Dispersion is one of the main indicators in analytics. It shows statistics on how much data deviates from the average value. If the results are close to the middle, then the dispersion is low, and if they are distant, then the dispersion is high. The higher the dispersion, the more unpredictability and risks for the business.
Let’s look at an example. Let’s take two groups of students. After the statistics exam, the students in the first group got either 4 or 5 points. And the students in the second group were worse prepared, so they have a larger spread: from 3 to 5. Although the average score may be the same, in the second group the dispersion is higher, because the spread between the lowest and highest scores is larger.
The points around which the spread is calculated are the average of the total array of indicators.
And the array is everything that surrounds email data us: test answers, sales amounts, investment costs. In order not to analyze an online store by hand on paper or in Excel, large and small businesses monitor statistics in one window using end-to-end analytics . It will show the path of interaction with the client and help to understand what the company’s weak points are.
End-to-end analytics
Evaluate the effectiveness of all advertising campaigns in one window from click to ROI
Invest in advertising that brings in customers
Read more
When to use dispersion
The following work with dispersion:
scientists (chemists, physicists, biologists);
statistics and analytics;
investors and traders;
engineers.
Another indicator is used in business. Dispersion will help:
understand how predictably a data set behaves;
assess financial risks;
predict the risks of an advertising campaign;
see the risks of investments, their profitability or unprofitability;
understand the spread of this cutting edge technology revenue among competitors and compare the company with them.
Dispersion is very similar to standard deviation. But its formula is simpler, so the question arises whether dispersion is needed. In this case, one indicator can be calculated based on the second. Dispersion is more convenient for statistics and when working with regression.
The advantage of dispersion is that it takes into account any fluctuations: both positive and negative. The squares of the deviations – we will analyze this in the formula below – cannot equal 0, so the appearance of stability is created.
But there are also disadvantages: if the results are far from the mean, squaring will greatly distort the data. If the user has no experience with variance, he may cn leads misinterpret the feature.
