Create salary bands for every role in your organisation
Intro
Any organisation benefits from having a structured approach to determining salary ranges for their roles. Bringing this structure and process to pay decisions will help you:
- Ensure pay is fair and there is pay equity across the organisation.
- Demonstrate to employees what their earning potential will look like as the grow within your organisation.
A typical salary structure will have the following components
- Job Family - are very similar to teams and relates to the specific type of work or skillset (’Talent Acquisition, Human Resources, L&D etc’), while Departments are broader (’HR’)
- Levels - How senior / experienced the role is. As seniority increases, so does the level of responsibility and consequently the salary. Levels are sometimes referred to as ‘grades’.
- Salary bands - these represent the minimum, middle and maximum pay rate for a given level and how salaries grow within a level.
Why are salary bands useful?
Salary bands support you in making fair and equitable compensation adjustments.
They can be tied to growth
- A salary band is a great way of giving an employee an idea of their earning potential within a level as they grow. As they pick up new responsibilities and increase their impact, they can being rewarded accordingly. Think of it as an objective way to move people through a level.
- This compensation clarity will also help motivate employees to set development goals in order to grow within their level.
- For more senior roles, it’s likely that employees will spend a lot of time within one level, so giving them clarity on where they sit within that level is critical, so they don’t feel like they are stagnating.
Help drive pay equity
- Salary bands help ensure that employees at the same level are being paid within the same range.
- In addition to this, by having a minimum and maximum rate for every job, and then placing employees into this range by selecting a sub-level, you can easily explain to employees why one is earning way more than the other despite having the same role and level.
Critical for financial planning
- Having a clear pay structure that all roles are aligned to allows for accurate financial planning for future roles and salary raises.
Helpful for hiring
- Salary bands are extremely useful when it comes to hiring.
- You can put a salary band on a job advert. I’m not going to go into why this is the right thing to do, but let my friends at Otta do the talking here.
- Giving you some flexibility on where to pay a new hire as you can objectively place them in the range, and not necessarily always at the bottom. For example they could be moving internally within the organisation and not start at the min point.
Defining which management approach to take with a direct report
- Identifying where a report sits in their role, helps managers determine how best to manage them, allowing for greater personalisation in management. Read more about the management styles to take and task-relevant maturity here.
In order to build salary bands you’ll need two things
- Reliable and relevant market data. This could come from various sources such as:
- ‘Give to get’ salary benchmarking providers like Radford, Figures, Pave, Mercer, Ravio, Figures
- Job boards
- Local recruitment surveys
- Your recruitment team who speak to candidates every day
- A ‘common language’, consisting of the elements listed below. These elements are also collectively referred to as the ‘organisational infrastructure’, ‘job architecture’ or ‘hierarchy of jobs’.
- The job families you map to and create will vary depending on the data sources you choose to use. For example each salary benchmarking provider will have a description for each job family that you can read and then decided which one matches your role most accurately.
- Job Families ensure you can identify the compensation data that most closely matches the job or team you are looking for, therefore giving you a more accurate data.
- The Job Family relates to the specific type of work or skillset (’Talent Acquisition, Human Resources, L&D etc’), while Departments are broader (’HR’)
- You can group by departments should you wish, however…
- “If you use Departments you’re essentially comparing apples to apples regardless of whether it’s an eating apple or a cooking apple; when using Job Families you’re comparing Granny Smith apples to Granny Smith apples.”
- The most common structure among scaling tech companies is a two-track structure as shown in the diagram below.
- You either take a track managing people or becoming a specialist managing projects and processes.
- Separate tracks for ICs and people managers allows employees to develop according to their interests and skillsets and ensures people are working where they will be most effective.
- Typically both tracks are valued equally, so the same base salary is used at each level, regardless of whether someone is on the IC or People Manager track in the same department. If this isn’t the case in our organisation then it’s not a problem.
- Levels define how senior / experienced the role is.
- Job levelling criteria is used to define the overall ‘profile’ and ‘size’ of every role and level in the organisation. This is sometimes referred to ‘scope’, ‘responsibility’ and ‘influence’ depending on the levelling framework you are using.
- Choosing your levelling framework - Each salary benchmarking provider has their own levelling framework. You can view some of the most popular ones here.
5 simple models to help build your salary bands
There are a few ways to build salary bands for any role. I’ve put the 5 most common models together in one spreadsheet.
