How to Build a Salary Structure with Simple Regressed Midpoints

[@segunakiode]

This is awesome - simple and easy to understand. I just learnt a new way to regress midpoints in a pay structure design. Thanks!

[@ATRI_Total_Rewards]

Cool. Yes it's really simple.

[@ma.jardishamarcellana1511]

This is so helpful! Thank you so much! I'm just curious as to the .8 and 1.2 values. What are these numbers and why .8 and 1.2 specifically?

[@ATRI_Total_Rewards]

The .8 and the 1.2 are used to set the. Salary range width. Salary range x .8 will determine the range minimum. Midpoint x 1.2 will be the range maximum. Companies will manage salaries to stay between minimum and maximum. There is no “correct” salary range width. In some countries, ranges must be wider, such as +/- 30% or more. This model example ises +/- 20%. I hope this helps!

[@BuiQuynhNhu]

Thank you very very much!!!

[@ATRI_Total_Rewards]

Hi Bui, you are most welcome!

[@chandurante5166]

thank you very much!

[@ATRI_Total_Rewards]

You are welcome sir!

[@MeandYoutube516]

I find your explanations very easy. I would love to build a career in compensation and benefits. Please can you advise on specific excel skills that i can learn? Thanks

[@ATRI_Total_Rewards]

Go to www.my-ATRI.com and look at the competency framework.

[@ElizaPolly-v5e]

What if the company hasn't set up grades/levels yet for their employees/jobs, but you have employee salary and market data? Do you have to create the levels by combining jobs with similar market midpoints?

[@ATRI_Total_Rewards]

That is a good first step. Look for natural levels as you group together similar jobs

[@sumitkulthe2391]

Can we get the working sample excel file? Looks good

[@ATRI_Total_Rewards]

I will make the working sample available very soon through a new ATRI Total Rewards Community site loaded with resources. If you subscribe you will be notified here with the link/QR to the new site. Thanks for your interest!

[@thabomgcina8259]

How do you manage the midpoint differential, i.e. increase or reduce

[@ATRI_Total_Rewards]

Good question. There is no single correct answer. In the US you can have a consistent midpoint differential (grade to grade change) whereas in hierarchical cultures and developing countries, I generally find it necessary to segment my regression, with a flatter midpoint differential (e.g. 10% grade to grade) in lower grades and a steeper progression in higher grades, such as 20-30%.

[@DMediaInc]

Hey there. I think I missed something. Could you please tell me where the 0.8 and 1.2 came from?

[@DMediaInc]

Never mind I figured it out. This is the 40% range for the known midpoint. I got it now.

[@ATRI_Total_Rewards]

These set the range width. Using .8, the salary range minimum would be 20% below the midpoint, and with 1.2 the max would be 20% above midpoint.

[@manneed2]

Should the midpoint regression be equal at all grades/levels or it can be progressive?

[@ATRI_Total_Rewards]

Hello, and apologies for the late reply.. the midpoint regression can vary. I sometimes must create a hockey stick-shaped midpoint progression with 10-15% mid-to-mid differentials in lower grades and higher (like 20-30%) differentials for the upper half. There is some art to structure design.

[@ifeanyiohaka8097]

very interesting. I will need your help. don't really know how to do alot. teach me more please or refer me to an institute where i can learn practical application of HR functions

[@ATRI_Total_Rewards]

Hello sir. See www.my-atri.com which is the total rewards institute I created. Follow and subscribe because in June I will also introduce a site with online learning content and community for you to meet others eager to learn comp & benefits.

[@ifeanyiohaka8097]

@@ATRI_Total_Rewards Thank you and God bless

[@coachj9075]

what formula did you use for the job matching?

[@ATRI_Total_Rewards]

human judgment is needed to select the best match between a company job and a survey job. And it's possible there is no good match. Look for a survey job that is at least 70% the same as a company job, in terms of the nature and level of work.

[@Stephanieslinkard]

When I've needed to do job matching for positions with no 100% survey equivalent I've combined the two most closely matching to create a hybrid comparison position and taken the average salary data from that new hybrid. All companies may not prefer to do it that way but mine did. When I did that, I made sure to include a footnote that the survey position was a hybrid and that no 100% comparison existed.