Pay analytics & compensation research

You have the data.
We build the
analysis.

Send us what you already hold. Offers you made, salaries you pay, people who left.

We turn it into a pay structure you can defend, anchored to public official statistics. You see the finished deliverable first. You decide afterwards.

And if the data cannot leave your building, we ship you the model instead and you run it yourself.

See a full deliverable Get in touch

No licensed survey data. Those licences forbid redistribution, so a report built on them cannot show its own working. Ours can.

Why the analysis matters

A raise is only a raise after inflation.

In 2022 a company in Warsaw awarded its team eight percent. Prices rose by 14.4 percent that year. It was a pay cut, delivered in a letter that said congratulations.

Consumer price inflation, 2022 to 2024

Annual percentage change. Source: World Bank, retrieved August 2026.

4% merit budget

8.0

4.1

2.9

7.9

6.8

3.3

14.4

11.5

3.8

United States2022 · 2023 · 2024

United Kingdom2022 · 2023 · 2024

Poland2022 · 2023 · 2024

Everything below the dashed line is a real increase. Everything above it is a real cut. In Poland in 2022 a four percent budget was worth minus 10.4 percent. The four percent line is an illustration, not a published figure. Your own budget, applied to your own locations, is the first thing we compute.

The work

A complete deliverable, not a capability deck.

Three working models you can pull apart right now, then a finished deliverable. Everything here runs on sample figures. The point is what happens when the numbers are yours.

Live models

Move the sliders. The analysis recomputes.

These are the same models we run against client data, stripped back to the controls that matter. Nothing is submitted anywhere.

Country

Year

Merit budget you awarded 4.0%

0.0%

real change in pay

You gave

4.0%

Prices rose

0.0%

Inflation is real data: World Bank, consumer prices, annual percentage change, retrieved August 2026. Real change shown as the simple difference, the convention used in merit planning.

And here is what a finished deliverable looks like. A full pay architecture for a fictional company of 180 people across three countries, sent in full before anyone asks.

Executive summary page showing the cost to bring everyone to range minimum and the saving from capping base pay — The two numbers leadership actually asks about

Compa ratio governance matrix mapping performance rating against position in range — The matrix that replaces negotiation with a rule

Pay transparency readiness assessment across six required capabilities — Readiness, and the one capability we do not claim

If it looks like something you would put your name on, tell us what you are trying to solve. If it does not, no conversation was needed.

Two ways to work

If you cannot share the data, take the model instead.

Pay data is the most sensitive data in the building. Plenty of leaders want the analysis and simply cannot send salaries to an outside domain, whether that is policy, legal, works council or instinct. That objection usually ends the conversation. Here it does not.

Option one

Send us the data

You share offers, payroll and attrition under an NDA. We build the analysis and hand back the finished deliverable.

Fastest route. Days rather than weeks, because nobody is learning a new tool.
Sharper analysis. We see the whole distribution, so we catch what a template cannot.
An extract, never system access. You choose the fields and export them. We do not connect to your HRIS.
Names are not needed. An employee ID, a grade, a location and a salary is enough for almost everything.
Data returned or destroyed on completion, your choice, confirmed in writing.

Right when the data can leave and you want the answer, not the machinery.

Option two

Take the model. Keep your data.

You send us nothing. We build the model against sample figures, ship you the whole package, and you drop your own numbers in behind your own firewall. We never see a salary.

An Excel model with live formulas, not pasted values. Change an input and every number downstream recalculates.
No macros. Macro enabled workbooks get blocked by enterprise IT, so there are none.
A dashboard as a single HTML file. Double click, it opens in any browser. No install, no server, no admin rights.
Works fully offline. Nothing phones home, nothing is transmitted. It will pass a security review because there is nothing to review.
A worked example pre filled with sample data, so you can see it running before you touch your own.
A one page runbook and a version stamp with file hashes, so a year from now you know exactly what you are looking at.

Most firms cannot offer this. When the product is a licensed dataset, the model is the asset, so handing it over ends the business. Ours is public data and transparent logic, which is precisely why we can give it to you.

What we can build you

Every one of these ships either way, as a finished deliverable or as a model you run yourself.

Interactive

Interactive dashboards
Interactive web pages
Scenario and what if calculators
Excel models with live formulas
Data entry templates with validation
Budget planning tools

Analytics

Attrition analytics
Heatmaps by grade, function and location
Compa ratio and range penetration
Offer acceptance curves
Cost to remediate modelling
Merit budget and real pay analysis
Pay gap grouping structure

Reports

Portfolio reports
Board and CFO decks
Print ready PDF reports
Annual salary guides, white label
Benchmarking summaries
Methodology appendices

Frameworks

Salary bands and job architecture
Level descriptors
Compa ratio governance matrices
Promotion and increment policy
Interview scorecards
Employee explainers
Manager one pagers

This is a starting point, not a menu.

