Most large organizations call themselves data-driven, yet their teams still spend most of their time cleaning and locating data instead of building models, dashboards, or AI products.
Singapore Summit: Meet the largest APAC brokers you know (and those you still don't!).
IG Group has tried to solve that problem with a new “extended medallion architecture” that impressed Google enough to become the subject of a co-authored white paper and a separate customer success story.
- Tesla Options Go Daily in Market First from IG Amid Campaign to Raise UK Retail Investment
- With Pat Cash Onboard, IG Addresses Falling UK IPOs and Retail Engagement
- IG Ends PayPal Payments for UK Users Following FCA Discussions
The core issue is familiar. Traditional data platforms often follow a medallion pattern, with bronze for raw data, silver for cleaned and standardized data, and gold for business-ready outputs.
For FX and retail trading , this is mainly about speed and consistency. IG’s new data model should help its FX, CFD and crypto teams get cleaner numbers faster, so they can launch features, reports and risk tools more quickly and react faster when markets move.
Data Teams keep Control while Business Teams Move Faster
For clients, that work should show up as more reliable information across the app, web and statements, and a quicker rollout of new tools and AI-driven features. In simple terms, IG aims to fix the data pipes so traders and investors get clearer, more dependable information and improvements sooner.
In many firms, a single central data engineering team controls all three layers. That protects quality but creates long queues as every domain relies on the same group for new datasets, metrics, and changes.
Keep reading: IG Japan Confirms Potential Data Exposure of 163K Clients, but No ‘External Leak’
IG Group’s extended approach keeps central control over the bronze and silver layers and over any shared gold datasets. It then adds separate “domain gold” projects for teams such as risk, finance, or marketing .
These teams draw from the governed silver data and build their own aggregations and features without changing the core model or overloading the central team.
The design is platform agnostic but already runs live on Google Cloud, using services such as BigQuery, Cloud Storage, and dbt. A senior Google field engineer stress-tested the architecture against real-world scenarios before Google agreed to publish the detailed white paper and case study.
Cutting Heavy Data Preparation Burden
For IG Group, the model aims to cut the heavy data preparation burden, sharpen governance, and let business teams ship products faster. For the wider market, it offers a practical template for balancing control and agility as firms move deeper into AI-led workloads.
Several large financial institutions are also rebuilding their data platforms to support AI and real‑time analytics, but most do so quietly and in partnership with cloud or data vendors, without sharing full blueprints. They talk about “governed data platforms” and “AI-ready data estates”, yet the underlying designs usually remain internal.
In that context, IG’s move stands out because Google has stress-tested its in‑house pattern and then published it as a named reference architecture with both a white paper and a customer success story. For a retail and FX broker of IG’s size, having its own data design treated as a reusable model by a major cloud provider is still the exception rather than the rule.
Ironically, as firms double down on data-driven strategies, IG Securities (IG Japan) has disclosed a data handling lapse that potentially exposed the personal information of up to 162,879 clients internally, alongside 29,734 records stored on an external server without prior approval.
The firm said there is no evidence of any external breach, attributing the incident to contractor oversight failures, weak access controls, and a misclassification of sensitive My Number data.
