Retail & Consumer · Data Foundation and Intelligence
One Data Foundation, Three Use Cases, About $7.4M a Year
A Fortune 500 automotive aftermarket retailer replaced legacy platforms costing about $2M a year in licensing, then built catalog, customer and fleet intelligence on the new foundation. About $7.4M a year, with $1M already banked.
~$7.4M/yrWhat the program is worth every year~$1M/yrAlready banked in net cost saved
14Business domains unified on one foundationThe Challenge
The Foundation. The retailer's data strategy ran on a legacy IBM data warehouse and a rigid Master Data Management (MDM) platform. Together they cost about $2M a year in licensing. Rising data volumes capped analytics on the old hardware, and blocked any move to real-time, AI-ready data. The catalog, customer and fleet signals the business already owned sat largely invisible to decision makers.
Every year on costly, inflexible infrastructure was a year of opportunity cost. Demand, customer and fleet signals that could have driven decisions sat unused instead. Moving from static reporting to real-time data products was not optional for a business that wanted to act on what it already knew.
Catalog Gap Intelligence. Every time a customer searched for a part the retailer did not stock or surface, that demand vanished unseen. Roughly 1.3 million not yet available searches a year were invisible to merchandising. Dashboards told them what had happened after the fact. They could see only what sold, never what customers wanted and could not find.
That was direct revenue leakage, and a catalog shaped by hindsight rather than live demand. Without a way to see gap searches in real time, merchandising was always reacting to yesterday's demand instead of today's. Each of those searches was demand the business had already earned and could not fill.
Customer 360. Ten or more customer journeys each ran on their own data, their own trigger and their own campaign. The same customer appeared as ten fragmented records. Plays fired independently, and sometimes against each other, with no single validated view of the customer. This ran across a digital-sales estate worth roughly $1B.
That meant wasted marketing spend, inconsistent experiences and missed reactivation. Without one validated customer context, every play optimized for itself rather than for the customer relationship as a whole. Wasted spend and mixed signals landed on the same person.
Fleet Operations. Years of Internet of Things (IoT) telematics from the delivery fleet were collected but analytically invisible. Safety, trip and asset data sat in separate silos across eight operational data domains. Operations managers saw exceptions hours late, if at all, and had no unified view across safety events, trips and asset utilization.
That meant avoidable safety exposure, inefficient routes and unmanaged cost across store, distribution center and delivery operations. Seeing an exception hours after it happened and catching it as it forms are two different businesses to run. An event worth acting on was history by the time anyone saw it.
What We Did
The Foundation. We started by migrating the legacy IBM data warehouse into Snowflake and Google Cloud, a hybrid cloud move built for cost control. The foundation unifies 14 business domains under a Data Vault 2.0 governance model, so every domain lands in one place with one set of rules.
We then tested build versus buy on governance, proposing a four-cluster Data Vault 2.0 Master Data Management design in place of the incumbent platform. Alongside it we stood up three intelligence use cases, Catalog Gap, Customer 360 and Fleet and Operations.
Catalog Gap Intelligence. We started by correlating three signals the retailer already owned: gap searches, vehicle-model lookups and off-highway demand, matched against catalog coverage and supplier data. That produced one demand-gap picture instead of three disconnected feeds.
We then contracted those signals as governed data products, each with freshness guarantees and a merchandiser approval path built in. The approval path is the part that matters. It is what makes the data safe to automate against rather than only safe to report on.
Customer 360. We started by mapping overlap, sensitivity and which signals actually predict behavior, turning ten pipelines' worth of logic into one validated customer context. The work was about deciding what the retailer genuinely knew about a customer, not copying ten existing definitions into one place.
We then contracted that context as one Customer 360 product, with masking policy, lineage and an audit trail. Those controls are what let anything act on customer data automatically. Without masking and lineage, automated retention is a governance problem rather than a capability.
Fleet Operations. We started by connecting safety events, trip patterns, vehicle assignments and asset utilization into one picture, with anomaly baselines learned from the data itself rather than set by hand. That gave operations one definition of normal to measure exceptions against.
