On-Demand Webinar · 1 hr 2 min

Differentiation Through DataOps in Financial Services

Simon Trewin, co-founder of Kinaesis, joins Chris Bergh to work through how financial institutions move fast without breaking things: virtual environments and continuous deployment for delivery speed, automated testing across end-to-end pipelines for quality, and what regulation demands of both. Recorded February 2021; updated August 2026.

Presented by Chris Bergh

What you'll learn 6 points
  • Kinaesis builds its financial services DataOps practice on six pillars: a Target driven by a business vision and mapped as user journeys, Instrumentation of the data flow at every step through profiling, data quality and monitoring, Metadata that keeps business definitions and models connected to the data, an extensible Platform that can absorb new business demands, Collaborative analytics across consumers, IT and data owners, and Control through version management, release management and exception handling.
  • Kinaesis defines the target with a five-part process it calls S.C.O.P.E.: Storyboards of the user journey, Content represented with the correct context, Output written down and agreed across stakeholders, Process for interacting with the data pipeline, and Estimate as a cost-benefit analysis of any pipeline change.
  • Data projects in large financial organisations do not naturally iterate or move fast: infrastructure takes a long time to procure, sources are complex and have competing priorities, controls and governance add layers, small data gets big quickly, pipelines are complex, and the methodologies in use were borrowed from software engineering.
  • A large banking group running a BCBS 239 programme was behind schedule with eight months left before non-compliance. Kinaesis added a focused team of five or six consultants, mixing data specialists, risk subject matter experts, analysts and technical specialists, and broke the work into trackable iterations.
  • That engagement produced 300 metrics across seven lines of business, from one month of upfront analysis followed by six months of delivering metrics incrementally. It identified 207 reconciliation breaks, fixed billions of pounds of reporting errors, produced 2,000 rows of metadata and a three-month evidence trail, and left the bank with a new operating model.
  • DataKitchen frames the financial services work as four areas improved iteratively: decreasing the cycle time of change, lowering error rates in production so customers trust the data, improving collaboration within and between teams, and measuring the process. A top 5 US bank applied it to self-service, building governed data sandboxes for more than 1,000 non-IT users with legal rules on data usage and lifetime, monitoring of usage, and a path where an idea earns the right to be reimplemented centrally.

Slides

45 slides

Questions from this session

What are the six pillars of the Kinaesis DataOps approach?

Target, Instrument, Metadata, Platform, Collaborative Analytics, and Control. Target means the pipeline is driven by a business vision mapped as user journeys. Instrument means profiling, data quality and monitoring at every step, with quality information always presented alongside the actual data. Metadata keeps business definitions connected to the data, the platform is extensible for new demands, analytics is collaborative across IT and data owners, and Control adds version control, release management and exception handling.

What is the S.C.O.P.E. definition process?

S.C.O.P.E. is how Kinaesis defines the target of a data project so stakeholders are not disappointed later. Storyboards map user journeys and visions, Content represents facts, reference data, transactions and process metadata with the right context, Output writes down and agrees the reports, analysis and data to be delivered, Process describes how people interact with the pipeline, and Estimate is a cost-benefit analysis of the change.

Why is DataOps needed in financial services?

Regulatory controls have added layers of complex data processes and the regulators keep getting more sophisticated, faster transactions have made the market intensely competitive, big technology firms are moving on incumbent income streams, and large financial organisations carry a lot of legacy. Meanwhile there is an expectation that machine learning will improve decisions. Data projects in these organisations do not naturally iterate or simplify, which is the gap DataOps closes.

How did DataOps help a bank meet BCBS 239?

A large banking group's BCBS 239 programme had grown too large and was behind schedule with eight months to go before non-compliance. A team of five or six consultants broke the challenge into trackable iterations, delivering key risk indicators incrementally rather than in one release. The result was 300 metrics across seven lines of business, 207 reconciliation breaks identified, billions of pounds of reporting errors fixed, and a compliant operating model.

How does a bank give business teams data without losing control of it?

Through governed self-service sandboxes. A central IT or data group gives a business team a prepared data analytic environment with the data sets and tools it asked for, monitors and governs its use, then takes it back or changes it. A top 5 US bank ran this for more than 1,000 non-IT users, following legal rules on data usage and lifetime and tracking usage throughout.

What is DataOps?

DataOps is the set of technical practices, cultural norms and architecture that enable rapid cycles of experimentation and innovation in delivering new insight, low error rates, collaboration across complex sets of people, technology and environments, and clear measurement and monitoring of results. It covers both the technical environment and the people, process and organization around it.

Where to go next