this client uses Clickfunnels.
Most such funnel solutions are built by marketers who pretend at science.
It’s entirely plausible that you find a “winning” opt-in page variation which increases lead conversion but reduces sales revenue. Which is counter intuitive…no one wants to increase opt-in leads at the expense of revenue!
Moreover, traditional split-tests don’t make sense on personalised advertising networks like Google and Facebook Ads.
See, an old-fashioned landing page split-test assumes traffic is random. Therefore, rotating evenly, showing two versions of a landing page in effect, splits visitors into two groups–those who saw version A (the control) and those who saw version B (the variant). Then you can measure and compare the two rates of conversion. When there is enough data to support it, you can declare one version the winner.
Trouble is, traffic networks AUTOMATICALLY optimise and learn the following…
(1) which ad to show
(2) when to show that ad
(3) and, who should see that ad
Meaning, traffic, when it arrives at the landing page, is NOT random.
Fact is, response can be influenced by two factors simultaneously…
(1) is that one version of the landing page is naturally better than the other
(2) is that different individual users are more inclined to respond to one version than the other
If you rotate landing page versions in a traditional split-test while getting modern, non-random internet traffic, then Google thinks it’s receiving conversion feedback for a single landing page when in fact it is feedback about two pages.
Your split-test and the Google optimiser end up interfering with each other’s results, making your traffic shit and your split-test wrong.
On the other hand…
With “Google Campaign Experiments”, you actually split traffic at the outset by duplicating an ENTIRE campaign, making changes to the duplicate, and then splitting only your budget between the two versions. This means you now have two very similar campaigns, with traffic INDIVIDUALLY optimised to each landing page.
Not only that…
You can also measure the overall cost vs revenue of your marketing process that includes page A as its landing page (the control) and a similar marketing process that uses page B (the variant) as its landing page.
Plus, such an experiment is actually tested in the real context of how you intend to use it. So you won’t see one result while split-testing and then get a different result when you roll out the winning change.
1. The variant BEATS the control in absolute lead count
(it gets 150% more leads than the control)
2. The variant BEATS the control in lead cost
(it gets leads at 75% of the cost to the control)
3. The variant BEATS the control in revenue
(it makes 142% more money than the control)
Here’s the real kicker…
The control actually gets a higher absolute opt-in conversion rate (7%) than the variant (3%). Completely opposite result!
If you wanna optimise your marketing, look at the WHOLE picture and measure the full impact of your change.
Your marketing is highly interdependent.
Optimising pieces or steps individually doesn’t necessarily mean that together, you have an optimal sales process. This is why the Super Traffic Machine DIY Kit is designed as a complete building process from start to finish, tied together by a single philosophy.