Ali Nawab

Context is everything, and simple statistics with excellent context will outperform the most compex analytics if there is limited context. always understand the why, and make sure the effect sizes back up your why.

One time at my work, we had two products. One of them had a lot of attention, marketing and was considered the star of the portfolio. However, the effect of every 1 FTE added to this product team diminished after the first 3 people. In contrast, adding 3 FTE to an underrated product yielded in higher adoption, and customer retention, and scaled up to 10x the "star" product. At the starting point, both products looked the same, but in 18 months the effect size was 10x more.

Visual aids can indeed be powerful tools for explaining complex concepts like effect size. For example, a straightforward bar graph displaying small, medium, and large effect sizes (0.2, 0.5, and 0.8) can effectively illustrate the magnitude of differences between groups. This approach bypasses statistical jargon and emphasizes the practical significance of the data, making it accessible and understandable for a broader audience. Using visuals in this way helps communicate the true impact of findings, whether in business, healthcare, or other fields, fostering clearer and more impactful decision-making.

Effect size cuts through the noise to reveal the true impact. Take Ozempic, a diabetes medicine, however has been the rage as a weight loss aid. Statistical significance tells you it helps people lose weight, just like exercise but that requires consistent effort. With "limited" effort, the effect size shows the magnitude – whether it helps you shed a couple of pounds or transforms your body. It’s the difference between a minor tweak and a major breakthrough. People losing 8-10 kgs in a matter of weeks without hard work is a sizeable effect size. While there are likely second and third order effects of this approach, the effect size is clearly visible on the users' waistline.

Effect size is like the weight of the evidence. It's not just about knowing something happened, but how big a deal it is. Imagine you hear about a new diet that works. Statistical significance tells you it’s effective, but effect size tells you if you’ll lose 1 pound or 20 pounds. It’s the difference between a minor change and a game-changer. Effect size measures real-world impact, making sure we don't get excited over trivial results.

find people you love to work with, because you will spend a lot of time with them. Through the highs and lows. The best people in our team have worked together for over five years, and have moved from project to project. The shared trust helps the team move forward at lightning speed.