We build the models and tools that get it out: pricing a new part from the ones you have already quoted, cutting collection rounds without cutting tonnage, holding stock that matches demand, bidding what a slot is actually worth.
New: Really European? We traced who owns the products on Europe’s buy-European directories, from company registers.
Quote Matcher
measured on 1,000 parts held out of 4,000
price anchor, median error
±0.0%
against ±61% from a regression baseline
8 of them, all the same shape underneath: a set of options, a constraint, and a better way to choose than the one in use.
Price a new enquiry from the parts you have already quoted.
Contract manufacturers hold years of quotes and almost never use them. An estimator searches by memory, so the same part gets priced three different ways in a year. Matching a new enquiry against historical parts on physical attributes — material family, envelope, tolerance, finish, quantity — turns that archive into a price anchor, in seconds, with the reasoning shown.
Run it on a partRebuilt model turned a projected loss into profit on the same fleet
Component-level forecasting across a multi-warehouse network
Auction bidding at scale, and a consumer marketplace of our own
A rebuilt operating model that turned a projected loss into a profit
Rejected findings published alongside accepted ones, with reasons
Browser extension fixing legacy ERP exports, used daily
Operating models, scenario analysis, and the paper that carries them
Aerospace machining
Estimators priced new enquiries by searching old quotes from memory. Nothing was reused systematically, so similar parts left the building at inconsistent prices and every enquiry cost hours.
Waste and recycling
The operating plan showed a loss and the conclusion was that the contract was unviable. The plan rested on three assumptions that had never been checked against the vehicles actually in the yard.
Global online marketplace
When placement is sold by auction, bidding low puts you under the fold on a page the customer had already decided to buy from, and bidding high means paying retail for demand that was coming anyway. What a placement is worth varies by listing, by market and by device — none of which is visible from the bid side.
Live in your browser, nothing uploaded
The part-matching engine, running as a public demo. Drop in your own quote history and it matches against that instead — the file never leaves the tab, because the whole thing runs on your machine.
Also live, also in your browser
Every engagement starts with a file somebody apologises for. Drop one in and the Data Clinic names what is actually wrong with it — duplicate rows, two decimal conventions, dates in two orders, the same value spelled three ways — and hands back a cleaned copy.
For financial entities
Drop in the report package and it runs the EBA's own rule set — keys, closed lists, LEI check digits, package format — then reads the register back: where the concentration sits, which providers are designated critical, where the data rests. In your browser; nothing is uploaded. In the regulators' dry run, 93.5% of registers failed at least one check.
Rejected already, or not sure what goes where? The register guides cover the rejection codes, the templates, provider identifiers and each regulator's deadline.
Dataset
Public registers that share no identifier, joined. 3,300 legal entities resolved to corporate groups, rolled up to the parents they report, and matched to EU research funding. Free to download, with the rows it could not place published alongside the rows it could.
20 published pieces, mostly things that had to be worked out for a project and were written down so they would not have to be worked out twice.
Anyone can write a number on a website. The difference worth paying for is whether it survives being looked at, so everything here is built to be looked at.
Analytics Ascent is a small team. Engagements have covered aerospace machining, waste and recycling, hardware manufacturing and online marketplaces — and the people who write the code are the people you talk to.
The interesting problems arrive as a spreadsheet nobody trusts and a question somebody needs answered on Thursday. That is the normal starting point, not a reason to wait until the data is tidy.