BattleFin: how the alternative data marketplace works

BattleFin runs a series of alternative data events called Discovery Days and a platform called Ensemble for finding, testing and buying datasets. This guide describes how each works for a data buyer, from BattleFin's own pages, and then turns the scorecard it publishes into a five-question plan for a short provider meeting. It is written for someone at a fund or a corporate data team who has been invited to an event or is comparing ways to find data. Every statement about BattleFin was checked against the pages linked here on 4 October 2026.

Fokals is delivered direct, by REST API and as bulk files, which a buyer loads into its own systems with its own loader, whichever way it met its other providers.

What BattleFin offers

BattleFin's home page describes its aim as helping investment firms find signals, assess data and form relationships, through events and software. It states 1,100+ data buyers, 800+ AI and alternative data providers and 3,000+ alternative datasets. Its pages describe three parts.

BattleFin's announcement of the acquisition describes Exabel as a platform that covers the path from data to insight, with ready-made models and dashboards over datasets already mapped to tickers and room for a user's own data. The rest of this guide is about the events and Ensemble, which are the parts a data buyer meets first.

BattleFin has a separate page for corporate buyers in retail, consumer goods and technology. It lists access to alternative data providers, benchmarking of peer usage, collaboration on custom pilots and a spotlight on new work in areas such as geospatial intelligence and sentiment analytics.

How the events work

The events page says the one-to-one meetings are booked in advance through Ensemble and take place each afternoon, in slots of 15 to 20 minutes that include five minutes to move to the next meeting. Panels and sessions run through the day, and the page also describes AI workshops in which buyers, providers and technology platforms work on research tasks together. The page for finance buyers adds a data science track in which buyers build predictive models and map and tag KPIs, and says events span one to three days. The home page says buyers and providers are screened before they are matched for meetings, and that events include live demonstrations of products. The same pages describe the provider side: showcases for new providers, workshops and a place in BattleFin's Alt Data Roundup newsletter.

That page also lists two annual packages for buy-side firms. One covers the three US events, in Miami, New York and Chicago. The other covers all five global events, adding Asia and London. Each includes up to five tickets per event, the choice of one-to-one meetings with providers, and access to the Ensemble database for one month before and after each event. The page states the price of each package, and it is the place to read the current ones.

How the platform works

BattleFin's demonstration page describes Ensemble as a marketplace, network and software tool for alternative data that takes care of ingesting, monitoring, cataloguing, securing access to and delivering datasets. The membership page sets out the buyer's path in five parts.

1. A catalogue, which BattleFin says its event series and an in-house team of evaluators keep supplied.
2. A profile page for each dataset that sets out what it is, how it is specified and why a buyer might want it.
3. A sandbox, a secure research environment, in which a buyer can try a product's complete history before buying.
4. A purchase through an agreement and purchase process on the platform.
5. A scheduler for meetings with providers, in private video rooms that BattleFin provides.

A buyer also completes a buyer profile so that the platform can show the most relevant datasets. The membership page lists sandbox access, a monthly review call with a data hunter and access to workspaces as membership options, and its text states no price. BattleFin's Ensemble page adds that each data product is evaluated in a standardised process and lists five parts of a Data Quality Scorecard: a metadata summary, univariate analysis, bivariate analysis, time series analysis and data completeness.

Fifteen minutes with a provider

A slot of 15 to 20 minutes, less five to move on, leaves 10 to 15 minutes of conversation. That is enough for five questions of two or three minutes each. The Ensemble page does not define the scorecard headings, so the table below uses them as an agenda and gives, for each, the question to ask and what a documented answer looks like. The examples are Fokals datasets from the hiring dataset, and every answer can be found in the data dictionary and the methodology.

The questions are short, but the tests behind them need a file. Ask for a sample for the companies you track before the meeting, or use the sandbox if the dataset is in Ensemble, and run the two checks below on it.

-- sample_daily is the provider's sample, loaded with one row per company and day.
-- Keys: the count should be zero.
select count(*) as duplicate_keys
from (
  select company_id, day
  from sample_daily
  group by company_id, day
  having count(*) > 1
) d;

-- Completeness: the share of rows that carry each field.
select
  count(*)                                    as rows_loaded,
  count(median_salary_usd) * 1.0 / count(*)   as share_with_pay,
  count(isin)              * 1.0 / count(*)   as share_with_isin
from sample_daily;

Read the second query with the table beside it. A low share with pay is expected where companies do not publish pay, and a low share with an ISIN is expected where a company is private, because a row has an ISIN only when the company or its parent is listed. What you are looking for is a share that the provider cannot explain.

After the meeting, send every provider the same three requests: a sample for the companies you track, the data dictionary and the sourcing statement, plus a due diligence questionnaire if your compliance team uses one. Run the same two queries on each sample and compare the results side by side, because a table of five providers' duplicate counts and field shares says more than five descriptions. Keep each file with the date you received it, so that a later review can show what each provider sent.

Where Fokals fits

An event series is a way to meet many providers in a short time. Fokals is delivered direct, by REST API and as bulk files, and licensed directly, by written agreement, for internal use, embedding in a product or redistribution, and its sourcing statement and methodology are public documents. A buyer who has already met the providers can test a Fokals sample the same way: sample data for the companies you track is sent on request, with the data dictionary and methodology.

How BattleFin compares with another marketplace is set out in Eagle Alpha versus BattleFin, and the tests scouts run on a new dataset are in what alternative data scouts look for. The entry on the data marketplace explains the two kinds a buyer meets.

Frequently asked questions

The queries and code on this page are examples to adapt. Test them in your own environment before you rely on them.

What this page says about the products it names was checked against their public documentation on 4 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.