Datacook
Datacook is a French Customer Data Platform that collects, unifies, cleans and activates customer data using AI matching and twenty predictive scores, aimed at marketing and CRM teams in consumer-facing companies with large customer bases.
What is Datacook?
Datacook is a Customer Data Platform published by the French company FIDELIZ, based in the 9th arrondissement of Paris, and sold under the Datacook brand. The publisher positions it as the first AI-native CDP and as the French Customer Data Platform powered by AI. It was founded by two scientific profiles: Elodie Vaillant, CEO, from marketing and CRM, and Baptiste Pigaux, from data science, presented as CTO on the team page and as CDO in the company's press release.
The platform is described in five blocks. Reconciliation comes first: a symbolic recognition AI identifies tables, fields and values, absorbs any format and sorts them into a data model, then reconciles identities with fuzzy-matching algorithms to produce a 360-degree customer view. Data quality follows: continuous correction, reformatting, repair of broken joins, de-duplication and categorisation, with outliers reported and alerts raised on missing values. Self-integration is the central claim: Datacook says it is the only CDP on the market that self-integrates into any environment and therefore needs no integrator, connecting through API, SFTP, feeds, batch and an interface contract. Scoring and segmentation come next, with twenty predictive customer scores covering affinity, value, churn, cross-sell, next purchase moment and promotion sensitivity, feeding dynamic predictive segments presented as going beyond RFM and activable in mainstream campaign tools. The fifth block is human: a dedicated team, weekly committees during the first half-year and monthly afterwards, quarterly workshops, business recommendations and quarterly summaries, an approach the publisher calls compagnonnage and which its press release names the HIT methodology, for High Innovation Targeting.
The home page advertises seven use cases with claimed multipliers: conversion x3, second purchase x4, campaign response +50%, margin x2, persona and territorial mapping, LTV and contactability +30%. These are publisher claims, published without methodology or scope. Three client references are named on the site, Aesio in health insurance, Best Western in hotels and Hippopotamus in restaurants, with Digital Group added in the press release. Distribution is direct and quote-based: no online sign-up, no public price, and a single call to action, the demo request. The site is entirely in French, and the publisher claims data hosting in France.
What it does
- Collect customer data scattered across every available source, whatever the format
- Standardise, clean and de-duplicate the records, and repair broken joins
- Reconcile identities by fuzzy matching into a single customer repository
- Rebuild the data dictionary and the data model automatically
- Restore a 360-degree customer view across every touchpoint
- Score customers on twenty predictive indicators and build segments by use case
- Push the segments back into the client's campaign tool to trigger actions
When to use Datacook / When not to
A quick filter to help you decide if Datacook is the right fit.
When to use Datacook
- Marketing and CRM departments of consumer-facing companies with large customer bases, in sectors close to the named references: health insurance, hotels, restaurants and retail.
- Organisations whose customer data is scattered across several systems and is incomplete or poorly structured, since the publisher claims no preparatory clean-up is required.
- Marketing teams with no dedicated data team, who want analysis, scoring and recommendations delivered alongside the software rather than a tool to operate alone.
- Companies already running another CDP that has not delivered the unification or the activation they expected: the publisher builds its whole argument on that situation.
- Organisations subject to the GDPR that want their customer data hosted in France and prefer a publisher established and identifiable in France.
When not to use Datacook
- Teams looking for self-service: there is no online sign-up and no free trial, the entry offer is a paid three-month package, and the only entry point is a demo request form.
- Buyers who need a public price to size a budget or shortlist vendors, since both offers are quoted only after a sales conversation.
- Engineering teams that want to assess an API or a named connector before contacting sales: no API documentation and no third-party integration is published.
- Non-French-speaking teams: the site and its documentation are entirely in French and no interface language is declared for the platform.
- Organisations wanting a hands-off product with no recurring vendor involvement, as the offer is built around weekly then monthly committees and quarterly workshops.
How to use Datacook
A typical end-to-end flow, from setup to results.
