Blog Viewer

The E.A.R.T.H Principle: How to Run an Enterprise Analytics Team with Less … and Stay Sane

By Sharon Flynn posted 06-11-2021 10:06 AM

  

Recently, I worked for a large financial institution on the client side, leading a global enterprise-wide team that served 12 million customers and more than 46,000 employees. My team focused on digital analytics; there were other teams actively using digital data, but we were the digital data product owners.   

My team consisted of just seven people serving more than 900 direct users of the data. 2,000 dashboards and nearly 900 KPIs later, our team had an internal NPS score of 50 without doing overtime. How did we accomplish this?

While leading that team, I developed the E.A.R.T.H acronym to meet two objectives:

  1. Successfully run an enterprise analytics team with less resources
  2. Keep my sanity! 

Five Principles for Efficient and Valued Digital Analytics 

  • Empathy
  • Automate
  • Relentless
  • Transparency 
  • Harmony 

History

Following an internal reorganization, our team moved from the marketing department, where it was established six years earlier, to the larger digital channels division. This shift promptly caused us to become a global, enterprise-wide team. 

In short, our responsibility quadrupled—and we had to work even quicker than before for our clients. 

Most importantly, our resourcing (team size) did not change. With potentially the entire enterprise now in our mandate, how could we avoid a future of unending and  overwhelming ad-hoc requests? 

We are analysts … so we analyzed the problem. 

Figure Out Where You Are to Get Where You Need to Be 

We sifted through the information we had to truly understand our current status and evolving future state. We then looked at project tracking and intake logs (if you are using Jira or other project management tools, this is an excellent place to begin to understand your intake). Additionally (or worst case in lieu of intake) we looked at meeting logs, downloaded our calendars, and categorized the meeting blocks. All email exchanges were categorized for a year. Finally, we downloaded the user access logs of the self-service tools we managed and cross referenced them against the new teams.

Next, because the previous data sources were backward-looking, we also wanted to understand the new landscape by:

  • Surveying all existing users of data and new teams we identified as potential power users 
  • Creating contemporaneous notes of interactions, conversations, meetings, corridor convos, and “running into that guy in Starbucks” chatter 
  • A nice mix of quantitative and qualitative data 

Principle One: Empathy 

To fulfill the needs of your customers and move the team vision forward, you must understand the needs of your customers from their perspective. 

It is easy, especially when we are swamped, to criticize the behavior of others. But looking empathetically at the problem uncovers dysfunctional incentives and gives you the power to change behavior.

Remember, we already have:

  • Analyzed user activity 
  • Categorized the FAQ that came in via email 
  • Examined ad-hoc requests coming in
  • Surveyed our user base including verbatim
  • Collected quantitative and qualitative data to understand our customer base

We took this data and asked six very basic questions of it:

1. Who is asking questions of the data and team and who is not

This helped us identify which teams are going in and interacting with the database, and who was conspicuously absent. 

2. What are they asking? What are they doing with the data?

Was it clear what the request was for? Can we elevate the data request to be more efficient and effective?

3. When are they asking? 

It might seem obvious, but is there a predictable cadence here to help us plan for spikes?

4. Where are they in the organization chart? Product? Marketing? Design? 

What about their position could inform us of their data needs, and why? Again, is there a published cadence, product roadmap, or something else we could plug into? 

5. Why are they asking in that way? 

Are they asking for the same “quick number” over and over? Could that be automated? Are they overcomplicating the ask? It’s important to try to understand where they’re coming from when they ask questions.

6. How valuable is this deliverable to the overall organization’s objectives?

As subject matter experts, we needed to focus our hours on high-value work, whereas “nice-to-haves” had to be self-serve. 

Principal Two: Automate 

Find all “copying and pasting” activity and stop it. Revisit current tools and max out their automation capabilities. 

QA via tools like Tag Inspector can free up thousands of hours of analysts’ time. 

Even Microsoft Excel can be more efficient. Ensure your team understands every aspect of tools like VLOOKUP, Power BI, and API plugins—and then use them all.

“Copy & paste” is your business case for a new solution. Reach out to the community via DAA, Slack, Twitter, or LinkedIn and reach out to your internal community of similar teams. Your problem is rarely unique; there is a solution, so do not settle.

Principal Three: Relentless 

Be ruthless in critiquing how the team works. For example, I have a philosophy that a regularly produced report that creates ad-hocs is a failure.

The Good Ad-hoc
  • Build and evolve your reporting 
  • Facilitate knowledge building 
  • Align with your expertise either in complexity or time
  • Align clearly with a decision and outcome 

Good ad-hocs act as bespoke analysis or needed revisions of a living process. 

The Bad Ad-hoc
  • Variations on data in your report that are “special snowflakes”
  • No clear alignment with overall business decisions and outcomes 
  • Should be self-serve 

Bad ad-hocs need to direct requester to self serve or refine request to be a “good ad-hoc”. 

The Ugly Ad-hoc

  • Might require some escalation or a tough conversation 
  • Unrealistic delivery timeline
  • Use relationships to jump the queue 
  • Obscure reasoning 

As an aside I dislike "Ad-hoc" and prefer to use bespoke. 

Principal Four: Transparency  

I am a huge fan of public-facing tools outlining your workflow. Project management tools like Jira and Asana allow stakeholders to self-serve the answers to the question: are you busy or just mismanaging your time?

Transparency prevents a dysfunctional appearance and facilitates an unemotional, priority-focused allocation of resources. It is also useful for building business cases for additional staffing or tools to boost capacity that is measurable and seen as impactful to the business. 

Principal Five: Harmony

Once you have an understanding of your users, have empowered them, and democratized your intake process, you then can work to make the team as balanced as possible.

A balanced team is critical, so it’s important to hire the best of the best. Here are some qualities all teams should value:

  • Share the same sheet music – Shared, clear goals allow a lean team to have direction, autonomy, ownership, and purpose in their working lives 
  • Cast for passion – Understanding where each team member aspires to go in their career allows you to align projects to those goals ensuring mutual benefits and high retention 
  • Teams should be safe, supportive, and respectful
  • Shared vision for future
  • Remove obstacles – As a leader, it’s your job to remove obstacles to success, from slow computers to combative stakeholders 
  • Ban invisible work – absolutely ZERO overtime. This only works with transparency of resource allocation; however, consistent overtime is a red flag as a leader and needs to be stopped as soon as possible. 

#CareerAdvice

Permalink

Most Recent Blogs

  • Posted in: Women in Analytics

    I am very proud to serve the DAA’s Women in Analytics community and greatly admire the contributions the members of this ...

    1 person likes this.
  • Posted in: Women in Analytics

    Recently, I worked for a large financial institution on the client side, leading a global enterprise-wide team that served ...

  • Posted in: Women in Analytics

    Volunteering with the DAA’s Women in Analytics mentoring program has been an eye-opening experience for me. The DAA states ...

    1 person likes this.
  • Posted in: Women in Analytics

    I am delighted to bring you a guest blog from Tammy Kim-Newman is Business Developer and Tech Talent Specialist at the University ...