{"id":4362,"date":"2019-10-23T17:14:46","date_gmt":"2019-10-23T16:14:46","guid":{"rendered":"https:\/\/www.theoaklandgroup.co.uk\/?p=4362"},"modified":"2019-10-23T17:14:46","modified_gmt":"2019-10-23T16:14:46","slug":"building-effective-data-teams","status":"publish","type":"post","link":"https:\/\/weareoakland.com\/blog\/building-effective-data-teams\/","title":{"rendered":"Building an effective data team"},"content":{"rendered":"

We recently attended the Northern Data Forum where the big issue being discussed was what it takes to build an effective data team. Dave Naylor, CTO of ResearchBods and Heather Dawe, Lead Data Scientist at Hiscox UK talked about their experiences of doing just that. It was really insightful, so we thought we\u2019d share some of key takeaways with you.<\/p>\n

We all know what bad looks like, so don\u2019t copy it<\/strong><\/p>\n

Data science is littered with examples of poorly constructed teams, often a result of having been set up too prematurely. Businesses want to do cool things with data but aren\u2019t sure exactly what, which means they don\u2019t put the tools in place to help their data teams deliver.<\/p>\n

We\u2019ve all seen it happen, so we should know what to avoid.<\/p>\n

The formula for creating a team and helping them deliver<\/strong><\/p>\n

The success of any data-driven innovation relies on the ability of the business to change. Once a data strategy is in place the way to implement it is through lots of incremental tactical changes.<\/p>\n

Embracing change is also essential to setting up a data science team.<\/p>\n

Building teams in any area of a business has always been difficult and the usual norms and conventions have existed, long before 1965 when Bruce Tuckman put forward his forming-storming-norming-performance<\/em> model for how groups form.<\/p>\n

It\u2019s no surprise that these phases are as relevant today as ever and are crucial for a team to grow, face up to challenges, tackle problems, find solutions, plan work and deliver results.<\/p>\n

But while the team dynamics are the same, the data science discipline is relatively immature, and therefore has fewer constraints and norms than other more established specialisms. For example, job titles aren\u2019t as well defined as in marketing or sales and data science has a huge breadth of talent, from the PHD superbrain to the Excel wizard. This makes it harder to set up teams as job roles can be ambiguous and the people lack customer practice. Combine this with a business unsure of the best way to integrate and make the most of their new data team, and bedding in can be tricky.<\/p>\n

The key here is patience. Going back to Tuckman\u2019s phases of group development, you can expect all the usual personalities to emerge, but it may take time. Only once you come through this period and a team dynamic forms will your pecking order naturally fall into place.<\/p>\n

Always have the big picture in mind<\/strong><\/p>\n

Before creating a data team be clear about what you want to achieve \u2013 this is great advice. Building data-driven capability requires solid foundations, which is why 80% of the work is in making the data usable.<\/p>\n

Decide where your data team fits into your business. There\u2019s no right or wrong answer here but try and avoid creating silos. Your data team is likely to interact with many functions of your business so think where\u2019s best for them to live.<\/p>\n

Then consider the optimum team size. A little like Cheers<\/em>, you want to go where everyone knows your name. Once a team exceeds 20 people the benefits of cross sharing start to break down and groups can form, so grow slowly.<\/p>\n

Building a data-driven capability<\/strong><\/p>\n

\"\" \"\" \"\" \"\" \"\"<\/em><\/p>\n

Be the best leader you can be<\/strong><\/p>\n

If you\u2019re responsible for creating and leading an effective data team, you\u2019ll need to demonstrate the skills to match. Here are some of the traits all great leaders share:<\/p>\n