- Request a demonstration through the demo page, which is the only entry point the site offers
- Discuss scope with the publisher and obtain a quotation, since neither offer is priced publicly
- Choose between the two offers on the pricing page: the Forfait 3 mois package or the Licence SaaS
- For the Forfait 3 mois, send a simple export of transactions, campaigns, loyalty and web browsing data
- Receive the package deliverables: analysis of the business and of the customer base, profiles and behaviours, acquisition, loyalty, retention, repeat-purchase and LTV levers, value pockets, recommended marketing actions, and one test campaign with its ROI
- For the Licence SaaS, open access to the raw data and to the campaign tool or CRM through API, SFTP, feeds or batch
- Let the platform rebuild the data model, clean and reconcile the records and produce the single customer view
- Read the dashboards in the Datacook interface and work the continuously updated predictive segments
- Activate segments from the campaign tool, orchestrate across devices through connectors and export on demand
- Keep the ongoing advisory rhythm: weekly committees in the first half-year, then monthly, plus quarterly workshops and business summaries
Pros & Cons
Pros
- Self-integration is the publisher's central promise: it claims no external integrator is needed, whatever the client's environment
- Tolerance for incomplete and poorly structured data is claimed, with no preparatory work asked of the client before onboarding
- The short three-month entry offer allows a judgement on a test campaign and its ROI before committing to a licence
- Human support is part of the offer, with weekly then monthly committees, quarterly workshops and business summaries, which is unusual for a CDP
- Scoring depth: twenty predictive scores and behavioural segmentation rather than static criteria
- Data hosting in France is claimed and GDPR compliance is asserted without reservation on the product side
- The publisher is identifiable and financed: company name, trade register number, share capital, registered office and publication director are all published, three client references are named across three sectors, and a EUR 1 million round was closed in November 2023 with Bpifrance among the investors
Cons
- No price is published anywhere: both offers are quote-only, so the budget cannot be estimated before a sales conversation
- No free trial and no free tier are announced; the entry offer, the Forfait 3 mois, is a paid package
- No public API documentation and no API page in the sitemap, so integrability cannot be assessed before contact
- No third-party integration is named for Datacook: only generic channels are cited, API, SFTP, feeds and batch, so compatibility with a specific CRM or campaign tool has to be checked case by case
- Legal documentation is minimal: no separate privacy policy, no product terms, no DPA, no sub-processor list and no published retention period; the legal page rests on a legal basis predating the GDPR, states that the site collects no personal information while describing the collection of IP addresses, and requires data rights to be exercised by signed letter with a copy of an identity document
- The site and its content are in French only, no interface language is declared, and a single email address serves the whole site with no identified support channel
- The advertised multipliers, x3, x4, x2, +50% and +30%, are claims published without methodology or scope; the claim of being the only self-integrating CDP on the market is not supported by any third-party source; and the top 10 CDP ranking in which Datacook places itself first is published by Datacook
Pricing & Plans
No price is published. Datacook displays no amount anywhere on its site, and its pricing page presents what it describes as the whole catalogue: two offers. The first, Forfait 3 mois, is a paid three-month package and not a free trial. The second, Licence SaaS, is priced according to the volume of data integrated. Both are sold by quotation after a demonstration request, and the site announces neither a free plan nor a free trial.
- Forfait 3 mois (three-month package). The client provides a simple export of transactions
- campaigns
- loyalty and web browsing data. Datacook returns an analysis of the business and of the customer base
- customer profiles and behaviours
- acquisition
- loyalty
- retention
- repeat-purchase and LTV levers
- Licence SaaS
- presented on the pricing page as depending on the volume of data integrated. The client opens access to its raw data and to its campaign tool or CRM through API
- SFTP
- feeds or batch. Datacook provides unlimited use of the platform
- orchestration across all devices through connectors
- continuously updated predictive segments activable from the campaign tool
- on-demand exports
- dashboards in the Datacook interface
Data, GDPR & hosting
A consolidated view of how Datacook handles your data.
GDPR overview
GDPR compliance is claimed without reservation on the product side: the product FAQ answers "Absolutely", and the publisher's own comparison calls compliance native and non-optional. The published legal page, however, rests on a pre-GDPR basis: it cites French law 78-17 of 6 January 1978, law 2004-801, article L.226-13 of the criminal code and the European directive of 24 October 1995, and never names the GDPR. Access, rectification and objection rights must be exercised by written and signed request with a copy of an identity document; no dedicated email address or form is offered. The same page states that the site is not declared to the CNIL because it collects no personal information, while another section describes the collection of IP addresses. No DPO is named, and the publisher being established in France, no Article 27 representative applies.
Who owns the data?
Datacook publishes no contractual clause on the ownership of the customer data entrusted to the platform. Its single legal page addresses only visitors to the website: the publisher states that it collects the referring URL, the internet service provider and the IP address, and undertakes that no personal information about a site user is published without their knowledge, exchanged, transferred, assigned or sold to third parties, with one stated exception, a sale of the company, the acquirer then being bound by the same obligations. Intellectual property in the site itself is claimed by DATACOOK. Ownership of the transactional, CRM and behavioural data processed by the CDP is simply not documented and has to be settled in the contract.
Reuse rights
On the website side, the legal notice states that DATACOOK collects personal information only for the needs of certain services offered by the site, that the user enters that information and is told whether each item is mandatory, and that navigation cookies do not identify the user. On the platform side, the CDP ingests transactions, campaigns, loyalty, web browsing, server logs and behavioural, demographic, preference and campaign-response data, then standardises, cleans, de-duplicates, reconciles by fuzzy matching, enriches, models, scores and segments it, so that the client can push actionable segments back into its own campaign tools. The site publishes no reuse or licensing clause for the client, and says nothing, in either direction, about whether customer data may be used to train the publisher's own models.
