Snowflake Composable CDP and Marketing: From Raw Data to Effortless Hyper-Personalisation

You have a brilliant marketing campaign idea, and you write a detailed brief for the IT department to target a very specific audience. And then, the wait begins. Two weeks later, the data team sends you back a segment that doesn’t quite match your request, starting a never-ending back-and-forth process. Frustrating, isn’t it? However, this is the daily life of many marketing teams constrained by siloed and rigid data architectures.

The model of a composable Snowflake CDP promises to solve this fundamental problem by connecting marketing tools directly to the data source. Mathieu Lavedrine, Vice President of Customer & Technology at imagino, explains how this architecture radically transforms the collaboration between IT and Marketing and unlocks new use cases.

 
“Historically, data belongs to IT because marketing does not speak the language of data (SQL). The challenge today is to bridge this gap by giving real autonomy to business teams.“
 

What are the historical frictions in marketing data flows that you see among customers?

The fundamental problem lies in a technical dichotomy: marketing speaks the language of audiences and ideas, while IT speaks the language of the database. The classic briefing process between these two teams is inherently inefficient. Going back and forth to adjust a segment often takes weeks. In addition, a traditional architecture involves copying and transforming the data several times: from the source to the Data Warehouse, then to a traditional CDP, and finally to the campaign management tool. This “spaghetti dish” of data streams that go in all directions generates enormous latency and significant storage costs. Finally, this technical complexity pushes marketing teams to self-censor: faced with the technical difficulty of importing new data or creating new segments, they give up on innovation.

 
“With the traditional CDP, we have multiplied the silos and the copying of data three or four times. As a result, the campaign tool is often limited and expensive.”

How does the Snowflake composable CDP approach simplify the data infrastructure?

Traditional CDPs failed because they copied all the data to create a unified foundation, a cumbersome and unnecessary process since marketing often only needs specific data for its campaigns. The approach allowed by a composable Snowflake CDP coupled with imagino is to connect directly to existing data. This is the end of data copying. If a piece of information is in Snowflake, it is immediately queryable with the application of the imagino semantic layer.

In concrete terms, the tool translates the technical data to make it intelligible for marketing: a “0, 1, 2” code in the database will be displayed as “Sir, Madam, Other” in the imagino interface. This approach gives unprecedented autonomy to marketing (visualisation, filtering, self-service segmentation) while drastically lightening the workload of IT. Marketing manages 80% of its tasks alone, and IT can focus on the 20% of requests with very high technical added value.

 
“With Snowflake and imagino, the customer suddenly realises that adding a new field of customisation no longer requires an IT project lasting several weeks. The data is already there, it’s just visually exposed to marketing.”
 

Do you have any examples showing the benefits of enabling data with a Snowflake Composable CDP?

The speed Snowflake allows is literally a game-changer. During our demonstrations, we connect to a Snowflake test environment containing a billion lines of transactions! Marketing can make a request about this ultra-massive volume and get an answer almost instantaneously. The calculation of the RFM (Recency, Frequency, Amount) score is a great example. Previously, in retail, implementing this score required a complex IT project lasting several months with heavy pre-calculations. With a composable Snowflake CDP, all you have to do is define the RFM score as a simple rule… in less than an hour! Above all, this agility makes it possible to adapt it in real time: if inflation changes the value of the average basket, or if you want to calculate an RFM only for a specific category of products, marketing can adjust its criteria on the fly, without asking for a new technical project.

Another critical case is the integration with customer service: direct access to data makes it possible to instantly exclude a customer with an open after-sales ticket from a promotional campaign, for example. This is the ABC of a good customer experience, made possible by the absence of latency.

Beyond the technical aspect, why is this imagino-Snowflake partnership so advantageous?

For Snowflake, the tool is moving from a business intelligence solution (where the past is analysed) to a centerpiece of the operational infrastructure that generates revenue (marketing). The more data is used to personalise communications and actions, the more the investment in Snowflake pays off.

Another major advantage is in terms of security and sovereignty. imagino is the only campaign management platform that can be fully deployed in Snowpark, Snowflake’s ultra-secure application hosting environment. For sectors dealing with particularly sensitive data (banking, insurance, health, etc.), it is the absolute assurance that the data never leaves their governed ecosystem.

 
“The big paradigm shift with the composable Snowflake CDP is that the campaign tool now comes to data, and not the other way around. This instantly solves the security and governance challenges of the most demanding companies.”
 

How can Snowflake’s Artificial Intelligence capabilities be harnessed through imagino? And what advantages can we expect from it?

Historically, marketing was based on binary queries and structured data (e.g., “Spent more than €100”). With Snowflake Cortex AI, we can now deal with unstructured language. Take for example the comments left by customers. Cortex AI is able to analyse these texts, understand their nuances (sarcasm, disappointment, anger, etc.) and assign them a sentiment score. The strength of imagino is to make this AI data immediately actionable for marketing. Without having to reinvent complex new tools, a marketer can create a hyper-personalisation campaign: specifically target customers whose sentiment analyzed by Cortex AI reflects disappointment, and automatically send them an apology message with a retention offer (for example). Snowflake’s AI translates the raw data, and imagino activates it to recreate links.

 
“For years, leveraging customer feedback was an obstacle course for marketers. Today, AI does the semantic analysis at the source, and makes it possible to trigger the right action, at the right time, without any technical effort.”