predictive model, e-commerce
A customer retention model on a large multi-country e-commerce dataset. It clearly outperformed the baseline and turned into concrete advice on acquisition and retention spend.
Los Angeles, California
Most companies already have the data. Few can see in it where margin grows and costs come down. We help you find those answers and put them to work.
What your data can show
We go through your sales, costs and operations data and show — in plain numbers, with ranges and assumptions spelled out — where the money is leaking and how much is realistically on the table.
Which products, customers and channels earn their keep, and where pricing or mix leaves margin behind.
Where spend grows faster than results: suppliers, overtime, returns, discounts, waste.
Which customers buy more, which are about to leave, and what brings them back.
Stock that sits too long, items that run out too often, and the cash tied up in both.
Demand and cash-flow forecasts with honest ranges, so plans stop being guesses.
Early signals in the numbers before a problem shows up in the bank account.
Every finding comes with the data behind it and an estimate of potential impact as a range. We do not promise results we cannot measure.
Start with a free data reviewWhat we do
We find what is broken, duplicated or missing in your data and agree on the metrics that matter — so every report tells the same story.
Learn more →We build reporting your team opens every week: clear metrics, refreshed automatically, in the tools you already pay for.
Learn more →Forecasts, customer and pricing models that turn history into decisions about what to stock, whom to keep and what to charge.
Learn more →Deep, data-led studies for niche markets, starting with poker staking funds and tournament economics.
Learn more →Why us
You work directly with the people who do the analysis — no hand-offs, no account managers. We stay with you from the first data review to the moment your team runs the result on its own.
We are not tied to any software vendor. We recommend what fits your data and budget.
Every number comes with its source, its assumptions and how sure we are of it.
Code, models and dashboards are handed over. No lock-in.
Industries
Margin by product and channel, customer retention, pricing and promotions.
Stock levels, supplier performance, demand forecasting.
Profitability by client and service line, capacity, cash flow.
Analytics for poker staking funds: tournament selection and makeup forecasting.
→Work & research
predictive model, e-commerce
A customer retention model on a large multi-country e-commerce dataset. It clearly outperformed the baseline and turned into concrete advice on acquisition and retention spend.
public dashboard
An interactive dashboard on airline delays and cancellations. It shows how we lay out operational metrics so people actually use them.
market study
Our study of online tournament economics by buy-in, field size and room. Methodology and limitations will be published with it.
How we work
01
We agree on the business question, the data and what "done" looks like before any work starts.
02
A fixed price for a fixed scope, in writing. If the scope changes, we say so first.
03
We define upfront which metrics will show whether the work paid off.
04
Regular check-ins with interim findings. No black box, no surprises at the end.
05
NDA by default, access only to what we need, data removed after handover.
06
Code, notebooks, dashboards and a walkthrough, so your team can carry on without us.
Technologies
We build on the stack you already have, or recommend a simple one if you are starting from spreadsheets.
Team
Dmitry Shichko ran a retail and service business for over a decade before moving into analytics, so we read your numbers the way an owner does. A dedicated head of research leads our research practice.
We will look at your situation, help you weigh the options and recommend the most sensible next step — even if that step is not working with us.