All the data, in one platform. The answers still missing.
ASICS collects data from every omnichannel touchpoint, plus the systems around it, into a global platform on Snowflake. Customer journeys, conversion, product launches, the loyalty programme. The platform holds it.
Their product teams still had to turn that into insights. Which step in the journey loses people? Which launch worked, and why? Where does the loyalty program pay back? Answering that takes someone who knows the data model and sits close enough to the product to know which question is worth asking.
The data was never the problem. Getting from the data to a prioritised roadmap was.
Two engineers in the teams, dbt in the lakehouse, a semantic layer on top
The order matters. First get answers to the teams that need them, then make the platform easier to work on, then open it up to everyone else.
Embed in the product teams
Two analytics engineers, full time, with the teams that own the customer journey and the loyalty programme.
Read the data together
Conversion ratios, drop-off in the journey, launch performance. Cross-referencing sets that had not been read against each other before.
Hand findings to the roadmap
Findings go back to the product teams as options they can act on. They decide what gets built.
Add dbt to the lakehouse
Proof of concept and launch on Snowflake within eight weeks of the decision.
Get dbt adopted
One team using it is a pilot. We are helping the analytics organisation put it to work across their own models.
The semantic layer
On top of the lakehouse, so AI agents, analytics engineers and product teams can ask a question in human language and get a defined answer back.
QA every answer
We compare what the semantic layer returns against the factual data. An answer nobody trusts is worse than no answer at all. Automate to scale.
Four layers between raw touchpoints and a plain question
Snowflake lakehouse
The global data platform ASICS already runs. Omnichannel touchpoints and the surrounding systems land here.
dbt transformations
Modelling, testing and documentation in one place, so a transformation can be reviewed and reused instead of rewritten.
Semantic layer
Business terms defined once. AI agents, analytics engineers and product teams query it in human language.
QA against the source
Every semantic answer is checked against the factual dataset behind it. Trust is the whole point of the layer.
The same semantic layer is connected to the Agentic Runway, so the agents running there work from the same defined business terms as the analysts.
Capacity, a working platform, and a layer anyone can query
Analytics engineering, embedded
Two engineers in the product teams, producing findings the teams act on rather than reports they file.
dbt on Snowflake, live and spreading
Eight weeks from decision to production, plus the adoption work to get the wider analytics organisation onto it.
Semantic layer, architected and in build
The design, the implementation work, and the QA loop that keeps its answers matching the source data.
Insights, outside-in view and implementation experience
We are one of two AWS Retail Competency partners in the Netherlands, and Lambda Service Delivery certified.
Embedded analytics engineering
Senior engineers who work in your product teams, on your platform, in your rhythm.
Semantic layer architecture
The design pattern behind our agent work: defined business terms, one layer, answers you can check.
More of our work
All cases →
Sitting on data without insights?
We put analytics engineers inside your product teams, and build the layer that lets everyone else ask their own questions.