A Percentile is a concept in statistics that identifies a specific data point in relation to a group of data. In the context of salaries, the way to think about it is: the 80th percentile means that only 20% of all salaries for that role are higher than that number.
With Carta as our source for Market Data, we have access to the 25th, 50th, 75th and 90th percentile salaries for every role, at every level. With those numbers, we are able to extrapolate to determine salaries for every percentile from 70th to 90th. This is the range of percentiles we use for salaries at Buffer.
We apply a unique percentile for each area. This is partially a result of our migration from Radford to Carta as our Market Data source. Due to the differences in the market data between these sources, certain roles fell at different percentiles for Carta market data than we fell within Radford. Percentiles have been determined by making sure we are keeping salaries mostly consistent with where they were. The percentiles we arrived at are reflective of the existing state of salaries. Aligning percentiles across the company is one of the plans for Future Iterations of our salary system.
We have a range of percentiles as shown in the table below:
You can also view the list of percentiles directly in our Buffer Benchmarks spreadsheet.
One important thing to note on the varying percentiles is that this only impacts the data input that we bring into the Buffer Benchmark at Level 5. This is due to sample size challenges in our Market Data, which would cause problems and reduce flexibility if we used the market data for all levels. Therefore, once we pull in the Level 5 market data, we establish our own salaries at each level that using consistent percentage increases across all roles and levels.
In this way, while we have some variation in the percentiles we’re using across teams this only impacts a single data point in the overall salary system. From that data point, we then apply consistent increases between every level across the team.