#announcements

@Ari Hoffman @Rob Ayre didn't y'all do a survey about compensation in the recent past?

We worked with our friends over at the Customer Marketing Alliance on that one! Audrey, you can find the report here: https://www.customermarketingalliance.com/customer-marketing-salary-report-2022/ this should help to give you some guidance and help you along!

I believe @Kaily Baskett has also done some research on this?

@Audrey Zigmond click to make full screen/zoom/etc

@Kaily Baskett this is wonderful. Looking through it, wondering, how many people contributed to your sample size?

Hey Ari, apologies in advance for the novel! But I looove statistics so I can't help myself. Sample size is indicated by n, and for this question n=82. I know people hesitate when they see sample sizes like this, but my background is actually in statistics and psychology research, and I can confidently say this data is representative of the CMA population. Any sample size greater than n=30 is sufficient for deriving statistical significance. If you're nerdy like me, here's the deeper dive: The guiding principle for a sample size of n>30 is called Central Limit Theorem, which is the statistical basis guiding modern behavioral research. The data represent a normal sample distribution of means, which is considered to be an unbiased estimator (as opposed to median or range which don't target a population parameter). These conditions are the ones required to be satisfied in order for CLT to be applied. In layman’s terms, we could go out and sample another 100 people or 200 people or 3,000 people and the trends in the data would remain the same because the results from the sample (n) represent a normalized view of the entire population (P). Now, that being said, like any good scientist I'm cognizant of biases. A potential bias in our sample population (n) could be motivation. For example, perhaps people motivated enough to take a career survey are more likely to have a higher salary. As with any study involving humans, there will always be variables we cannot control, such as motivation, but if we determine a pattern or a significant deviation (standard deviation is a whole other topic) from the data, it would be worth investigating in a follow on study.

You always amaze me, @Kaily Baskett!!! Your explanation is interesting. I don’t know enough to understand how you pulled your general sample size so I can’t know how the multitude of varying biases might influence it. I think it’s incredibly hard to pull a truly random sampling of 30 people that would actually satisfy CLT. I can almost guarantee that results will skew at 3000 vs 82 because of all the influenced variables in the the sample size, but that’s almost impossible to remove when we survey industry professionals the way we are. Even with the CMA Salary report I do notice there are some deviations with their larger sample size of 300. I’m not saying your results aren’t valid, the look generally correct to me (maybe a bit high but nothing substantial), just that I can almost guarantee the numbers will skew with a sample size of 3000 because of the larger influence of varying biases at a smaller scale (that goes for the CMA report as well).

Wow thank you SO much everyone who commented and shared their research. I feel really confident about this offer now. Hoping I can close it soon and share the good news.