Campaign/Adcenter help

[xtreme5000]

I've been running a campaign on Adcenter and it was doing pretty good. The first day, I spent around $30. This was a review type landing page. Here are my stats for day 1.

Bid: 3.50
Avg. CPC: 3.14
Avg. Pos: 2
CTR: 3.5%
Clicks:21
LP CTR: 62 %
LP- CTR Click: 15
Revenue: $124 with 4 leads.

So far so good, had a 68% ROI. Thought I had a winner, so I let it run for a bit more.

After a few more ad optimizations I got a CTR of almost 9%, but Adcenter won't reduce my CPC costs. According to adcenter tools, my ctr is almost 2-3 times what others are getting for the same keywords.
Also , my LP CTR has dropped to 31%. I let it run for a total of 100 clicks and Day 2 3 and 4 have so far been losses with only 1 or 2 leads.
Here are the few questions I had:

Should I just attribute the first few leads to dumb luck, or is there something there? How do you know at which point you have real data and when its insignificant? For example, do you determine your ad's ctr at 100 impressions, 1000, 10000? What about your conversion rates?

Also, how should I go about optimizing my LP? I'm guessing for any statistical differnce, I'd need at least 40-50 clicks for each LP? I was splitting my offer between two landing pages with diffent review text, but I found no statistical difference between the two. One got 14 and the other got 15 clicks. And both got equal amount of leads. Is there an easier/faster way to test multiple LP's instead of just two at a time? Basically a low-budget split test?


And how can I get Adcenter to reduce my CPC? I can't seem to find a minimum bid data like ysm or adwords. Should I just keep working on my ads?

Here are some boobies for your help:

 

[DelJean]

Should I just attribute the first few leads to dumb luck, or is there something there? How do you know at which point you have real data and when its insignificant? For example, do you determine your ad's ctr at 100 impressions, 1000, 10000? What about your conversion rates?

Also, how should I go about optimizing my LP? I'm guessing for any statistical differnce, I'd need at least 40-50 clicks for each LP? I was splitting my offer between two landing pages with diffent review text, but I found no statistical difference between the two. One got 14 and the other got 15 clicks. And both got equal amount of leads. Is there an easier/faster way to test multiple LP's instead of just two at a time? Basically a low-budget split test?

This is why a basic understanding of statistics (as a mathematical discipline) is so crucial for PPC marketing.

In general, 40 to 50 samples is too low to be useful, unless there's a very wide disparity between the two options (one LP gets 12 conversions, the other 2).

The higher the sample size, the higher your degree of certainity.

However, since taking a sample is costing you money, in the form of clicks, you need to be able to reach a reasonable degree of certainity (say 95%) as quickly as possible, in the fewest number of clicks. That's what's meant by "statistical significance". When is your sample large enough so that are you x% (90%, 80%, whatever%) certain that one landing page is better than another? If the pages are producing very close in results, you'll need lots of clicks. If they're not close, you'll need fewer.

That's where a good knowledge of statistics comes in. So you can figure out how big the sample size should be to give you a result you can trust.

Of course, this being the internet, someone has done the work for you by producing a free split test calculator and decision tool.

Split Test (A/B Test) Calculator | User Effect

Check it out. It's a good basic tool if you don't know your chi-squared from a hole in the ground.

 

[Sohan]

Those tits deserve a reply.

If you tweaked and increase your CTR, I assume you changed the wording? If so the new wording isn't performing as well obviously and isn't as related to the offer as it should be?

Try swapping back and try again. Over time you should slowly beat the mother fucking comp out the water.

 

[jenzkc]

conversion rates seem to vary by the day. one day you might have a 20% conversion rate on an offer, the next day it might be 0%. of course that depends on how hard you're promoting it, but it if you're not sending very much traffic, you'll see those extremes. also, CTR doesn't necessarily = conversions. i've often had better conversions from a .25% CTR than I have had from a 10% CTR. even if you have it targeted just right, getting 10% CTRs from your ads and an 80% CTR on your landing page, it doesn't mean it will convert better than a setup that gets lower CTRs all the way through.