I have lots of data
- Got a min and max for a level but not sure what the mid point is or how this connect with the other levels? This model will help you figure all of this out for you.
I don’t have much data
This model takes your existing market data and applies exponential regression to it which helps:
- Fill in the gaps where data doesn’t exist ✅
- Adjusting for when pay doesn’t progress ✅
- Where data is higher for level 5 rather than level 6, the model takes the highest amount as the max value.
- Where data doesn’t exist for level 1, it will forecast data for level 1 using data from the levels where it exists.
- Where data doesn’t exist at all, the model will fill that hole.
- The model needs a minimum of two data points.
- The model uses the
=GROWTH
function to carry this operation out. - In simple terms, the growth function calculates Y values on the basis of an exponential growth rate of given X values. In our case salary is our response variable (y) since it is our target predicted value, and levels are our explanatory variable (X).
- You can learn more about how the Growth function works here.
- Like exponential regression in model 2, linear regression analysis shows the relationship between an independent variable, such as the Level (x-axis), and a dependent variable, such as Salary Range Midpoint (y-axis).
- In linear regression, the function is a linear (straight-line) equation.
- Linear regression has been a traditional approach used in the past to predict salaries, however this infers that salaries and levels increase in a linear format which is typically not always the case.
- Unlike the other models, this one only requires one data point for level 1. This is very useful when you have no robust data at all.
- You create you range by modifying the midpoint progression from one level to another.
- See the FAQs for more information on the typical midpoint progression %’s.
Executive model
Why use these models?
✅ Help you create your own market benchmarks
If you have a hard to benchmark role (think ‘NFT Engineer’ or ‘Zebra Fish Lab Technician’ - yes, I’ve helped with these in the past) then you can use these models to build a salary structure that’s right for your organisation (all you need is the salary data for one level).
✅ Help you fill in missing data for levels
If you don’t have market data for every level, the models will fill in the gaps.
✅ Solve for when pay doesn’t progress
Sometimes market-based compensation data can be counterintuitive, where market data suggests that a lower level job is paid more than a higher level job. These models help fix this by using regression and simple progression formulas.
✅ Allow you to create salary bands and then determine the width of them
Define how wide your band needs to be for each level. For some job families you’ll want narrow bands and quicker progression in the lower levels, whereas with others you’ll want wider bands and a slightly slower pace of growth.
✅ Allow you to create overlapping salary bands
You have the ability to determine how much over lap there is between levels, or simply have not overlap at all. It’s your choice. If you’re unsure on whether to have any overlap between levels then read on!
Salary band terminology
- The lowest point in the range.
- Formula:
Minimum = Midpoint / (1 + (range spread/2))
- Example: £21,500 / 1 + (30%/2) = £18,700
- This is the exact middle of the salary range.
- The midpoint is set to provide a market competitive and fair salary for the organisation. This is sometimes referred to as the ‘market rate’ and will differ from business to business based on their compensation philosophy and the strategic importance of that role.
- For example some companies may target the 50th percentile as their midpoint, while other companies select the 75th percentile.
- You can also select different percentiles for different levels within a job family e.g.
- Levels 1,2,3 = 50th percentile
- Levels 4,5,6 = 75th percentile
- The data you input here, is used as the ‘anchor’ from which to create the min and max salaries in the range.
- The max salary in the range.
- Formula:
Minimum x (1+ Desired pay range spread %)
- Example: £18,700 x (1 + 30%) = £24,300
- This is the difference between the midpoints of each adjacent level i.e. the % jump in salary from one level to another.
- Formula:
(Midpoint of level above - Midpoint of level below) / Midpoint of level below
- Typically, lower-level jobs tend to have a smaller midpoint differential as the progression from one level to another is quicker. The differential increases for higher-level positions, but this does depend on the market data being used.
- Using pure market data for the midpoint, may well result in inconsistent midpoint differentials. These uneven midpoint differences happen unintentionally over time and very common when you use data submitted and aggregated from 1,000’s of companies.
Level | Typical midpoint progression |
Support position | 10%-15% |
Professional, Managerial | 15%-25% |
Executive | 20%-35% |
- The pay range spread aka ‘band width’, represents the the difference between the min and max salaries in a level.
- You do not need to calculate the spread, as this is something that you determine yourself for each level. Simply set the % to determine the width from the min to the max of a level.
