- Ekow Duker
Getting Real Value From Your Data Scientists
With many organisations looking to enjoy the benefits of digitisation and a data-led approach, finding and recruiting data scientists has been the hot topic of 2019. There is a plethora of online and physical courses that teach the technical skills a data scientist needs. But learning data science is one thing. Being effective in a workplace is another thing all together.
With this in mind, we partnered last year with a major Rwandan bank and insight2impact (i2i), a global resource centre that seeks to improve financial inclusion through the smarter use of data. Ixio’s role in this unique three way collaboration was to support and mentor the bank’s newly hired data scientist interns over the period of a year. We also acted as an experienced “bridge” between the data science team and the bank’s executives.
We were mindful that both the bank and the data scientists were “newbies” when it came to data science. The bank had only just embarked on its data journey and had never hired data scientists before. The data scientists for their part, were fresh out of college, and had little to no work experience. There was significant potential here for the bank and their data scientists to talk past each other and all of us worked hard to ensure this didn't happen. We’ve distilled below our key learnings from the engagement and the attributes that we believe contributed to a successful set of outcomes.

1 - Communicate a lot: We made extensive use of Slack, an “always-on” communication tool, as a means of getting all parties, many of whom were never in the same physical location, to share ideas, questions and get feedback. This ensured that roadblocks were removed quickly and the data scientists felt supported throughout their journey.
2 - Choose the first projects carefully: We deliberately spent a lot of time at the beginning of the project, deciding with the bank and i2i, what the objectives of the engagement were to be. It was important that by the end of the project, the bank’s data science team would be able to point to tangible outcomes that were also impactful for the bank. In the end the three objectives we settled on were:
Churn analyses to determine clients at risk of closing accounts
Sentiment analyses of social media posts to assess status of products and services of the bank in relation to competitors
A custom-built interactive dashboard for customer insight
3 - Have a committed executive sponsor: The data scientists worked closely with the head of marketing who was a vocal supporter of the project. His enthusiasm (and his seniority) were instrumental in keeping the project on track.
4 - Help your Data scientists grow: We were careful not to do the work for the data scientists, but rather pointed them in the right directions and explained why certain analytical approaches might be more useful than others. In doing this they grew visibly in confidence as they flexed their analytical muscles on increasingly complex tasks.
5 - Presentation and visualisations skills are critical: We spent several weeks iterating over final presentations and reports with the data scientists, helping them move from an arcane narrative filled with data science jargon, to a simple story that explained what they had done and why it was important to the bank.
The project culminated in the data scientist interns presenting to the CEO of the bank and members of the executive team at a breakfast meeting in Kigali. The response was overwhelmingly positive, especially when the data scientists revealed important, actionable insights, that had previously lain hidden within the bank’s data. The real measure of the success of this project is that, the interns were offered permanent contracts on the spot. This confirmed that experienced support is invaluable in making freshly trained data scientists more effective and employable. This was a welcome and well deserved validation of the work they had done.
None of this would have been possible without i2i’s Technical Assistance Lite programme. The Technical Assistance Lite programme supports the placement within a financial service provider (FSP), of data science interns alongside an experienced data science consultancy, for a predetermined period. The collaboration is primarily responsible for championing internal data science initiatives — essentially using both internal and alternative data sources, to draw actionable insights that improve financial inclusion and drive business profitability.
We are grateful to i2i and to the bank for their foresight in crafting an effective means of building data science capability within the organisation.