We then contracted the picture as governed operational products, each with lineage and policy, and with anomaly baselines the products maintain from the data itself. Those baselines are what an exception engine needs before anyone will trust it: a definition of normal that comes from the fleet rather than from an opinion.
The Solution
The Foundation. A Snowflake and Google Cloud foundation now carries 14 business domains under a Data Vault 2.0 governance model. Three intelligence use cases run on it, each contracted as a governed data product and built agent-ready.
Catalog Gap Intelligence. A governed, real-time demand-gap pipeline now runs across gap searches, vehicle-model lookups and off-highway demand, built agent-ready. A demand recovery agent can nominate catalog additions and quantify the revenue at stake, with a merchandiser approving every one before anything changes. When the retailer turns that on is a business decision, and the foundation underneath it is already paid for.
Customer 360. One governed Customer 360 product now carries masking policy, lineage and an audit trail, every journey works from it, and it is built agent-ready. A retention agent can conduct across the plays, with propensity flagging risk, dormancy confirming it, and the agent selecting the journey and offer per customer. The retailer chooses when to hand that judgment over.
Fleet Operations. Safety events, trip patterns, vehicle assignments and asset utilization now sit in one governed operating picture, built agent-ready. An exception agent can watch the products live, flag coaching-worthy patterns and anomalous routes, and route recommended actions with the evidence attached. Operations decides when to move from watching to acting.
The Outcome
The Foundation. The retailer already banks about $1M a year in net cost from the warehouse migration alone, and fourteen business domains now sit on one foundation with three intelligence use cases live on top of it.
The next phase is scoped and waiting: replacing the legacy master data platform is projected to save a further $1M a year, moving toward retiring the full $2M a year the old platforms cost. That upside sits on top of the program total. Decision makers now work from signals that used to sit unread.
Catalog Gap Intelligence. The pipeline surfaced an estimated $1.46M a year in previously invisible demand. At a conservative 25% capture rate that is $365K a year, and capture depends on the retailer restocking against the signal.
Merchandising stopped working from after-the-fact dashboards. The 1.3 million not yet available searches are down 9% as gaps close, and the team now shapes the catalog against demand as it happens.
Customer 360. Ten customer pipelines were proven down to one governed asset. On a digital-sales estate worth roughly $1B, the conservative revenue-upside estimate is $5M a year from less wasted spend, more consistent experiences and reactivation that lands.
The strategic shift is bigger than the number. Every customer journey now runs from one validated view instead of ten fragments, and the plays coordinate rather than compete.
Fleet Operations. The picture is worth an estimated $1M a year in risk and cost avoidance. Eight operational data domains were surfaced, and three operations groups, store, distribution center and delivery, now work from the same live view. Telemetry that sat unread for years is now visible to the people who run the fleet, and three operations groups moved from hours-late exceptions to catching issues as they form.
Related cases
This Practice Now Earns an Estimated $6M a Year More With No Added Headcount
A tax practice managing global mobility compliance for two multinational workforces of 500-plus worked reactively from spreadsheets and email. We built an end-to-end platform spanning pre-travel risk scoring, GPS tracking and unified reporting. Billable revenue rose 20%, an estimated $6M a year across two enterprise accounts.
Partners Now Save $4.8M a Year by Reviewing Less and Advising More
A Top-10 US accounting firm reviewed partnership tax returns across 150 federal and state form variants. We rebuilt the review process around agentic artificial intelligence (AI): an engine that surfaces only what changed. Partner review time per return fell 65%, manager review time fell 80%, and the firm saves an estimated $4.8M a year.
Up to $4M in Program Risk Avoided Before Any Rebuild Began
A global professional services firm ran 200-plus enterprise client engagements on a monolithic tax-filing platform more than ten years old. We proved a cloud-native architecture on Azure first, separating the data model, calculation engine, reporting and workflow into modernizable layers. A blind rebuild decision became a costed, tested roadmap.
Paying legacy license fees for analytics you still cannot act on?
Noblq works through your process with you and applies the right technology to the way you actually work, to automate what used to need people. You get a system shaped to your business, sized to the return, and owned by you.