I don't know how you can get data with such a tiny spend, to be honest.
From experience, I would have thought at _least_ $2000 per test.
I've been doing PPC stuff for 20 years. As soon as someone tells me they are doing PPC I ask them how they enjoy setting fire to all their money. Unless you have oodles of time to spend optimizing it and measuring it, you are guaranteed to just burn all your cash. It is very, very hard to win on PPC. One of the biggest problems is that unknowledgeable people in each domain are bidding on PPC ads without monitoring their spending or conversions and wasting their money, but their ads are bumping up the price of yours too.
geez. I've been thinking of doing some fake door tests but wanted to launch something like 10-20 experiments. Even at $500 that's 5-10k, which is not feasible. How can any startup afford these tests?
If I ran 10 experiments/week, with the $2k measure, you're looking at 80k/month, just to get some validation.
It depends on your cost-per-click, or cost-per-action. If you were in a niche that was pennies per action then you could do it on a smaller budget. It's just that you require a lot of data to get any meaningful results. My PPC traffic wavers up and down and it would be hard to know if it was the weather or some other factor unless you had enough traffic to average things out.
From experience, I would have thought at _least_ $2000 per test.
I've been doing PPC stuff for 20 years. As soon as someone tells me they are doing PPC I ask them how they enjoy setting fire to all their money. Unless you have oodles of time to spend optimizing it and measuring it, you are guaranteed to just burn all your cash. It is very, very hard to win on PPC. One of the biggest problems is that unknowledgeable people in each domain are bidding on PPC ads without monitoring their spending or conversions and wasting their money, but their ads are bumping up the price of yours too.