understand is why you are charged different amounts per click based on how much you bid...

example - 10,000 clicks each:
You bid 10 cents for a term, cost average is 9 cents per click.
You bid 15 cents for the same term, cost average is 12 cents per click.
You bid 25 cents for that term, cost average drops to 8 cents per click?

Then out of the blue you are told you need to increase even more ?

I've done controlled tests... the results don't always follow logic.

[coachm]

example - 10,000 clicks each:
You bid 10 cents for a term, cost average is 9 cents per click.
You bid 15 cents for the same term, cost average is 12 cents per click.
You bid 25 cents for that term, cost average drops to 8 cents per click?

I don't think you can understand such numbers unless you know what the competition is doing, both in terms of bids, and whether they are jumping in and jumping out.

I think a lot of people get confused and forget that what they actually pay is heavily influenced (up to the max bid), by what everyone else is doing.

So, it appears it isn't logical. I'm not sure if it actually is logical, but I know we can't really know until we have the info. google has, which we don't.

If you want to do a brain teaser, see if you can work out a scenario that explains the data you have, if you start considering the competition behavior (you can do this without actually knowing, just to find out if it's possible that's the issue).