- About BJSS
- Our Services
- Your Industry
- Enterprise Agile®
By Simon Hull, Head of Financial Services at BJSS
Last week, a few BJSS colleagues and I participated in the CDAO (Chief Data & Analytics Officers) three-day conference in London. The event was well attended and featured data and analytics leaders across several industries, with a significant presence from financial services, including NatWest, HSBC, UBS, BNP Paribas, Mastercard, Hiscox, AIG, MUFG, and the Bank of England.
There were many great keynotes, panel sessions and group discussions over the three days, and the opportunity to meet and talk face-to-face was welcome after a few years of virtual conferences.
As we heard about the success stories and challenges of data and analytics transformation, five key themes emerged for me as the critical areas of focus for organisations who wish to harness the power of data and analytics at an enterprise scale to drive positive business outcomes.
I've outlined the themes and the main takeaways below.
cultural change is the most important and difficult
It was almost universally acknowledged that people and cultural change is the most vital and challenging aspect of any enterprise-wide data and analytics transformation programme.
The promise to break down silos and democratise data and analytics will only be successful if accompanied by an organisational focus to improve data literacy, collaboration, and experimentation levels.
Within banking, for example, the idea of creating data literacy from board to branch was discussed. Executives need to get hands-on with data to understand and take actions based on the insights they receive, branch staff responsible for the primary capture of data need to understand the importance of data quality, and everyone in the company should become familiar and comfortable with data and analytics concepts.
Closer collaboration around data means moving on from a siloed world of data fiefdoms to one where business units freely share data and collaborate based on the same shared view of the customer. As data science experiments occur across the firm, the use cases, results and outcomes should also be shared to promote organisational learning and reduce duplicative effort.
Finally, innovation should be encouraged to ensure that experiments happen across the company, conducted by citizen data scientists operating within certain organisational guardrails that are defined centrally. Success should be promoted and celebrated to build momentum.
know your customer
When it comes to better understanding the customer, three areas came out as important - building the customer 360 view, how this can be applied at a real-time transaction level and how this can be implemented at a customer lifetime level.
Building the customer 360 view is a challenge for organisations that have evolved in business and technology silos, as there are generally many different customer identity systems in place. These identity systems must first be aligned to create one unique identity across the entire business. This will enable the various siloed data sources to be brought together and enriched with other external sources to provide a single view of everything you know about a customer.
Once this model is available, it can be used to improve customer service and experience, and two different horizons of interest were discussed.
The first is how to improve the customer experience in real-time for the particular transaction or activity they are in the middle of. This is all about enabling a joined-up journey across different channels or business areas, which is enabled by having shared customer-centric events that can be used to create seamless and contextual journeys.
The second is the bigger picture of a customer's lifetime journey. This is about using everything we know about a customer to understand what they are trying to achieve and why, and using this to predict their needs and identify life stages and milestones. For example, a change in savings habits may indicate that a customer is saving for a house and provide an opportunity for proactive assistance.
helping not selling
Being able to uniquely identify a customer, understand their real-time digital journeys, and start to predict future needs is one thing, but the way in which this information is used is key to the success or failure of the relationship.
It was mentioned across several talks that the priority was to focus on being helpful and providing education rather than on selling. Customers are generally put off by overt sales tactics inserted into their digital journeys, but if insights are based on predicted future needs, the customer is not ready to buy now anyway. So, in both cases, a focus on help and education is recommended.
There were many examples of this. For instance, an analysis of spending patterns in banking can enable banks to make suggestions to help customers take advantage of discounts or avoid duplicate spending.
Thinking further ahead, if it can be identified that a customer is on the path to a particular lifetime milestone (such as buying their first house) then a focus on educational content about the house buying process should be useful, well received and add value to the customer relationship.
privacy and trust
Privacy and trust, as expected, were also themes that were discussed extensively, with almost a third focusing in some way on data governance. As the use of data and analytics grows, so will the importance of being a trusted custodian of customer data. This is driven by both expanding regulations after GDPR and CCPA, increasing customer demands and the imperative to be a trusted organisation.
In the context of the previous theme around the focus on helping the customer comes a warning. Organisations must ensure that their interactions or messages are not seen as an invasion of privacy and interpreted by customers as a misuse of their data.
Steering away from sales and marketing messages in favour of contextual help and support should help with this. Also related was a discussion on the importance of good zero and first-party data collection rather than relying on third-party data.
Trust was identified as a critical competitive differentiator of the future. People want to do business with firms they trust and are even willing to pay a premium. Analysis shows that the top trusted brands perform better from a CAGR perspective.
Many organisations discussed trust, privacy, and the ethical use of data as a top-of-house priority amongst data ethics committees, including c-suite and board-level membership.
dataops and mlops
Finally, there was much discussion about the role of automation in scaling AI across the enterprise, particularly in the use of multi-disciplinary teams and the adoption of DataOps and MLOps practices to foster collaboration and improve the speed and quality of model development.
It was highlighted that DataOps, defined as "the application of agile, lean and DevOps to deliver outcomes for data", was essential in bringing producers and consumers of data closer together. Introducing automation into data delivery increases predictability and enables scale.
Data as code, with self-describing schemas and metadata enabling automated discovery, classification and lineage, is also essential. Doing this manually at an enterprise scale is a huge effort with a short shelf life. Much like cloud engineering, data engineering is getting closer and closer to software engineering, and some of those fundamental principles, approaches and practices are critical to delivering data effectively.
Building on DataOps, MLOps practices ensure an alignment with business needs and introduce automation into the overall model lifecycle to increase the pace of model development and continuous improvement. The lack of MLOps was highlighted as a key reason for the failure of AI to scale.
The above five themes are very much aligned with BJSS' approach to data & analytics. We support clients with data strategy and governance, data platform delivery, data science solutions, and ongoing data managed service support and maintenance. We take a user-centric approach, build security and privacy by design and utilise DataOps and DevOps to deliver rapid data-driven innovation.