Pricing and promotion strategy for profitable revenue growth, built on what your data actually shows.
I’m Fred Puech, a PhD economist and data scientist. For 20 years I have built pricing, promotion and revenue growth strategies for companies whose data is too large and too tangled for a spreadsheet: thousands of products, millions of customers, billions of transactions. I separate what your prices and promotions actually did from everything else that moved, and turn the result into decisions your leadership team can act on and defend.
Sales after a price increase
What actually moved the number
Change in same-store sales, decomposed by driver
Restaurant brands · 2027 pricing
Your 2027 price increase is a harder call than the last three.
For multi-unit restaurant brands and their PE sponsors: how much price to take in 2027, where to take it, and how to know within weeks whether it is working. The work takes 5 to 6 weeks and is done before prices go live.
The test
One objective: profitable revenue growth
Every question I take on, and every model I build, has to answer one thing: will this decision grow profit, not just revenue? If an analysis can’t change a decision, I don’t run it. If a decision won’t pay, I’ll tell you, including when the right answer is to leave prices alone.
That focus is why clients bring me in before a major price move, after one that didn’t land, and when margins are slipping without a clear cause. It is also why the work stays lean, fast and tied to the P&L.
The technical half
Data science that goes past the spreadsheet
Most pricing analysis compares before and after, averages across the chain and calls the difference the effect of price. That is how a promotion that only pulled sales forward looks like a success, and how a price increase gets blamed for a traffic decline that hit the whole category.
I work on your full transaction history, often billions of records across products, customers, locations, channels and time. I build econometric and machine learning models (panel regression, causal inference methods such as difference-in-differences and synthetic control, Bayesian hierarchical models, machine learning demand models) that estimate price elasticity, promotion lift, cannibalization, discount leakage and mix effects at the level where decisions are made: by item, customer segment, account, location or channel. When your data cannot answer the question, I say so, and I design the test that will.
Causal inference
Before/after overstates the damage
Weekly sales index, stores with the price increase vs. what would have happened without it
Models
Estimates
At the level of the decision
The executive half
Strategy argued at the executive table
The model is the starting point. What clients pay for is the decision: which prices move, where, by how much, which promotions to cut or redesign, and what result to expect, with the range around it.
I present that to CEOs, CFOs, boards and PE sponsors in their terms: margin, traffic, risk and timing. Your team keeps the logic behind every recommendation, so the next decision is easier than this one.
Decision brief
Price increase options, next 12 months
Expected gross margin impact with its 80% range, and traffic risk
Data complexity
Where simple analysis breaks down
Typical data volumes on past engagements
Fit
Where I do my best work
The common thread across my clients is not an industry. It is data complexity: hundreds to thousands of products, thousands to millions of customers, and transaction histories that run into millions or billions of rows. That is where simple analysis breaks down and where most pricing advice stops being reliable.
I have done this work in more than 20 industries, on both sides of the B2B and B2C line. The main ones:
Consumer
- Restaurants and QSR
- Consumer packaged goods
- Alcohol beverages
- Retail
- E-commerce and DTC
- Spare parts (for consumers)
B2B
- Industrial manufacturing and distribution
- Food manufacturing
- Chemicals
- B2B software
- Associations
- Spare parts (B2B industrial equipment)
Breadth is what makes the depth useful. A distributor’s 50,000 SKUs and a restaurant chain’s menu across 800 locations pose the same core problem: estimating how demand responds at a granular level, with noisy data, and turning the answer into rules people can apply. Having solved it in many settings, I recognize quickly which patterns in your data are real and which ones every business in your position tends to misread.
Timing
When clients call me
A major pricing decision is coming and the downside is unclear.
A price increase, a new price architecture, a discount policy, next year’s promotion calendar. You need to know what it will do to volume and margin before you commit.
Pricing Decision ScienceA price move or promotion didn’t do what you expected.
Sales softened, customers left, or a campaign underdelivered, and the next decision is already on the calendar.
Independent Pricing ReviewMargins are slipping and pricing isn’t fixing it.
List prices, discounts, promotions, loyalty and fees have been decided one at a time, by different people, and now work against each other.
