Growing Data Capability in the UK

To remain competitive in a global economy, our major cities need to focus on stimulating demand and the clustering of data capability in their centres. National government can lead on creating the right conditions to support the supply of skills but traditional organisations need to fulfill the demand by including enhanced data capability in their growth strategies and by recruiting new data teams in our cities.

“The data team, whatever they do, we just let them get on with it.”

I’ve heard a variant of this statement spoken a few times recently. It denotes a vague understanding and under-appreciation of the outcomes that good data capability brings to our organisations, whether that be improving customer experience, reducing risk, improving profitability or an operating surplus.

The first challenge for any traditional organisation dragging themselves forward into modernity is reaching that pivotal moment where the pin drops and the value of data are understood.

Simply put, data capability is an enabler of all of the cool stuff that is so overhyped right now by industry and government; the stuff that private investors and governments are throwing money at in order to grow technologies of the future. Blockchain, Machine Learning, AI? Data. Improved customer experience, personalised services, reduced risk and higher returns on investment? Data.

And the people you’re going to look to in order to help you harness data in order to achieve positive outcomes? Well, they’ll be forming your new data science team of course.

Business Intelligence vs. Data Science

When trying to explain the concept of a data science team to senior leaders for the first time, start with the comparison to what it is not. A common rebuttal from senior leadership is, “we already do data science – we have a team of people who access our data and give us insights to manage our business.” Most organisations are referring to their business intelligence team.

A business intelligence team is traditionally a reporting team in your organisation. They’ll have access to your key databases, provide you with periodic reports on how your lines of business are doing; monthly, quarterly, annually. Every once in a while, they might get freed up for a few days to help with a special assignment and do something more creative.

6-ways in which a data science team is different to a Business Intelligence team

  1. A data science team is a one-hundred-per-cent, project-delivery focused team.
  2. They bring you and your customers’ active value from what they do, rather than acting as a passivereporting function.
  3. They’re constantly applying the latest techniques in computer science, statistics, mathematics and creative storytelling to your organisation.
  4. They coalesce with other data scientists around the world to bring in the latest knowledge, software, and data to enhance your organisation’s outcomes.
  5. They come with existing domain knowledge or are very keen to learn how your business works.
  6. Working in an agile way, with fewer restrictions and the ability to explore, data scientists try, learn and deploy new models or code to your practices.

How do I try this out?

If you want to know more before trying this out you could speak to organisations in your sector who are already doing data science (depending on how friendly they are) or tap into any free meet-ups in your local area. For public sector organisations, there are already a few pressing ahead with data science, most notably DWPGLAand LB of Barking and Dagenham. For the private sector, easyJet is doing some interesting things with data science.

There are a number of companies offering data science internship or fellowship schemes, which can give you a taste of these skills for a couple of months. Pivago and ASI Data Science are two based in London that comes to mind, offering their services to both private and public sector organisations.

These types of organisation will typically charge you a fee to take on a Ph.D. qualified data scientist looking to switch into the industry. They’ll work with you to define a suitable project with the fellow, provide a coaching and wrap around service and leave you with some thoughts on whether to hire 8-weeks later when the project finishes. This reference is not an endorsement. Personally, I’ve not used either’s service but there should be references available.

An alternative route may be to market and offer your own paid internship with a local university. It may be cheaper but it will be more time consuming and would require you to have a degree of existing knowledge in-house capable of recruiting and managing the internship.

Great, I’ve tried it out and I definitely want to go ahead with building a team. What steps do I take?

Broadly, there are 5-steps you should take to establish your team. Kudos to Anna Borawska from QBE, from whom I’ve taken these 5-headings. I saw Anna present recently at a Data Science Festival meet-up, another great resource to learn about data science if you’re fortunate enough to be close to London. Why reinvent the wheel, here are some steps:

1. Get Buy-in

Create a corporate ambition for data science in your organisation. This will mean building the business case for investing in something that will initially seem high-risk and speculative to others. The spill-over knowledge benefits to your organisation (improved data culture, up-skilling of non-technical staff in data, enhanced customer experience) should be quantified and presented alongside financial benefits, such as an increase in revenues and reduced costs.

For public sector organisations, I’ve previously blogged about 9 ideas of where to try data science right now.

2. Build the Team

Design the team organisation and decide where the team will initially report to in your organisation. A transformation or strategic business unit may be a good initial home for your team as it expands out, given the project-focus, although it’s very important that the new team has a good relationship and understanding with IT.

Next, decide on roles and begin recruiting. Recruiting data science capability is difficult due to short supply of skills, so you may have to compromise on building a balanced team with complementary skills and development needs, rather than focus on recruiting all-rounders.

Copyright © 2014 by Steven Geringer Raleigh, NC. Permission is granted to use, distribute or modify this image provided this copyright notice remains intact

  1. Prioritise work

Pick the best initiatives that will deliver impact. This is about getting quick and visible wins under your belt to demonstrate to your organisation that their investment will pay off. The gradual expansion of your team will be dependent on demonstrable value.

  1. Deliver projects

Establish in-house methods for delivery that fit your organisation. This may take a few turns before you get it right. Delivery should be time-bound and agile. This is about demonstrating value in the business quickly and iterating solutions based on their feedback. Continuous improvement is essential.

  1. Establish Culture

Internally Improve data literacy across your organisation. Investing in data science shouldn’t be seen as a side investment. This is an opportunity to engage with others across your organisation as you develop a new culture for doing things. Just as digital design and service design have transformed how we approach work, data science is set to do the same.

Expand your internal support base through an inclusive approach to community building — offer up-skilling, share case studies and lead the conversation in how your new team can work with other parts of the organisation.

MIT’s Data Therapy project has written an excellent blog here with more tips on how to go about building a data culture.

Externally This new data science movement is offering up new options for how we approach work and collaboration. It is enabled and accelerated by a community of openness — open data, open software, open models. Organisations should consider establishing a culture that embraces openness and knowledge sharing with complementary organisations where this provides mutual benefits.

Closing reflection: On data and the future of work

The way we approach how we deliver products and services across all sectors is at a pivotal point right now. The upcoming paradigm shift will be huge. It will be comparable to the period in the 20th century where the world switched from ‘just in case’ mass production to ‘just in time’ manufacturing.

The period of change we’re living through will see us switch from a dependence on a few knowledgeable and experienced managers who enable the delivery of excellent services over time by parting their knowledge on others, to equipping all staff and customers with the near-perfect knowledge to make unconstrained and independent decisions.

The enabler will be data, models and automated decision-making. The change agents will be data specialists.

Jon Robertson is currently the Delivery Lead for Tombolo, co-organiser of City Data Hack and leading Data and Digital projects at Future Cities Catapult. Follow Jon on Twitter or connect on LinkedIn.

Leave a Reply