- Defining the width of your band is very important as it determines the growth opportunities for your employees within that level.
- Broadly speaking the range from Level 1-8 should represent a lifetime of progression in that particular job family.
- Levels 1-3 (narrower bands)
- Pay range spreads of: 15%-30% for lower levels.
- Typically your bands will be narrow in the lower levels where progression through a level is quick.
- Smaller range spreads can benefit lower level positions where the market rates are closer. If the range spread is too large it becomes harder to ensure the internal progression of entry hires to the next level, unless your making salary increases of 8-10% on a regular basis.
- Levels 4-8 (wider bands)
- Pay range spreads of: 30-40% for mid/snr levels.
- Pay range spreads of: 40-50% for execs.
- More senior roles will have wider bands, where time spent in a level becomes longer. Role responsibilities skills and expectations and therefore take longer to master and excel in. That being said, if the band is too wide, people may start to stagnate within that level as it takes too long to progress through it.
- Displayed as both a % and $, it shows the degree of overlap of bands across different levels.
- Formula:
(Max of lower level - Min of higher level) / Max of higher level - Min of higher level)
- To overlap or not overlap, that is the question
- No overlap aka a ‘stepped structure’
✅ Some people prefer this as it’s neater, and there is a clear step in salary from one level to the next.
❌ You would need to promote someone to the next level in order to increase their salary even if they are not ready which is risky.
- Overlapping bands
✅ When your salary bands overlap, your company has more flexibility. Employees can receive merit raises without taking on promotions they’re not ready to tackle. Employees have some runway to grow in their current level without a constant focus on promotions.
✅ Overlapping bands recognise the greater value of the input from a highly experienced/skilled individual at the top of their level compared to a newly-appointed employee on a learning curve at the lower end of the level above.
❌ You promote someone to the next level but their pay doesn’t actually go up that much, which isn’t particularly motivating for the employee.
Degree of overlap for senior roles
- Some organisations allow for a higher degree of overlap in the higher levels, where there can be a lot of variance in incumbent pay i.e. pay might be the same for an ‘Advanced’ VP and ‘New’ C-Level employee.
Too much overlap may occur when:
- There is too little difference in market rates between levels
- If you set the pay range spread %’s too high at adjacent levels.
FAQs
- Yes, you need them. Like it or not, they play a critical role in salary benchmarking.
- An organisation can have any number of levels. There is no magic number, however most of the salary benchmarking providers have between 8-12 levels, and these will depend on the size of your org.
- Below is a typical organisation structure consisting of:
- Two tracks
- 8 levels (IC1 - C-Level)
- Sub-levels - This is just another way to talk about the min, mid and max of a range. I personally find it as an easy way to objectively place an employee into either the min, mid or max point of a range. To determine a person’s sub-level we ask ourselves: “How established are they in the role within your organisation?”
I regularly meet orgs who feel that they are too small for levels or that they’ve managed to date without them and introducing them would damage the culture. I’m not going to dismiss these challenges, however if you want to carry out a salary benchmarking exercise you will need levels in some shape or form. If you don’t want to introduce levels or some form of seniority/hierarchy into the the organisation, then you won’t be able to carry out salary benchmarking. There is no halfway house.
- It helps ensure you have the right mix of junior, intermediate, senior and executive talent to reach your organisational goals
- Can aid career progression when used in conjunction with a skills based framework by helping define the routes to progression through the organisation.
- Helps solidify performance management processes by providing something meaningful to anchor conversations to
- Bring teams together - When levels across departments match one another’s scope of influence on product and people, it clarifies roles, suggests natural counterparts, and brings together employees from different disciplines to take ownership over shared problems and solve them together.
- Reduces subjectivity - Employees, regardless of role, are evaluated on the same criteria allowing for similar standards across all teams reducing subjectivity.
- Ease of building a career progression framework - Having standardised levelling framework makes expanding to a skills based framework a lot easier as level expectations for a role will have been defined already, meaning people can concentrate on building the skills that matter and not defining the size and reach of the level.
Levelling framework | Career progression framework |
• Defines the scope and responsibilities associated with the different levels.
• Not values or performance based.
• Definitions are role agnostic which allows for the evaluating and levelling employees on the same criteria.
• A levelling system that can be linked to external market compensation data to ensure objective based salary benchmarking so you can compare apples to apples | • Outlines key skills and competencies expected at a certain level.
