[bersius]

Agree that procedural knowledge is important. But I think the critical decisions that make a startup a success or failure don't have a specific procedure as they're unique problems that nobody has every faced before. I also believe that moving out of the box is the only way to truly take what you've built and make it much bigger than it is.

[chehoebunj]

lets take optimizing a facebook ad for increasing paid traffic.

1. take 50 images all else equal, put budget against the 50 ads, find top 5

2.take 50 headlines, rotate against top 5 images all else equal. determine highest performing picture headline combo by CTR.

... rotate in variations of body test, rotate in coloured borders on images...

Now, build an audience of facebook users using demographics, interests, job titles, pages liked.

This is procedural, and a better starting point that will save you thousands of dollars in paid traffic.

[bersius]

Completely agree, but what about understanding who that customer is, connecting the right features with the right group of people with the right marketing. That kind of information has a lot of theory but not a step by step guide guaranteeing success. I suppose this is more idea decisions and direction decisions rather than implementation but I think my point still applies.

[Cataclysmic]

Agree with the benefits of having procedures, and great example also - but I'd change one small thing "determine highest performing picture headline combo by CONVERSION RATE". In my experience high CTR images don't always have high conversion rates.