Daring Marketing — Let’s do the math on lovers, haters, and conversion rate

Gwendal

Bold messaging, pattern interruption, going after competition …We all want to try and experiment with more daring marketing content and copy. But is it worth it? And what’s the best way to optimize it. Let’s do the math!

I wrote about the types of risk-taking strategies here(What kind of risk-taker are you: Maximax, Minimin, or Minimax? A decision-making framework) but can we actually measure the impact of “edgy” marketing and how to limit the risk.

Let’s consider 2 marketing strategies:

  • Strategy/Messaging M1: Very classic.

50% of people “kinda like it” (converting at 15%), 50 % are not responding very well (converting at 5%).

  • Strategy/Messaging M2: Very edgy

25% of people love it, 75 % hate it. The love brings a conversion rate of 30%. The hate, 0%

Let’s look at 4 ways to handle it:

  • I don’t want to take any risk and will only use Marketing M1
  • I love risks and will only use Marketing M2
  • I am going to A/B test these 2 strategies
  • I am using AI/ML to do a smart distribution of content

TL;DR

Using an intelligent content optimization system wins, converting at 12%, followed by the no-risk strategy (10%), A/B testing (8.5%), and then risk only (7.5%)

Now let’s do the math!

1. No risk: All the visitors are exposed to Classic Marketing.

  • 50 visitors kinda like it and convert at 15% ==> 7.5 conversions
  • 50 visitors don’t respond very well and convert at 5% ==> 2.5 conversions

Results: 10 conversions, or a 10% Conversion Rate

2. All-In on edgy marketing

All my visitors are exposed to Marketing M2.

  • 25 visitors love it and convert at 30% ==> 7.5 conversions
  • 75 visitors hate it and convert at 0% ==> 0 conversions

Results: 7.5 conversions, or a 7.5% Conversion Rate

3. A/B testing

Let’s randomly serve Marketing A and B to 50 visitors

  • Population A (50 visitors) is exposed to Marketing M1(classic)

25 visitors (50%) kinda like it and convert at 15% ==> 3.75 conversions

25 visitors (50%) don’t respond very well and convert at 5% ==> 1.25 conversions

Results: 5 conversions, or a 10% Conversion Rate

  • Population B (50 visitors) is exposed to Marketing M2(edgy)

12 visitors (25%) love it and convert at 30% ==> 3.5conversions

38 visitors (75%)hate it and convert at 0% ==> 0 conversions

Results: 3.5 conversions, or a 7.5% Conversion Rate

Results: 8.5 conversions, or an 8.5% Conversion Rate

Even probably more important: We lost conversions on the test, ok…but now what? What’s the next step?

4. AI/ML approach

Now, let’s use a solution (like Cauzal AI) that is capable of serving the content based on what is most likely to convert.

It would be ideal to find the 25 edgy marketing lovers out of the population of 100 but that requires an AI model with 100% accuracy, which is not possible right now. So let’s say we capture 20 out of these 25 but still have 20 “haters” in the group.

  • Population A (60 visitors) is exposed to Marketing M1(classic)

30 visitors (50%) kinda like it and convert at 15% ==> 4.5conversions

30 visitors (50%) don’t respond very well and convert at 5% ==> 1.5 conversions

Results: 6 conversions, or a 12% Conversion Rate

  • Population B (40 visitors) is exposed to Marketing M2(edgy)

20 visitors love it and convert at 30% ==> 6 conversions

10 visitors hate it and convert at 0% ==> 0 conversions

Results: 6 conversions, or a 12% Conversion Rate

Results: 12 conversions, or a 12% Conversion Rate