If the thing you need is not on the list, describe it. We will tell you honestly whether we can build it, roughly what it would take, and if the answer is no we will say so rather than reshape your problem into something we already make.

Describe what you need

Not sure which route fits? Tell us what the constraint is and we will say which one we would pick in your position.

The difference

Four things a firm built on proprietary data cannot do.

None of these are claims about working harder. They are structural, and they are the reason this practice exists.

We publish the method

A firm whose advantage is a licensed dataset cannot show you how a number was made without destroying the asset it sells. Ours is public, so the derivation ships with the work. Every source named, every adjustment stated, every date recorded.

We work under your name

White label by default. Your logo, your foreword, your client relationship. We have no interest in being visible to the people you serve, which is exactly why firms that publish their own research use us.

We tell you what we did not assess

Every deliverable carries a limits page. Pay equity testing, individual placement, statutory compliance, all named as out of scope rather than quietly implied as covered. Clients tell us it is the page they read first.

You see the work before you commit

We build the analysis on your data and show you the finished deliverable. Then you decide whether there is a conversation to have. Nobody is asked to buy a description of something.

What we build

Four things, done properly.

01 · Pay architecture

Salary bands and job levelling

Grades, ranges, midpoint progression and overlap. A compa ratio matrix so managers apply a rule instead of negotiating. Level descriptors that settle a promotion argument in writing. Geographic differentials that move the whole range, so a promotion means the same thing in every country.

Built from your payroll, offer and attrition data

02 · Research

White label reports and salary guides

The annual guide your firm publishes, produced for you. Built on public official statistics with the derivation documented well enough that a sceptical reader can check it, which is what makes a prospect forward it internally rather than skim it.

Your brand on the cover, no attribution to us

03 · Readiness

Pay transparency preparation

You cannot publish a pay range you do not have. We build the structure that makes disclosure possible, quantify what it costs to bring everyone above their range minimum, and sequence the rollout so a job posting is not what tells your team.

Includes the remediation cost, computed on your headcount

04 · Communication

The explainer employees actually read

Most companies build a structure and never explain it, so nobody believes it. We write the plain language piece that answers the eight questions people really ask, including the awkward ones about why a colleague earns more.

Written to be handed straight to your whole company

The method

Your data, anchored to public statistics.

Five steps. Each one is recorded in the deliverable, so any figure can be traced back to a source, a date and a stated adjustment.

STEP 01

Anchor

BLS Occupational Employment Statistics, ONS Annual Survey of Hours and Earnings, Eurostat, national statistics offices. Industry cuts, never national averages.

STEP 02

Age

Public wage data is always stale. We age it using the Employment Cost Index, never headline CPI, which is the most common analytical error in this work.

STEP 03

Level weight

Grades mapped to percentiles of the adjusted distribution, interpolated log linearly because wage distributions are skewed to the right.

STEP 04

Blend

Your offers made, accepted and declined. Your attrition. Weighted by how much evidence there actually is, not by preference.

STEP 05

Publish

Sources with retrieval dates, adjustments applied, assumptions written down, and an explicit list of what was not assessed.

Step 04, in plain terms

Five Your data is an anecdote. The public figures do nearly all of the work.
Thirty Equal footing. Your evidence and the published statistics carry the same weight.
Sixty Your data decides. The public anchor becomes a check on it.
A hundred The market you actually compete in is now visible, and it is rarely the market the averages describe.

This is what keeps the work honest in both directions. It stops six data points being treated as truth, and it stops sixty being ignored because a published table felt more official. If your sample is thin we say so, and the structure is marked provisional.

Who we work with

Firms that publish, and firms that are scaling.

Executive search firms

You publish an annual salary guide because your market expects it. We produce it, under your name, so your researchers can stay on searches.

HR and reward consultancies

Capacity is the bottleneck, not demand. We are the production layer behind band structures and job architecture your clients never see us build.

HR technology vendors

The State of Compensation report your category expects. Built so a prospect can check it, which is what makes it travel.

Companies of 80 to 400

The size where pay decisions stop scaling one at a time and every offer quietly becomes a precedent.

What we hold ourselves to

Four words, and what each one costs us.

Values are cheap to write and expensive to keep. These are the four we are prepared to be measured against, defined so you can tell when we have failed one.

Speed

Days, not months. A pay structure inside a week, a research report inside three. The analysis is engineered rather than assembled by hand, which is why the clock is short and the price is not a proxy for effort.

Excellence

Every number traceable to a source and a date. If we cannot show you the working, we do not ship the number. And when something was not assessed, it is named on the page rather than left for you to assume.

Execution

Finished artefacts, not recommendations. You get the model, the deck and the page, ready to put your name on. Nobody has ever been helped by a slide that says consider developing a framework.

Capability

If it can be measured, it can be built. Tell us the question and we will build the thing that answers it, or tell you plainly that we cannot. Both are more useful than a maybe.

You have the data.We build theanalysis.