High-performance infrastructure to activate your customer data

Batch Data: the data infrastructure at the core of the Customer Engagement Platform

The best campaigns are won before you hit send: with clean data, available the moment you need it. That's the job Batch Data does. Connect plugs into every source you have, from your apps to your data warehouses. Unify brings it all together in one place: profiles, behaviors, and catalogs. Activate puts it to work in real time.

Batch isn't a customer data platform (CDP) in the traditional sense. It connects, unifies, and activates your data directly, with no export step to a separate activation tool.

Batch Data Connect

Connect wires Batch into your entire stack, in and out. Collect events from your apps and servers in real time through SDKs and the API. Sync your data warehouses and files with Cloud Sync. Plug in your business tools with ready-made connectors for CDPs, analytics, and ad platforms.

Batch Data Unify

Unify pulls everything into a single customer profile. It reconciles anonymous and identified journeys through cross-device identity resolution, keeps every behavior queryable in real time, and links your business Catalogs to profiles. From there, it keeps enriching each profile continuously, from simple aggregates to predictive scores.

Batch Data Activate

Activate puts that unified profile to work, in real time. It drives advanced customer segmentation, personalizes every message, and triggers marketing and transactional scenarios the moment a behavior or attribute changes. No intermediate sync. No delay.

European infrastructure

Our infrastructure runs entirely in Europe, on our own bare-metal servers. Your data stays under European jurisdiction, out of reach of the Cloud Act. It's an architecture built to meet what Data, IT, and Security teams actually need.

Built for scale

Batch processes 20 billion events a day and keeps 1.2 billion active profiles live every month, across more than 600 bare-metal servers. A distributed architecture, built on Kafka, Cassandra, ClickHouse, and Golang, for performance that holds steady at scale.

Batch Data Connect: connect every source, find every interaction across your systems

Collect your first-party data in real time, straight from your apps, sites, and servers. Sync data from your warehouses, files, and platforms on whatever schedule you choose. Then sync deliveries, opens, clicks, profiles, and segments back to your data stack. That's customer data integration in practice: data moves both ways between Batch and the rest of your stack, no duplicate copies.

Batch fits into your existing architecture. No new data model to learn, no lock-in.

Real-time collection

From your apps, websites, and servers, Batch ingests profiles, attributes, events, and subscriptions in real time.

SDKs built for production

Batch SDKs rank among the top 100 most-used in the world, and they keep pace with every shift in web and mobile tech. They slot into your existing dev and release workflows: standard frameworks, standard package managers, detailed changelogs, and migration guides for every upgrade.

Automatic device-level data collection

Batch SDKs pick up the technical and contextual data you need for targeting automatically. No extra tagging required.

• Locale

• Language

• Country

• City by IP

• App version

• Last visit

• Last location

Integrations built for your stack

iOS Swift · Objective-C

Swift Package Manager · CocoaPods · XCFramework

Android

Kotlin · Java · Gradle · Maven Central

Cross-platform

Flutter · React Native · Expo · Cordova · Ionic

Web

JavaScript SDK. Web Push through a JavaScript snippet or a service worker. Runs on websites and Progressive Web Apps alike.

Profile API: connect all your server-side flows

The Profile API feeds Batch with Profiles, attributes, events, and subscriptions straight from your backends, streamed or in bulk.

• Profiles created on the fly

• 300 updates a second in real time

• 10,000 profiles per call

Feed Batch from your data stack

Connect your warehouses, files, CDP, and analytics platforms once. From there, your data syncs into Batch on its own, no custom pipeline to build or maintain.

Cloud Sync for data warehouses

Batch connects natively to your modern data stack. Feed it straight from your data warehouse, no extra development required.

• Sources: Snowflake, BigQuery, Amazon Redshift, Databricks, ClickHouse, Microsoft SQL Server

• Features: Sync scheduling, incremental mode, setup dashboard

Cloud Sync for files

Connect your SFTP repositories and cloud object storage, then set a cron schedule for each import. Batch fetches and ingests new files on its own.

• Automated imports

• Multi-file support

• Sync scheduling

Marketplace connectors for your business tools

Connect Batch to the tools you already run. Turn on ready-made integrations with your CDP, analytics platforms, data clouds, and business tools, no custom connector required.