Pricing Decision DesignHow we work together
Engagements
| Engagement | What you get | Typical length |
|---|---|---|
| Independent Pricing Review | An objective read of what a past price or promotion decision actually did, separated from mix, seasonality, competition and market effects. A clear view of what to keep, fix or reverse, and what the next decision must account for. | 2 to 12 weeks |
| Pricing Decision Science | One upcoming decision designed on your data: the options, the expected impact on volume and margin with its range, a recommendation, and a plan to measure the result. The models stay with your team for the next decision. | 8 to 12 weeks |
| Pricing Decision Design | Your pricing rebuilt as a system: price architecture, discount and promotion policy, the models and tools behind them, and the governance that keeps them working (decision rights, a pricing calendar, guardrails, post-decision reviews). I lead the design and work alongside your team until it runs without me. | 6 to 24 months |
Evidence
What it has delivered
| Client | Situation | Result |
|---|---|---|
| Consumer Quick-service restaurant chain | Guest count down more than 12% after mis-targeted price increases | 12 consecutive quarters of price increases without traffic loss |
| Consumer Direct-to-consumer brand | Price increases, discounting, loyalty and shipping fees working against each other | 400 bps incremental gross margin over 24 months |
| B2B B2B online parts distributor | Thin margins, tens of thousands of SKUs, cost-plus pricing | 7% pocket margin improvement with no volume loss |
| B2B B2B logistics provider | Contract discounts negotiated case by case, weak policy compliance | 4% margin improvement; discount model now prices every new customer |
| Consumer Retailer | Sales softened more than expected after a price increase, with the next move pending | Pricing validated: the decline came from elsewhere, no broad demand erosion |
| B2B B2B industry association | Flat target-margin pricing, some services never repriced | 15% of members repriced, 5% revenue improvement |
| B2B Industrial equipment manufacturer | Price increases not reaching the invoice | Margin-floor and repricing rules worth 1.3% to 4.2% of revenue |
| Consumer Private equity firm | Two acquisition targets, unclear pricing power | One investment declined on growth-potential grounds in 4 weeks |
Clients
Who I work with
CEOs and CFOs
Get a clear recommendation with its expected margin impact and its risk, ready for the board.
CMOs, sales and commercial leaders
Get price lists, discount structures and promotions that grow volume and margin together, not one at the other’s expense.
Heads of pricing, revenue growth and analytics
Get a peer who can go deep in the data and help carry the result upstairs.
PE operating partners
Get an independent read of pricing power, in diligence or in the first year of ownership.
Method
How I use AI
I use machine learning models in the analysis where they outperform simpler methods, and I validate them like any other model. I use commercial AI tools for desk research and to find public data. I do not deliver AI-generated recommendations: every recommendation is mine, built on your data and explained in terms you can check.
Free 60-minute working session
Bring one decision.
In a free 60-minute working session, we look at one pricing or promotion decision you are facing, what your data can and cannot tell you about it, and what a good decision would take. No pitch deck.
FAQ
Frequently asked questions
What does Keenalytix do?
Keenalytix is an independent pricing practice led by Fred Puech, PhD. It builds pricing, promotion and revenue growth strategies for companies with large product, customer and transaction data, in both consumer and B2B industries, based on econometric analysis of the client's own transaction data.
Who is it for?
CEOs, CFOs, commercial and pricing leaders at mid-market and enterprise companies with hundreds to thousands of products, thousands to millions of customers or large transaction volumes, in restaurants, retail, consumer goods, manufacturing, distribution and software; and private equity firms evaluating or owning those businesses.
How is this different from a pricing software tool or a large consulting firm?
Software gives you a model's output; large firms give you a team. Keenalytix gives you a senior economist who builds the models on your data and makes the recommendation himself, with a single test: does it grow profit.
Do you use AI?
Machine learning models are used in the analysis where they outperform simpler methods, and AI tools help with desk research. Recommendations are never AI-generated.
How do engagements start?
With a free 60-minute working session on one pricing or promotion decision. Engagements then run from a 2-week review to a multi-month pricing transformation.