I made a jump in my career from computer-aided drug design to the burgeoning business intelligence field around the time of the dotcom bubble. Learning the ropes in several boutique consultancies - and getting exposed to a wide variety of industries and analytic use-cases - was invaluable experience in those early days.
I developed my interests in innovation, applied modern analytics and architecture during my long service with a large financial services organisation in London. The company championed internal mobility, so I applied my knowledge of data warehousing, visualisation and self-service analytics to lead cross-functional, global teams in the delivery of analytics in many of the company’s key business areas: risk, global banking and markets, finance, retail banking and more.
At Alteryx, I’ve been extremely fortunate to develop my analytics career in several directions: first, by driving customer engagement with our platform, but also working across the product division to help shape the future of analytics technology through continued investment in innovative and open-source projects, as well as building relationships with partners and the wider analytics ecosystem.
Perhaps, most enjoyable of all, I get to evangelise the benefits of data science and advanced analytics, helping to develop the untapped potential of analysts within organisations.
I was extraordinarily fortunate to attend a year-long leadership course based at Sandhurst Military Academy during my time in financial services. Sandhurst’s motto of “Serve to Lead” has remained with me ever since, and I firmly believe in a real distinction between analytic leadership and analytic management in our industry.
To put this into practice, I donate my analytic leadership as a data ambassador to DataKind UK – a data science charity that supports analytic exploration and solutions for charities and social enterprises across the UK. It operates by co-ordinating groups of data scientists to work on specific problems over a period of weeks or months. In this respect, data science is most definitely a team sport.
I’ve found nothing so rewarding as “serving to lead” these groups of talented individuals over numerous data dives. The camaraderie, the talent and the energy in the room would be highlights in anyone’s career.
In such a technical industry, it makes such an impact when you can explain concepts, approaches and best practice without over-complicating or overwhelming an audience. To this end, my role models are Peter Jackson and Caroline Carruthers who’ve written excellent, accessible and approachable books about the expanding world of the chief data officer and data-driven Transformation – truly inspirational.
What surprised me most was how loudly the analytics industry become enamored with automation to build on previous investments in data engineering (to store and analyse large volumes of data at a reasonable cost of time and money) and the groundswell of self-service data analysts and citizen data scientists.
Collectively, we have shifted our capabilities (and expectations) to conceive a data-driven experiment and then being able to unleash collective creativity to develop insights and deploy analytical models into a governed, maintained environment in a fraction of the time of previous years.
Now, automation allows that model to be accessed, refreshed or deployed without needing armies of IT folks sitting behind every analyst or data team – this is analytic acceleration.
In the year ahead we should expect to see sustained growth in the use of advanced analytics (especially predictive and prescriptive analytics) at the core of the most successful digital transformation programmes.
Organisations will realise that the highest monetisation of their transformation efforts comes from predictive models driving insights and prescriptive models automatically driving business actions.
Whether it’s financial forecasting, supply chain management, pricing optimisation or next best offer analytics, these are the high-value, high-impact use-cases that can pay the bills for entire analytics departments.
The biggest opportunity may yet be the biggest threat that we collectively face as a global society – the application of advanced analytics and AI to hyper-personalised data can result in micro-targeting not only for retail brands but also for political actors to manipulate events such as elections and the global democratic machinery.
The opportunity arises for a wider conversation around AI ethics across our industry. Organisations like DataKind report that the vast majority of analysts and data scientists haven’t signed up for a code of ethics, and often don’t have formal systems in place to ensure that experiments are properly scrutinised for bias or negative outcomes.
This leads to a deeper conversation about trust across the data and analytics ecosystem. Do we truly understand data source lineage and quality? Can we automate our data pipelines to reduce manual errors? Is our final model explainable and interpretable? Capturing these attributes brings us closer to a complete picture of our information flow from data to insight to action.
It’s so critical to remind clients that technology is really not at the core of a business transformation strategy (after all, technology doesn’t explicitly deliver the value) but instead it’s fundamentally about how data assets are curated across an organisation, and then put to use to deliver actionable insights.
When I discuss digital transformation with clients and prospects, there’s often an “analysis paralysis” associated to finding the right place to start the programme.
While it’s tempting to tackle some of the biggest challenges head-on (we often call these disruptive transformations), there are significant benefits from starting out with innovative transformations, taking a single business process, identifying manual or other choke points, and making an analytic process intervention.
In this way, we can take existing processes that run for hours and transform them into analytic steps taking literally seconds. The payback (in computation, reconciliation and human resources) is highly visible, immediate, and often addictive. Delivering innovation this way often fits well into agile transformation project cycles too, with multiple changes being released on a regular basis.
After several cycles, teams build up serious analytic “muscle memory” which (along with treating data as a tangible asset to the organisation) means that teams are ready to tackle those more disruptive transformation challenges along with trusted partners.
Clients are now demanding more and more standardisation to these higher-value projects: there’s a growing body of knowledge and industry domain expertise around best approaches to analytics in retail, healthcare, supply chain (and others).
The technology challenge now centres on how to deliver vertical analytic solutions that deliver best-in-class performance, while being flexible to localised differences in environment (on-premises, cloud, hybrid, etc) and being approachable to the widest range of analytic users.