I had a very interesting conversation today with an ideal customer for Interact.
This customer has someone on their team who is very excited about using Claude to make a completely custom quiz. I can pretty much imagine the reason why, and it is a very understandable reason.
Probably what happened is they got into Claude and built a very complex, in-depth quiz or assessment, because that is what AI in general, and Claude specifically, pushes you to do.
It pushes you to make something really detailed. The more detailed and specific you can make your assessment, the more value it is going to provide to the person going through it, or so it seems.
Now, there is a very specific reason why that is not true. It is detailed in the bestselling psychology book Thinking, Fast and Slow.
Fast thinking and slow thinking
The whole premise of that book is about fast thinking and slow thinking. Fast thinking is what you do almost all the time throughout the day.
It is what you do when you are making snap decisions and are probably fixated on a million other things. You have to make decisions throughout your day, and that is what people are doing 99.99% of the time.
Slow thinking is when you actually stop, reconsider something, look at all the factors, and make a decision.
The problem with a complex quiz is especially apparent in a place where someone discovers a quiz or assessment on your website. Maybe they visit after clicking the link in your bio to learn more about you after seeing your content.
They are not necessarily in a place where they want to stop, look at all the factors, and make a decision.
The quiz has to generate leads
Another issue is that, in order for your quiz to be valuable, it needs to generate leads and it needs to be connected to the rest of your marketing.
We have data now, after running Interact for the last 13 years, showing that people do not pay for our product in the long term unless they are successfully using a quiz for lead generation. That is a dangerous problem.
That becomes a huge roadblock, because connecting to an external integration using Claude is going to be difficult. You are going to run into issues and roadblocks.
If you want to segment your leads based on quiz data, that is a whole other problem. Then, if anything ever goes wrong, and most integrations update their APIs frequently, you will have to go back in and update your quiz.
You could also just host it on Interact, where it is about $30 a month at the low end to host your quiz, and we have a pretty generous plan.
A quiz built with Claude is also not going to have reporting or analytics.
The simple quiz formula works best
I have seen a lot of people go down this road and build out a really complex assessment. They look impressive. But the engagement is lower.
The actual formula for the highest-converting quiz is a very simple one.
You have somewhere between three and five unique outcomes. These are different paths that apply to people who are in different places or have different personas.
Then you have five different questions asking people about themselves, their interests, where they are at, and who they are. Based on how they answer those questions, they are guided to one of the outcomes.
You write those outcomes yourself, from your experience, and you have a simple lead capture.
It is actually the very simple quiz formula that works best.
Unfortunately, Claude tries to make you overcomplicate everything and spend a lot of time building it out. The integrations are not handled, so you have to figure out how to make it live somewhere.
At the end of the day, you are going to get a lower conversion rate.