CDP & data activation

Segment, mParticle, Tealium, Zeotap, Treasure Data, mediarithmics, Commanders Act, Hightouch, DinMo

Analytics & attribution

Amplitude, Mixpanel, Piano Analytics, AppsFlyer, Adjust, Branch

Marketing platforms

Salesforce Marketing Cloud, Piano Composer

And more

→ Discover all our partners

Find all your Batch data in your systems

Export profiles and segments, push events to your analytics tools, and sync the engagement signals Batch generates back to your data warehouse. Your data stays available in your stack, for your own analyses and models. This kind of customer data integration runs both ways: what Batch produces flows back into your stack too.

Every send produces engagement signals

Sends, deliveries, opens, clicks, views, every interaction with a message, Batch captures them automatically. That data is instantly available inside Batch for targeting and analysis, and it streams to your stack too.

• Sent

• Delivered

• Opened

• Clicked

• Bounced

• Unsubscribed

• Direct Open

• In-App interactions

Cloud Sync: bidirectional sync with your stack

Cloud Sync doesn't just feed data into Batch. The same connections to your warehouses and file storage also send what Batch produces back out to your stack.

• Cloud Sync for data warehouses

• Cloud Sync for files

• Bidirectional sync

Export your data on demand or in real time

Export Profiles, segments, and events through the API, or straight to your storage infrastructure. Events collected by the SDKs can go straight to your analytics tools and other destinations too.

• Export API

• Cloud Sync

• Event dispatchers

A toolkit for developers

Validate payloads, catch errors instantly, and see exactly what data Batch actually has. Diagnostic tools cover the whole ingestion chain, from the first API call to checking a profile live in production.

• Detailed API error handling

• Profile View

• Data Catalog

• Attribute and subscription inspection

• Event timeline

Batch Data Unify: unify identities, behaviors, and business data around Profiles

Batch reconciles the signals from every channel into one data model. Identity, attributes, subscriptions, events, and business data become immediately queryable for targeting, personalization, and orchestration.

Your customer data, reconciled

Cross-channel identity resolution

Batch matches a single user's interactions using whatever identifiers are available at each step: account ID, email, phone number, and so on, all deterministic. Apps, websites, forms, in-store checkouts, any touchpoint, they all feed the same profile the moment an identity link is established.

Anonymous → identified reconciliation

Batch starts collecting activity from a user's very first visit, on an anonymous profile. Once they log in, that activity merges into their identified profile. Screens viewed, products browsed, events fired before login, none of it gets lost or duplicated.

Inspect any profile

Find the attributes, installs, subscriptions, and events tied to any user, right in the Batch dashboard. Profile View lets your Data, Support, and Product teams see exactly what Batch knows about a profile, at a glance.

• Profile View

• Attributes

• Subscriptions

• Event timeline

Granular behavioral history

Batch keeps every interaction in its Events Store, timestamped, with every associated attribute. Nothing gets aggregated away: raw data stays available for targeting and analysis. 36 months by default, and you can adjust that to fit your privacy constraints and use cases.

All your business data

Every product, piece of content, store, or offer you have, ready to activate.

Structure your business objects in Catalogs

Products, content, stores, agencies, flights, offers, they all live in Catalogs, with every attribute attached.

Reference your objects by ID, no duplication

Your events and profiles carry a SKU or a product ID. The actual properties, price, stock, image, live in the catalog. Batch resolves the reference at send time, so you update the object once instead of copying its fields into every payload.

Store millions of items

Store your entire product catalog in Batch: millions of items, with every variant and market-specific version.

Batch enriches your customer data: from predictive scoring to everyday aggregates

From the events your SDKs and APIs send, plus your engagement data, Batch enriches everything by creating new targetable attributes, no work required from your technical roadmap.

AI Predict: 15 built-in predictive scores

Available natively in the Batch CEP, these scores help you optimize CRM campaigns and lift performance: decisive churn date, top at-risk customer, promotion sensitivity, and more.

Targetable aggregates out of the box

Batch computes aggregates straight from user behaviors, over whatever rolling window you set: 7, 30, 90 days, or anything else. Number of purchases, average basket, visit frequency, you name it. Marketing teams build them with simple operators: sum, count, average, min/max.

