google adwords - starting from scratch

estimating without historics - any tips?

[herbie]

hi,

i'm planning our companies first paid search campaign - and the boss needs some projections (i.e. how many clicks we'll drive and for how much budget) but i have zero historics. i was wondering if anyone has any pointers...

using the overture tools and a few assumptions i've estmated how many clicks and cpcs we might drive over the overture/yahoo network. i happy enough with this but i can't see a way to do the same with google adwords. any ideas?

someone suggested to me that as a rule of thumb i could double both the expected clicks and cpcs from overture to "guesstimate" the google ones (because google has more traffic and is more aggressive to bid on). i.e. if 10 clicks at 50p each looks feasible for overture then the google network could probably do 20 clicks but at £1.00 each? is this good advice?

thanks v much, herbs.

[humblebeginnings]

Mmm, very difficult to predict anything at all. Guess a lot depends on your product, keywords, bidding, ad-copy, landing page, etcetera.
All I can say is that I get far more clicks from Google than from Yahoo. If you have some sort of serious budget, and if you can bid relatively high, it will make things a lot more easy.

[resham5]

You could also use the AdWords Traffic Estimator to get atleast some idea of where your headed. In my opinion you should consider less than half of the number the estimator gives you and probably double the cost.

[trannack]

I think Resham is pretty spot on. For me I have always found that the estmator is grossly wrong - it can go in either direction. Sometimes I get loads more clicks than the estimator provided - other times significantly less. I would suggest you get the maximum budget you can and run quite an agressive campaign for the first month. Once you have a months worth of figures under your belt re-adjust accordingly. Remember, particularly at the moment, it is taking anywhere from 8 days to over three weeks for a new campaign/adgroup to be fully operational, therefore the first few weeks may not be a true representation. But generally speaking, for me, I get about double the traffic in Google to Yahoo. Good Luck, and let us know how you get on.

[Green_Grass]

IMHO there is no way to estimate. As you fix a budget and fix the CPC, the Google algo will make efforts to exhaust your budget. It will take a while for the algo to find the optimum sites ( for Content) and optimum impressions for Search but soon it will start to use up more and more of your budget. If your keywords are popular, you may try the 'recommended budget' feature in adWords. It may not be spot on but will give an indication of the no. of searches and keyword popularity.

You need to take the plunge before you can give estimates to your Company.

[AJSchott]

Using Google's Traffic estimator will give you a rough idea... but perhaps you should turn the scenerio around for your boss... How much is a conversion worth to the company? A conversion being defined as a lead, sale or whatever the desired action.

For example, let's say the net profit from a conversion is $100. Estimating a modest 2% conversion rate would mean for every 200 clickthrus you should get four sales. Depending on the cost of your specific keywords, the $400 net profit will need to cover the cost of 200 clickthrus to break even.

[ronburk]

If you can estimate accurately before trying it, then you didn't try very hard. I can nearly always significantly improve my results over time as I explore what ad text gives me the lowest CPC and the highest CTR.

The best way to get an estimate is to start with a modest budget (for many sectors, a teeny-tiny budget can give you enough info to get a good ballpark estimate of what's possible) and start getting those real numbers ASAP.

[herbie]

thanks everyone for the advice - this really helps. cheers v much!

[QualityNonsense]

I'm with AJSchott - I'd ask the boss what he can afford to pay for a lead/sale. I've yet to use a traffic estimator that provides remotely accurate data. Even at an order of magnitude level, Google's tool is often way off (see numerous past threads for an explanation).