About Fair Offer
Fair Offer is a compensation model for US-based tech companies.
May 08, 2025
IntroductionSome FYIs about Fair OfferWhy is it free?Why shouldn’t I just use benchmarks?I’m a Total Rewards / Finance / People Operations Leader. Can I use this for my team?I’m a compensation platform / consultant / other. Can I license Fair Offer?
Introduction
"All models are wrong, but some are useful." — George Box
It’s an open secret that benchmark databases are hit-or-miss: great coverage for some roles (mid-levels in engineering, product, etc.) and some locations (SF, NY). But how much equity should you grant a Staff Engineer in Memphis at a Series B company? "Error, low sample size."
Fortunately, it turns out this is a very solvable problem with data science. We analyzed thousands of data points across public and proprietary data to design a predictive model that produces 4+ million combinations: 200+ titles, 600+ locations, and 30+ company stages (seed-stage to FAANG). Check out our white paper & FAQs below for more detail.
For a full breakdown, please read our series of white papers:
But in short, Fair Offer is based on the following insights:
- At a given company, within a given job family: both total compensation and base salary follow a fixed “progression” for each level (e.g., L7 Sr. Manager makes 1.18x an L6 Manager in base salary, and 1.23x times in total compensation)
- This progression varies between maturity / stage / size of company, and is the primary determinant in compensation philosophy (dictated by the “peer group”). E.g., a 10 person company that is bootstrapped or seed-stage company might have a 1.16 base / 1.20 TC progression, while Google has a 1.21 base / 1.35 TC progression.
- As a direct result, entry level roles are similar regardless of company stage, but the more senior the role, the more compensation diverges.
- Different job families have the same progression when considering the peer group, but their levels are offset from each other (L5 of function A = L6 of function B, L6 of function A = L7 of function B, etc.) The benefit is that by gathering data or patterns about just one job family, you can extrapolate the entire compensation framework.
- Cost of labor across geographies varies, partially correlated to cost of living. So for example within the US, if a location is 30% cheaper than NYC, the same role might pay 10% or 15% less.
Some FYIs about Fair Offer
Fair Offer is not a benchmark, it is an algorithm.
Taking an algorithmic approach allows Fair Offer to extrapolate 5+ million combinations, from only a few thousand data points of training data.
This means that each individual result isn’t backed by a range of data points; rather, the entire algorithm is trained on compensation data points that we’ve observed in the market. On an ongoing basis, we’ll continue to refine the algorithm, and it will be updated due to trends. For example, we expect that the inputs and weights will be adjusted over time due to inflation, cost of living / labor across cities, increase or decrease in demand for specific roles.
We are currently on the third public version, published May 2025. We published v2 in Jan 2024, and v1 before that in early 2023. We will continue to evolve & improve over time.
Please provide feedback! Expect a new version of the Fair Offer Standard published every ~12 months.
It is meant to be a “template” that companies customize & implement as a framework.
While it might be a useful estimate for a role in a pinch, the best way to run compensation is to implement a leveling guide. Think of it like a “benefits policy” — just like you might look at an industry survey for employee health benefits but then the key step is to set your policy, Fair Offer and benchmarks in tandem are a more holistic way to inform your leveling & compensation policy.
It is heavily tailored towards startups & tech companies based in the US.
This is because these companies allocate a significant portion of compensation to equity — which increases complexity, creating the greatest room for improvement. While there are probably parallel algorithms for investment banking, or electricians, or retail workers, the scope of the current Fair Offer version is to cater to companies that consider equity to be a significant component of compensation.
Why is it free?
Because we care about transparency and making an impact in the market. We think every employer should find it easy to make balanced, reasonable, market-competitive offers without spending tens of thousands of dollars; and that every employee should feel and know that they’re being fairly compensated without spending hours on research.
Plus, we hope you spread the word, and Fair Offer will be helpful leadgen for Autograph’s paid product (Headcount Planning & Management).
Why shouldn’t I just use benchmarks?
Benchmarks are an important input! But they do not (by themselves) allow hiring managers / recruiters to generate a clean offer out of the box, for four major reasons:
- Spotty data, especially for globally distributed workforces
- An isolated benchmark without the context of a comp philosophy or framework makes it harder to communicate compensation to employees and candidates… not to mention the compensation “journey” as the employee grows and company advances.
- Wide, imprecise ranges resulting in muddy data (e.g., in salary bands — $150-220K base; or company stage — $100-250M raised)
- Benchmarks go stale quickly because each combination (e.g., "Staff engineer in Memphis TN at a Series B company") needs new data to be updated. However, a data science approach allows you to detect trends in the underlying factors (staff engineer trends x Memphis trends x Series B trends) to better predict what a competitive, market offer should look like. Skate to where the puck is going, not where it's at.
Based on hundreds of conversations with HR teams, compensation experts, hiring managers, and executives, we’ve devised a solution to these deficiencies in how compensation is handled today.
I’m a Total Rewards / Finance / People Operations Leader. Can I use this for my team?
Absolutely! Fair Offer is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license. You’re welcome to use it in your company (you just can’t directly profit off of it, e.g., by modifying or licensing it to someone else).
We would also appreciate if you were to credit us when you do use it, in line with the license :). The primary reason for this is because we are aiming to make it not just a handy tool but a compensation Standard, with the goal of enabling transparency in the compensation market. Whenever someone receives a Fair Offer, they should be able to immediately think “ah okay, I understand what this means, and I don’t have to stress about whether it’s in line with the market.” This will create the same benefits as a standard home purchase agreement, or a standardized NDA, etc. Both parties can arrive at an agreement sooner, in a more trusting way. To make this easier, you can even click the
Share button in the top right corner of any result.I’m a compensation platform / consultant / other. Can I license Fair Offer?
Let’s chat about it! Contact us here.
Share article
Write your description body here.
