Artificial Intelligence – Testing the Robot

When Mike Rhodes presented at DMDU Connect 2020 Artificial Intelligence – Less Artificial More Intelligence (How to play nice with machines in 2020),  he explained the importance of experiment.

We think we are right for the most part, so we should probably take our creativity and test it.   We need to experiment more.  When we experiment, we expect to be wrong, we think we are going to fail.    We were told,  “íf it ain’t broken don’t fix it”.  That was a great rule for the old world but now we need to break it, and try new things.

We will fail and possibly a lot.  How do we mitigate risk?

Here are three tools to use:

• Contingency planning – what if. Run through some scenarios. What if this happens, play through those things in your head first.
• Pre-mortem. Sit down and assume it has failed and ask why did it fail.  This will trick the brain to thinking differently. Once you know potential problems then you can put things into place to mitigate risk to ensure these things do not happen.
• Little bets. Don’t bet the house.  How do you take baby steps.  Talk to the bill payer and ask them if this is something interesting and if they would pay for it.

 

TIE CREATIVITY TO BUSINESS GOALS

As much as possible, tie your creativity to business goals.  How do you experiment?  There are lots of goals we can have for our business and hopefully people and planet are as important to your business as profit, but profit is typically the short term goal post that we can use when we are running experiments to see if they worked or not.

Let’s take a look what the Economist did.

We need more Alchemists.  We need more people thinking about new creative things to do.

Example – the Economists – 2 options Online $59 in print $125.00

Alchemist said offer both at same price, mess with peoples heads.

Offer results:

68% chose cheaper results

Alchemist – offer both same price

No one took up print option.

16% cheap 84% print and web

An alchemist came up with this idea – a computer could not do so.

Having seen this study when I was running a course with my co-author Perry Marshall, I insisted we test something similar to this.  We had three options:

Basic pack $950 –  3 people took up

Basic + pack $990 – 125 took up

VIP (6 only) $4990 – 6 took up

How do you take this idea into your business. How can you have a price anchor, or a decoy, or indeed both?

Wine – most restaurants will have the cheapest to the most expensive.  Smart restaurants will have it the other way having the most expensive first.  Once you get past the $3k Grange and go down the list, you start feeling like a cheapskate if you go down to the cheapest bottle.

In the google world we can experiment with these three things:

BIDDING

Use google smart bidding options and test against some form of manual bidding.

TARGETING

We can look at targeting option like smart shopping and say let google decide who should see my shopping ads v. me controlling everything, lets test that with a little bet.  Don’t go all in on Google AI. Experiment with it in your business.

MESSAGES

Computers not quite there yet for the most part. GPT3 examples – writing creates amazing text.  This area is pretty safe in marketing.

We need to remain mindful that our ideas may not be the best ever.  Test them, experiment, make little bets.

Blog Post created from Mike Rhodes’ DMDU Connect LIVE 2020 Presentation.

If you would like to watch the full presentation, CLICK HERE and start your FREE 7 Day Trial.