Data retention & training
Hosting summary
The publisher claims hosting in France for the platform, in the block it devotes to itself in its own comparison of CDPs, where it writes that data is hosted in France with native, non-optional GDPR compliance. Separately, the legal notice declares the host of the website itself as Ex2 Inc., CP 20031 BP Pharmacie Brunet Chicoutimi, Quebec, G7H 7S2, Canada. A DNS resolution carried out on 2 September 2026 places the site's IP address, 91.134.70.153, on AS16276 OVH SAS in Dunkirk, France. These three statements are not equivalent and the site does not reconcile them: the hosting claimed for the platform, the host declared for the website and the observed IP address should all be clarified with the publisher before any contract. No trust, security or infrastructure page exists on the site, no hosting region other than France is named for Datacook, and the source of the France claim is a promotional article published by Datacook rather than a contractual commitment.
Things to keep in mind
Risks and trade-offs to weigh before adopting Datacook.
- Pricing is invisible: no budget can be built without going through a salesperson, and there is no public reference point against which to compare the two offers.
- The site states three different things around hosting: the legal notice declares Ex2 Inc., in Quebec, Canada, as the host of the website; the publisher's own comparison claims data hosted in France for the platform; and the site's IP address resolves to OVH in Dunkirk, France. Have the publisher reconcile the three in writing.
- The legal page states that the site is not declared to the CNIL because it collects no personal information, while the same section describes the collection of IP addresses, and the legal basis it cites predates the GDPR, which the page never names.
- No DPA, no sub-processor list and no retention period are published: for a tool that processes an entire customer base, these have to be demanded contractually rather than assumed.
- The site says nothing, in either direction, about whether customer data may be used to train the publisher's models: settle it explicitly in the contract.
- The performance figures displayed, x3, x4, x2, +50% and +30%, are claims published without methodology or scope, and the top 10 CDP ranking that places Datacook first is edited by Datacook itself.
- The publisher's registered name is FIDELIZ, distinct from the Datacook brand, so check which entity signs the contract; on the same note, Baptiste Pigaux is presented as CTO on the team page and as CDO in the press release.
Setup & Integrations
Technical difficulty
Presented as low. The publisher claims no integrator is needed: its AI recognises and absorbs formats on its own, and no data preparation is asked of the client, even for incomplete or poorly structured data. For the Forfait 3 mois, a single export is enough. For the Licence SaaS, the client opens API, SFTP, feed or batch access to its raw data and its campaign tool or CRM. First results are announced within weeks, against a few weeks to several months for a CDP deployment in general. No public documentation allows verification before contact.
Deployment
Behind Datacook
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Datacook.
Frequently asked questions
What is Datacook?
How does a CDP differ from a CRM?
Is an integrator required to deploy it?
How much does Datacook cost?
Is there a free trial?
Where is the data hosted?
Is the platform GDPR compliant?
Is there a public API, or a mobile app?
How long does it take to become operational?
Which clients are referenced, and how is the publisher contacted?
Should you pick Datacook?
Datacook holds a narrow and clearly stated position: a French Customer Data Platform for marketing and CRM departments, sold with contractual human support rather than as a self-service product. Its differentiating promise is self-integration without an integrator and tolerance for messy, incomplete data, the client sending what it has and the platform doing the reconciliation. Sovereignty is the second axis, with data hosting in France and GDPR compliance claimed by a publisher that is small but financed and identifiable, and that names three client references across insurance, hotels and restaurants.
Against that, the public file is thin on almost everything a buyer usually reads before a first meeting. Pricing is entirely opaque: two quote-based offers, no amount anywhere. Legal and technical documentation is scarce, with no separate privacy policy, no DPA, no sub-processor list, no retention period, no API documentation and no named integration. The performance multipliers on the home page are claims without published methodology, and the top 10 CDP ranking in which Datacook comes first is published by Datacook itself.
The decision therefore cannot be made from the website. It happens after a demonstration and a quotation, and the Forfait 3 mois exists precisely to remove doubt at limited commitment: one export in, an analysis and one test campaign with its ROI out. Before signing, ask for what the site leaves open. Which legal entity signs, since the registered name is FIDELIZ and not Datacook. Where customer data actually sits, since the legal notice declares a Canadian host for the website while the product claims hosting in France. Whether a DPA and a sub-processor list exist. What retention periods apply. And whether customer data may be used to train the publisher's models.
- Choosing a selection results in a full page refresh.
- Opens in a new window.