“It’s all math and psychology. Much attention is given to psychology, but math is absolutely critical. Good math can save mediocre persuasion, but bad math will sink the best sales pitch every time.” -Perry Marshall

Choosing The Right Audience: A Statistical Strategy For Increasing The Efficiency of Your Ad Campaign With Fine-Tuned Custom Audiences

Kian Xie, M.A.

Reduce ad spend, create consistent results, and simplify the campaign optimization process by selecting your initial custom audience with visual, data-driven clarity. I'll walk you through the mathematical and organizational process behind this strategy, using Facebook Ads as an example.

Note: Some elements of this walkthrough are missing details concerning mathematical rigor. * View full explanation >>

Kian Xie, M.A.

I believe that all industries can benefit from working together, building bridges between worlds that would otherwise stay separate and insular. All skills are transferable, and the most difficult problems can be solved by synergizing expertise from seemingly unrelated disciplines.  

My 11 years of professional experience in marketing, mathematics, education, the arts, and personal development coaching equip me with versatile tools and limitless perspective to solve problems and create visionary breakthroughs in a wide range of fields.  

Top 5 Gallup Strengths: Activator, Futuristic, Strategic, Ideation, Relator

Contact Me With Your Inquiries, Insights, And Opportunities:

kian@kianxie.net | (401) 545-2557

Currently seeking full-time employment at a stable, growing, innovative company with the highest level of integrity and a strong commitment to positive social change.

Areas of expertise: marketing data analytics, statistical analysis, technical training and program development

Full Explanation of Mathematical Detail:

*The measure of "standard deviation" mentioned in this video would be more accurately described as a "pseudo-deviation". Using the sum of all audience sizes as Sigma(f) in the formula does not technically yield a true standard deviation because all of the audiences listed intersect with each other. Determining the true standard deviation would take a much more time-consuming procedure, although this process can be automated to yield rigorous results when working with a database. If you are restricted to using minimal technology, you can still use this "pseudo-deviation" with good results, or you may choose to experiment with different strategies to determine an effective "pseudo-deviation" to delineate categories. To determine the size of the audience of the whole list, input all the keywords listed into Audience Insights as a set, to create a custom audience which is the union of all keyword-based audiences listed. Then, use the midrange of that audience size as Sigma(f). This will yield a smaller "psuedo-deviation" measure than what I used in the above video, so you may choose to reduce the width of the category constraints as much as makes sense to you (for example, in increments of 0.25 standard deviations), based on the spread of your data. You can choose any width that works - what matters is that you are able to separate your lists into suffiiciently varying categories, so choose a consistent category width that yields a sufficient amount of variation to make a clear decision. Back to top^^