Data Leadership Challenges I Failed At So You Don't Have To! PART I

Author

Tristan Burns
January 24, 2024

I’ve worked in data for about 14 Years…

…and I’ve been in data leadership (in some form or another) for almost half of that time.

When I started out as a data leader I didn’t have a clue what I was doing.

I met many challenges in that time. Some of them I conquered.

Others conquered me.

This post series is about those. The failures.

Below is Part 1 of Data Leadership Challenges I Failed At So You Don’t Have To

Below I discuss the first 2 (out of 6) challenges I faced and failed at and the resulting impacts.

Luckily today, with hindsight, time and inspiration from others, I have a better idea about how I would approach these challenge now and wanted to share that with you in the hope that my experience helps you in the face of similar challenges.

1. Trust in data being 100% accurate.

🏔 Challenge: There was an assumption from stakeholders that data had to be perfect in order to use it in the decision making process. I exacerbated this by often saying I wasn’t confident enough in the data for me to have an opinion on certain decisions.

📉 Impact: It resulted in my spending too much time on data quality and neglecting supporting the business through insights which has both operational and reputational downside for the data team.

Companies need to move fast and make data informed decisions. Data perfectionism kills this and limits businesses ability to innovate. It also creates distrust in the data.

👨🏼‍🎓What I learned: Data quality is important but data perfection comes at the expense of speed and innovation. It is also harms the goal of data-centric decision making.

As with any decision, we may need to move forward without a full picture if we wish to progress. Whilst we must strive for high quality data, it can’t come at the expense of progress.

🚀 Today’s Approach: Perfect is the enemy of good and a fools errand.

Get to at least 51% confidence with the insights that you have and commit. Over time your data quality an intuition will improve. Moving fast with some but not all the info is better than not moving at all.

(N.B. I’m not talking about a confidence level in a statistical sense)

2. Shadow data teams popping up

🏔 Challenge: Due to the small size of my data team it wasn’t always easy to stay across the data needs of the entire business. As a result “shadow data teams” popped up.

This is where another business domain hires their own data analyst who’s role is typically to build reports and answer basic questions.

I often had no visibility of these hires taking place until the person was in role and already working.

📉 Impact: Shadow data teams are dangerous because there is often no alignment between them and how data is handled elsewhere in the business. It is also possible that the individuals hired don’t have the right experience or skills to succeed.

This creates a tonne of friction when multiple reports and varied interpretations of the data start to emerge. This causes confusion and mistrust in both teams and more broadly.

You may not even know shadow data teams exist in larger orgs.

👨🏼‍🎓What I learned: I failed here because I fought against this trend.

When your team is too small to support the entire business it may be inevitable that shadow teams emerge. When other teams have surplus budget and yours doesn’t it should come as no surprise that this happens.

Your first instinct will be to refuse to cooperate with shadow teams and take this as a personal insult.

But this will just hurt you and your teams standing in the business in the long run. Cooperation and transparency is the key here.

🚀 Today’s Approach: Faced with this challenge today, I’d first of all be completely honest with the limitations of my small team and acknowledge that we are struggling.

I would request to be looped in on any plans to hire data people and try to involve myself in their hiring, even if I’m not going to 'people manage’ them.

I would strive to install a dotted line between myself and them on the org chart and insist they participate and cooperate closely with my team in order to create a hub and spoke model.

Check back in next week for part 2 were I look at how I failed in regards to:

  • Failure to educate stakeholders on simple data concepts, &

  • How I failed to handle data tool adoption.

⚡️Whenever you are ready here are a few ways I can help you: