In tables within Analyst and Developer, occasionally users will add up the data within a column and find that it does not equal the Total presented in the table. While this can create confusion for users, there are reasons for this experience based on the data that has been selected.
In general, a total should be consistent for a set of filters, regardless of the data sort. For example, in the Job Postings Table, whether looking at Job Title, Industry, Occupation, Company, or Skill, the total number of postings is the same for a given set of filters. This is the total for the filter set, and should be consistent with the headline number in Job Posting Analytics/Job Posting Competition reports.
Total Postings Count Differ:
Reason 1: Additional rows not present in tables
A common scenario, especially for Job Title and Company, is that the sum of the rows is less than the total.
It is important to remember in these cases that the tables are limited to 1000 rows. In these cases, a user should ask if it is reasonable that there are more than 1000 of the category. The Lightcast Titles taxonomy is more than 70k strong, so we should expect, especially in a broad search, that there are more than 1000 matching titles.
Reason 2: Postings matching multiple entries
In this scenario, common for Skills, a single posting matches multiple rows in the table. In this example, we're looking at demand for skills over a 2-month period in the state of Massachusetts.
Over this timeframe, there are just over 408k postings. Most of these postings call for multiple skills, however; if you export and add the total number of skills, the sum is over 4 million skills requested in those two months - nearly 10 per posting.
In this case, the total is grounded in the total number of postings - it is far more important to know that 36% of postings requested Communications skills (146k out of 408k) than to compare that 146k to the number of skills requested across those postings.
Total Employers Competing Count Differ:
A situation similar to the above - double counting in rows - can occur with the "Employers Competing" metric.
Take, for instance, this report on the postings for Lightcast over a short time period. Because this report used the filter "Company = Lightcast", we'd expect the total number of Employers competing to be one - as is seen in the total. However, in each individual row - for each occupation - there is also a single employer.
While this example is a bit extreme, it demonstrates what is happening. With any filter set, you likely have individual employers competing for multiple occupations; as a result, the Total Across All Occupations should be lower than the sum of the rows.