The right model for the right problem

Predicting churn from a purchase history and writing an email subject line: two very different jobs. Batch uses whichever technology handles each one best.

AI Predict: predicting behaviors

Machine learning, with the algorithm chosen for each score.

• Trained on your data, siloed per customer

• Quality tracked via AUC, lift, and drift

Every AI Predict model trains on your brand's own data, in your own training environment. No learning ever crosses between accounts. Batch's team handles feature engineering, algorithm selection, and training, end to end.

AI Assist: generating content

Powered by the best third-party LLMs available, chosen for each use case and updated regularly.

• Anthropic, OpenAI, Google, Mistral

• No personal data ever transmitted

• Quality controlled through evaluators and golden sets

AI Predict: models and use cases

• Top at-risk customer: Classification, LightGBM

• Subscription churn: Classification, LightGBM

• Decisive churn date: Time-to-event, LSTM (deep learning)

• Lifetime Value: Regression, LightGBM

• High-potential customer: Classification, LightGBM

• High-potential prospect: Classification, Logistic regression

• Product affinity: Classification, Logistic regression

• Cross-sell affinity: Classification, Logistic regression

• Promotion sensitivity: Classification, Logistic regression

• Product recommendation: Ranking, ALS (collaborative filtering)

• 2nd purchase date: Regression, LightGBM

• Repurchase date: Regression / Classification, GBTRegressor / Logistic regression

• Promotion recommendation: Ranking, LightGBM

• Best send time: Statistical computation

• Best channel: Statistical computation

AI Assist: use cases

• Conversational Agent: Analysis, message creation, recommendations. Third-party LLM + RAG

• AI Home: Analysis & recommendation. Third-party LLM + RAG

• Writing A/B test variants: Generation. Third-party LLM

• Translating a campaign: Translation. Third-party LLM

• Adapting tone: Rewriting. Third-party LLM

• Naming scenario steps: Summarization. Third-party LLM

Transform, monitor, and govern your data

Schema and governance, built right into the platform.

Stay in control of your data schema

Types, names, sources, archiving, every attribute Batch collects gets typed, named, and tied back to its source. The Data Catalog shows you what's actually in the platform: which attribute, fed by which flow, holding which values.

You stay in control over time, too. Rename an attribute that got a bad name at implementation. Spot two fields doing the same job. Archive the ones left over from a deprecated app. All of it happens right in the dashboard, no support ticket, no ETL rewrite.

Monitoring & alerts: your implementation under watch

Batch watches your flows and attributes around the clock. A drop in ingestion volume, a field going empty, duplicate attributes popping up, you'll know in time to act.

Govern data throughout its lifecycle

Consent, preferences, retention rules, deletion, manage all of it in Batch through a dedicated set of tools: Privacy center, Inactive Profiles, GDPR API, and more.

Batch Data Activate: an activation mode for every use case

Activate all your data, with no new technical work

Customer segmentation sits at the core of Activate. Dynamic segments, behavioral targeting, one-off or mass sends, advanced personalization, predictive scores: your marketing team builds campaigns on their own, using the data you've already connected.

Dynamic segments

Filters run against the whole base at execution time. Nest them, reuse them, from one scenario to the next.

Static audiences

Populations built elsewhere, imported and combined with any other criteria you want.

Real-time reach preview

Watch your targeted volume update live as you change your criteria, before you ever hit send.

Query builder

Your CRM team builds customer segmentation as precise as an advanced SQL query, without writing a single line of it.

Behavioral targeting

Query the Events Store directly, over whatever time window you set at targeting time.

Retargeting

Turn your campaign interactions into criteria: non-openers, link clickers, profiles you've already reached.

Automation triggers

Fire on event arrival or on an attribute change.

Marketing and transactional messages

One platform for both: fast delivery and stronger traceability for transactional, consent management and performance tracking for marketing.

Real-time personalization

A Jinja-like engine built for performance and real time, easy to use, with a preview dashboard.

