Local optimum vs global optimum (5)

I don't post a ton due to a combination of my natural forum laziness and how I currently make my money. That said something I see happening a lot is this:

Newer marketers get caught in the local optimum instead of searching for a global optimum.

What is the local optimum? Essentially you take a set of possible solutions say 10 ad copies, 10 pictures, 10 LPs. You find the best possible solution from those three variables. (or the best plateau from that series)

Why is this bad?

Lets say you run a POF ad campaign-- all your pictures are of beautiful blonde haired women. Your CTR is .2% and you are running at a solid 40% ROI. Pretty good.

so you find 10 more pictures of beautiful blonde haired women-- your CTR sticks within a .1-.3% You are trying to squeak that 40% ROI up to 45%. Are you happy with this? Sure, I guess but...

You decide to throw a curve ball into the mix and you test a super hot asian woman and your CTR jumps from .2% to .4% and your ROI jumps with it (80% ROI?) Yeah you're happy with that but you are still within the same topological space and thus sticking in the local optimum.

What if a radically different approach could yield better results? What if instead of a picture of a beautiful blonde women-- a picture of an overweight hairy man was producing a 3% CTR and your ROI could increase at the same magnitude. By sticking within the local space of beautiful women-- you'd never come across this hidden gem and, for me, thats great. If you miss it maybe I won't and while you're trying to earn dem ones, I'll be earning dem Ms.

In short what I am suggesting is-- try and test radically different approaches-- do things that are completely outside of the norm and see what you can get.

Disclaimer for the pedantic: The terms are taken from mathematics and my use of them isn't perfectly in line with their mathematical counter-parts. Yes, I am aware of it, no I don't care.

Dog I don't even know what pedantic means.

Good insight though. Purple cows, pattern interrupts and whatnot.

Had to look it up as well:



Interesting advice and indeed it seems that many newer marketers do focus on the local optimum

I mean of course it makes sense.

Most people are searching for lower risk lower reward options. This is certainly the case with new marketers that have limited budgets.

Any major 'global optimum' solutions that I've run into myself have been HIGHLY polarized. That is, they either completely bomb or give me 200%+ ROI's for months. The problem with finding them is that it requires a budget that's dedicated to absorbing losses 9/10th's of the time and you hope that you find an unusual angle strong enough to make up for it.

New marketers can't afford, through time or money, to find global optimal solutions.

Originally Posted by mrpenn

I don't post a ton due to a combination of my natural forum laziness and how I currently make my money. That said something I see happening a lot is this:

Newer marketers get caught in the local optimum instead of searching for a global optimum.

What is the local optimum? Essentially you take a set of possible solutions say 10 ad copies, 10 pictures, 10 LPs. You find the best possible solution from those three variables. (or the best plateau from that series)

Why is this bad?

Lets say you run a POF ad campaign-- all your pictures are of beautiful blonde haired women. Your CTR is .2% and you are running at a solid 40% ROI. Pretty good.

so you find 10 more pictures of beautiful blonde haired women-- your CTR sticks within a .1-.3% You are trying to squeak that 40% ROI up to 45%. Are you happy with this? Sure, I guess but...

You decide to throw a curve ball into the mix and you test a super hot asian woman and your CTR jumps from .2% to .4% and your ROI jumps with it (80% ROI?) Yeah you're happy with that but you are still within the same topological space and thus sticking in the local optimum.

What if a radically different approach could yield better results? What if instead of a picture of a beautiful blonde women-- a picture of an overweight hairy man was producing a 3% CTR and your ROI could increase at the same magnitude. By sticking within the local space of beautiful women-- you'd never come across this hidden gem and, for me, thats great. If you miss it maybe I won't and while you're trying to earn dem ones, I'll be earning dem Ms.

In short what I am suggesting is-- try and test radically different approaches-- do things that are completely outside of the norm and see what you can get.

Disclaimer for the pedantic: The terms are taken from mathematics and my use of them isn't perfectly in line with their mathematical counter-parts. Yes, I am aware of it, no I don't care.