• These are skills an individual can work to attain or improve (vs. a job scope which is defined)
• These can be both role agnostic e.g. communication, teamwork and role specific e.g. product knowledge, technical focus |
- We recommend implementing 8 levels (6 levels for IC/Management + 2 exec levels).
- This number of levels is also the most common number with the external salary benchmarking tools use. Therefore, introducing the same number of levels will allow for a quick and easy mapping exercise for the purpose of tying your levels back to pay.
- 8 levels will feel like too many levels at 20 people but by 100 people you will be glad you have that many.
- Even if you have no plans to hire an IC5 or IC6 soon, it’s important to show people already in that team that there is room for growth i.e. a path ‘may’ exist one day when there is a business need.
- It also enables future hiring, as if you do need to hire IC5/IC6’s you will need criteria to measure them objectively against. Having the levels from day one, even if no one fits into them allows you to build for future scale.
- You don’t have to! Sometimes a department like Customer Support will not use the last 3 levels and have most people in IC1. Inversely the Engineering team may have no IC1’s or IC2’s. Both are fine, however across the organisation you need levels that run from 1-6 to account for both depts.
- Just because the levels exist, doesn’t mean you have to fill them up!
- Google and Facebook have ~14 levels, where they have more granularity within their exec function given their size and global reach.
Here are some short descriptions about tracks.
Track 1 - "IC" (Individual Contributor)
- Focus: Becoming a specialist in their role
- IC will suit people if they enjoy:
- Building bigger and better systems and processes
- Knowing and understanding the micro details of how we get things done (and why)
- Quickly identifying how to solve problems in their team
- Getting creative with their technical knowledge
- Working collaboratively with more experimental colleagues to apply their knowledge to improve our ways of working
- Iterating and improving
- Looking for new opportunities and testing them out
Track 2 - "M" (People Manager)
- Focus: Hiring, team organisation, and helping people progress in the business
- Manager will suit people if they enjoy
- Managing others, delegating work, building a team
- Team organisation, including holiday management, role forecasting, and performance documentation
- Guiding others through complex interpersonal and communication challenges
- Mentoring, training, and leading others
- Coaching and supporting others through their careers, difficult situations, and day-to-day work
- It’s really down to the data you have access to and whether you want to differentiate between certain skills or not.
- For example some third party tools will provide you with a breakdown of ‘Front End’ and ‘Back End’ Engineers, while others bundle these into one job family called ‘ Software Engineering’.
- I’ve seen organisations roll up certain job families into the past such into the ones listed below and take averages:
- HR + Recruitment job families = People Team
- UX + UI + Product Design job families = Design
- Data Science + Data Engineering job families = Data
- Front end + Back end job families = Software Engineering
Before place employees into a salary band, you need to make sure that you have assigned them a level. Once they have been levelled, they maybe placed within a range using the following descriptions.
- Before using any of these models, you will need some market data. I have populated the models with sample data in order to demonstrate how they work.
- If you don’t have any market data, then you are welcome to schedule a call with me to discuss how to choose a salary benchmarking provider using the link here. There are lots of salary benchmarking providers out there, and below are a few that I have worked with:
- Pave / Option Impact
- Ravio
- Figures
- Radford
- Willis Tower Watson
- Salary.com
- Mercer
- TalentUp
- I always recommend checking the data with the hiring managers and any in-house recruiters who will know the role and market best. For many job families you may decide that the raw data behaves in the way it should. If this is the case, you’ll find that the raw data and regressed data will track pretty much exactly the same and there is no need to use anything but model 1 to create your salary structure.
- There are no right or wrong decisions to make when choosing which model to use. The choice you make will depend on the data you have and also the pay strategy you wish to implement. Feel free to book in time with me to chat about the models here.
- Key decisions:
- Which model best suits your needs? (FWIW, I personally favour Model 2 😃)
- Do you have plenty of market data to feed into the model you’ve chosen?
- Do you want overlapping bands at some levels and not others?
- How wide do you want your bands? i.e. the distance from the min to the max point at each level?
- The main disadvantage of the polynomial models is their sensitivity to outliers in the data set. Even a single outlier can significantly impact the results and render analysis useless.
If you’ve purchased a template and would like help customising it, then you can book in a 1 hour session with me as a jumpstart to using it and making it work for you particular needs.
60mins - £150+VAT