• React to a customer reaching VIP status: trigger on attribute change. No build, change detection is handled by Batch

• Combine three conditions and a behavior: query builder, nested conditions. No build

• Re-target Tuesday's non-openers: retargeting. No build

• Target customers in a city: built-in attribute criterion. No build

• Check volume before sending: reach estimate. No build

• Show today's price in the email: personalization via Catalogs. Catalog fed once

• Reuse your in-house model's population: Audience import. One export to schedule

• Target a table in your warehouse: Cloud Sync for Warehouses, incremental. One sync to configure

A European infrastructure

Your data under European jurisdiction

Your data gets collected, stored, and processed on bare-metal infrastructure Batch operates itself, in European data centers. The contracting entity is a French company. None of it ever trains a shared model.

• A US vendor with a European data center: US-law company, servers in Europe. Exposed, server location doesn't change that

• A European vendor hosted on a US hyperscaler: European company, US hosting provider's cloud. Exposed, via the host

• Batch: French company, its own servers in Europe. Out of reach

The Cloud Act lets US authorities request data from any company under US law, wherever that data physically sits. Getting out of its reach takes more than picking a European data center: neither the vendor nor the host can fall under that jurisdiction.

The same standard, applied to AI

• AI Predict: proprietary ML / deep learning models, your personal data siloed per customer, on Batch's servers in Europe

• AI Assist: third-party LLMs (Mistral, Anthropic, OpenAI, Google...), no personal data is transmitted

And all the fundamentals, of course

• SSO

• Fine-grained access control

• Access history

• 3FA bastion

• DDoS mitigation

• IDS/IPS

• Pentests

• Bug bounty

• Internal and external audits

• GDPR compliance

• No model shared across customers

A platform built for scale

20 billion events a day, 99.9% uptime

• 20 billion events a day

• 1.2 billion active profiles per month

• 99.9% uptime

• 100 million emails per hour

• 540 million push notifications per hour

• 500,000 SMS per hour

Distributed architecture

300 applications, spread across 5,000 instances, on a private network with no exposure to the internet.

Technologies chosen for volume and speed

Kafka for queuing, TiDB and Cassandra for the profile database, ClickHouse for events and send statistics.

Dedicated bare metal

Servers with no sharing with other companies, no heavy virtualization layer on top. Less performance variability, more stability under heavy load.

Environment isolation

Separate PROD and DEV environments, with advanced debugging tools.

Migrating to Batch

200+ CRM migrations from every platform on the market

Every CRM migration follows the same well-worn path, whatever platform you're coming from: Salesforce Marketing Cloud, Adobe Campaign, Braze, Airship, Selligent, Emarsys. Audit the existing setup, map the data model, migrate profiles and opt-ins, warm up your IPs, cut over.

→ Discover our case studies

700M active profiles migrated

Profiles, consents, catalogs, audiences, your data gets migrated and remapped into the Batch model. Your source systems stay right where they are.

88 NPS on migration

Direct escalation to technical teams, 96% of tickets closed in under an hour, a 3-minute median response time, in 5 languages. 97.6 support CSAT.

Your integrator already knows Batch

Working with an agency? They run the CRM migration project with dashboard access, a shared channel, and a Batch Implementation Manager keeping an eye on things.

Avanci, BeApp, BeTomorrow, Cartelis, Claravista, CustUp, Degetel, Dernier Cri, Jakala, Modeo, Niji, Relatia.

→ Discover our agency partners

99.7% email deliverability

Automated IP warming, reputation tracked daily, and direct relationships with Google, Microsoft, Yahoo, Signal Spam, and Spamhaus.

FAQ

Is Batch a CDP?

Not quite, even though it shares plenty of features with CDPs (customer data platform): identity resolution, a unified customer profile, enrichment. A CDP exists so your data team can distribute data across your stack. Batch's data platform is built to serve marketing activation use cases directly.

Do I need a CDP to use Batch?

No. Batch connects straight to your technical stack, both ways.

Inbound: SDKs and the Profile API for your apps and backends, Cloud Sync for your data warehouses, SFTP, and object storage. No ETL required.

Outbound: the Export API, bidirectional Cloud Sync, and event dispatchers send profiles, segments, and engagement signals back to your stack. Whatever Batch produces stays available for your own analyses and models.

And if you already run a CDP, Batch connects to it directly: Segment, mParticle, Tealium, Zeotap, Treasure Data, mediarithmics, Commanders Act, Hightouch, DinMo, and more.