Sign In

Josef Zach

True R&D, Sensory / Consumer Science Advisor.

Josef Zach is an experienced leader with an international reputation in the field of Sensory and Consumer Insights.1 He currently holds the position of Director Global Product Testing at Ipsos, a global market research and consulting firm.1

Professional Background

Josef Zach is known for his expertise in several key areas:

Global Responsibility:: He has a role with global scope at Ipsos, indicating his involvement in international projects and strategies.1

Innovation Implementation:: Zach is recognized for his ability to implement innovative approaches in his field.1

Team Development:: He has skills in developing and leading teams, which is crucial for his leadership role.1

Areas of Expertise

Josef Zach specializes in:

  • Sensory and Consumer Insights: This involves analyzing how consumers perceive and interact with products, particularly in terms of sensory experiences.1
  • Product Testing: As the Director of Global Product Testing, he likely oversees processes for evaluating products before they reach the market.1
  • Category Appraisal: Zach has shared insights on the importance of product selection in category appraisal studies.2
  • Preference Mapping: He has experience with preference mapping studies, which are used to understand consumer preferences within a product category.25

Professional Contributions

Josef Zach actively shares his knowledge and insights on LinkedIn:

  • He discusses the significance of proper product selection in sensory studies and consumer research.23
  • Zach emphasizes the importance of independent data for product optimization.5
  • He also touches on broader topics such as energy, sustainability, and climate action, suggesting an interest in these global issues.4

Josef Zach's LinkedIn profile and posts demonstrate his deep expertise in consumer science and product testing, as well as his commitment to sharing knowledge in his field.

Highlights

Aug 20 · linkedin

Understand correlation

If you want to understand the relationship between two variables, correlation is a good measure. You don't need special software to calculate it; you can do in Excel with the CORREL function. The result will be a number between -1 and 1. This measure is known as the linear correlation, also called Pearson's correlation coefficient.

This is often used when you want to find out what drives consumer liking. Use any other variables you have at hand, and run the correlation. This will give you an idea which variables are most highly correlated to liking. Whether they DRIVE liking, i.e. whether there is a causal relationship, is different matter that correlation analysis cannot determine. As discussed earlier, deeper analysis, good experience and follow up research are required to dive into this causality.

Anyway, allow me to ask: Do you understand how this calculation of correlation is done? I recently tried calculating it and learnt there is different methods to get to get there. Being able to do it is an eye opener, as it becomes clear that one single data point can dramatically influence the correlation of many observations.

In mathematical terms, the correlation of two variables x and y, is the covariance of x and y (s xy), divided by the product of the standard deviations of x (s x) and y (s y).

s xy you could also see as standard deviation of x*y, anyway it Covariance is the most common term. Covariance indicates the direction of the two variables but no more.

When you standardize the covariance (as done above), you get the correlation. This provides a measure of the strength and the direction of the relationship of the two variables.

The small table below lists an example with 10 observations and 2 variables. I started by entering a few low numbers that indicate a positive correlation. Then I changed one data point on the y-axis, and the correlation dropped significantly (see the red box). Next, I changed x and y values of one data point to 50. Wow, the correlation then went through the roof!

If you are as excited as I am, try to replicate this table in Excel. Play with the numbers and you can see what it means for the different pieces of the correlation. If you need help, let me know, I can share the formulae.

If you don't want to dive that deep, allow me to say: Don't blindly trust a correlation number in a table. Understand the frequency distribution of scores on the x and the y axes to see if there is consistency in the data, or if a few out of many data points are misleading you.

Visualizing that in a scatterplot can also help when you have many different data points on the x- and the y-axis.

Feel free to reach out for any questions. We will gladly help!

Aug 13 · linkedin

Towards the end of the scale

When striving to reach the action standard, the question often arises: How much better do we need to become to gain statistical superiority?

And for well-liked products, is it even possible to reach superiority when you are already at a very high level of liking, and your mean overall liking is close to the end of the scale?

From a business perspective you might need to ask yourself whether you support your intrinsic product sufficiently with a bundle of marketing activities, to ensure consumers are aware of what you are offering, willing to try, and happy to repeat the purchase after trial.

From a product development perspective experience tells us, there is always room for improvement. My recommendation is to think broadly, to discard all constraints, and to first explore any opportunities, even if practically difficult, or cost wise not imaginable. This way we will find your superior product, and in a second step you can explore whether affordable solutions can be found close to that ideal.

From a statistical standpoint we need to look into standard deviation. When analyzing data, the standard deviation is a critical measure of variability that indicates how spread out the data points are from the mean.

In certain scenarios, especially with rating scales, the standard deviation tends to decrease as the values approach the upper or lower limits of the scale. This phenomenon occurs because there is less room for variability near the ends of the scale, causing data points to cluster more tightly around the mean.

For instance, consider a rating scale for customer satisfaction ranging from 1 (very dissatisfied) to 10 (very satisfied). As ratings approach 10, the variation in responses tends to diminish because ratings cannot exceed 10. Consequently, the standard deviation becomes smaller, reflecting a high degree of consensus among respondents.

The shrinking standard deviation near the scale ends does not imply that achieving 'superiority'—defined as consistently high scores or performance—is impossible. Rather, it highlights that the range for improvement narrows as values approach the scale's upper limit. In practical terms, organizations or individuals aiming for superiority must focus on incremental improvements and consistency in high performance.

Understanding this statistical behavior is crucial for researchers and practitioners who interpret data from bounded scales. Recognizing that standard deviation decreases near scale limits can help in setting realistic expectations and designing effective strategies for achieving and maintaining superior performance.

Related Questions

What are some of the key projects Josef Zach has led at Ipsos?
How does Josef Zach approach innovation in product testing?
What is Josef Zach's background in sensory and consumer insights?
How does Josef Zach contribute to team development at Ipsos?
What are Josef Zach's thoughts on the importance of independent data in product optimization?
Josef Zach
Josef Zach, photo 1
Josef Zach, photo 2
Add to my network

Location

Greater Munich Metropolitan Area