Remember I recently wrote about how the traditional way people split-test is a waste of time?
I shared a different method that I use.
Where, rather than trying to find the ONE holy-gail ad (or landing page copy) that will convert for everyone who sees it, that you should instead use several variations on your copy.
Be always running at least two variants and in some cases more.
And rather than killing off the loser and keeping the winner, to just always keep adding new variants based on the “winner”?
Well some moron wrote in whining and spewing Gooroo-Speak about how WRONG I am.
“Keep an ad that does only 10% compared to one that does 30%?
“It’s not like the 10% ad is attracting ONLY the people who are turned on by it. It’s just a random sampling of half your prospects.
“If the other ad that’s also getting a random sampling of half your prospects is turning on 200% more of them relative to the first ad, you’re going to get 200% more sales from that first ad if you replace it with the second ad.
“Sure you might be turning off the original 10%. But you’re also turning on another 30% — so who cares?”
Do you agree with him?
First of all, let me be clear.
Split-testing when you have ENOUGH data to have statistical significance makes sense.
But most people are split-testing WITHOUT enough data.
They’re running like 100 clicks a day to two ads, and if one ad gets 10% and the other gets 30% conversion, they kill the 10% ad.
Although, it is VERY likely that an ad that performs at 30% conversion the first day will PROBABLY continue to outpull the 10% ad, even if the numbers shift about.
Click data is very precious. It costs money on a pay-per-click platform, and you can’t very well sit around for a month waiting to have enough data for statistical significance.
So you can keep trying to run a scientific test, but you’re losing time and money for the sake of a science that doesn’t really exist in marketing anyway.
Plus, the whole point he makes about the “random sampling” contradicts his next point that you could be getting 30% conversion on the better performing ad if you killed off the poorer performer.
Try it please.
Run two ads, one that performs at 30% the other at 10%…
Then kill the ad that gets 10% and let ONLY the ad that previously converted at 30% continue running on its own.
So that everyone who sees the ad ONLY sees the 30% converting ad.
Will you REALLY get 30% conversion?
Or will the performance of the ad drop?
It will drop. Guaranteed!
If there is not enough data to get statistical significance, this will always happen.
Just like tossing a coin – toss it 10 times, maybe 2 will be heads and 8 will be tails.
But does this mean that you have a 20% chance of getting heads when you toss the coin forever after?
No. You don’t have enough data to call your conclusion “statistically significant”
Toss it 1,000 times or 10,000 times instead – you’ll find that the REAL rate is 50% heads, 50% tails.
But who has time or money to toss the PPC coin 10,000 times per ad-pair to find out the two ads are EQUAL?
And after all this time wasted TESTING, they have not managed to create any “improvement” in their conversion rate.
So it’s idiotic to KILL the 10% performing ad, because it is likely to change.
But as long as the two ads are running, you’re scoring 30% AND 10% conversions and all the fluctuations in between.
And as long as there is not enough data to get statistical significance, there is NO good reason to decide the ad currently scoring 10% is a “loser”.
But there IS a good reason to build a THIRD ad variant based on the current top performer.
With three ads running:
– One that has so far scored 30%,
– another that has so far scored 10%,
– and a third ad based on Ad #1 that is expected to perform around 30% (more or less)….
…then you have just increased your conversion rate by some unknown amount without risking losing any ad that MIGHT prove at a later time to be the better performing one.
More advantages of this method:
You don’t waste TIME waiting for statistical significance.
And you don’t waste valuable clicks on just ONE test that may prove useless.
See, that’s why I tell you that marketing is NOT science.
It is a delicate art.
And nobody does it better than our team at YaghiLabs.
Our Done-For-You Traffic Consultations cannot be matched by anyone in the industry.
We build winning campaigns for our clients that get leads right out of the gate. And through our systematic delicate experimentation…we don’t waste your money or time as we ramp up your traffic performance.
Get a Done-For-You Traffic campaign built for you here…