Increase data team efficiency with a modern data stack

For high-growth companies, building a focused, priority-driven analytics team is mission-critical.
December 11, 2020

Building a data team at a high-growth company requires strict focus, a deep understanding of priorities, and establishing a foundation of trust, says Jacob Bedard, Director of Data Analytics at Dialpad.

When Bedard joined Dialpad, a cloud-based business phone company, he carried out a current state assessment, looking at the company’s areas of opportunity and areas for improvement in data insights and analytics.

He grouped his to-do list into four categories, generally following psychologist Abraham Maslow’s famous hierarchy of needs:

  1. Identify the most critical areas for business data needs
  2. Generate factors and insights that would further improve decisions
  3. Suggest ways to increase efficiency
  4. Conduct experiments that can add value

It didn’t take long before the business realized that the analytics team needed to be embedded in the foundation of all business decisions. Also, as requirements from the start, the team needed to produce accurate metrics consistently and ensure stakeholders felt confident in basic reporting.

How a modern data stack helps teams meet goals

A large part of building stakeholder trust in the data is laying the right foundation. In a presentation during the Modern Data Stack Conference 2020 hosted by Fivetran, Bedard gave his best practical advice for setting up a modern data shop — and keeping it modern.

First, let’s define the components of a modern data stack, which consists of (for a more detailed breakdown of the modern data stack, see our detailed overview).

  • An automated data pipeline tool such as Fivetran
  • A cloud data destination such as Snowflake, BigQuery or AWS Redshift
  • A post-load transformation tool such as dbt (also known as data build tool, by Fishtown Analytics)
  • A business intelligence engine such as Looker, Chartio or Tableau

With the foundation for analysis set and stakeholders confident in the data, the business felt comfortable moving towards data science and machine learning to take the business to the next level. Below are a few lessons Bedard learned along the way.

Focus on earning and keeping trust

During the assessment, Bedard learned his number-one priority needed to be gaining and keeping stakeholders’ trust. “We can’t stop right in the middle of an important business conversation and ask, ‘Are we sure we trust the data?’ or ‘How did this get put together?’ or ‘I think I saw something different last month,’” Bedard said.

Earn the company’s confidence by consistently demonstrating high data integrity and promptly responding to issues that arise, especially when it comes to accuracy or availability.

Prioritize the ‘nightmare scenarios’

Make a list of the worst possible situations your team may find itself in and determine how you can either prevent or swiftly react to them. Bedard thinks about “the things that would wake me up in a cold sweat in the middle of the night if I had a nightmare about them. And then I ask, am I actually ready for that, and will it go smoothly? Because these things, unfortunately, do happen.”

Whether your nightmare is miss-stated revenue figures, broken data pipelines, or reporting on obsolete data, plan now for when the worst happens.

Be nice to your future self

Consider how scalable your tools and processes are. Your systems should be built for future success. Metrics, input data, and processes will change over time, so make sure your data infrastructure is flexible enough to evolve, too.

This is where a modern data stack is crucial, Bedard said, since you can find the best tool for each specific job. For example, Dialpad uses Google BigQuery as its enterprise data warehouse and central data store, with Fivetran synching CRM data from Salesforce (among other data sources) into it.

Embrace change

Lastly, get comfortable with change. Your company and your customer base will grow, new leaders who join your company will swap out software and processes and you will have to phase out and replace tools to stay modern. Instead of being afraid of progress, embrace it, Bedard encouraged, as he left the group with a final message: “Things change fast.You’re probably going to have to pick up a new language over the course of your career. Just get ready for it.”

Watch all the Modern Data Stack Conference sessions

VIEW THE SESSIONS NOW

Start for free

Join the thousands of companies using Fivetran to centralize and transform their data.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Data insights
Data insights

Increase data team efficiency with a modern data stack

Increase data team efficiency with a modern data stack

December 11, 2020
December 11, 2020
Increase data team efficiency with a modern data stack
Topics
No items found.
Share
For high-growth companies, building a focused, priority-driven analytics team is mission-critical.

Building a data team at a high-growth company requires strict focus, a deep understanding of priorities, and establishing a foundation of trust, says Jacob Bedard, Director of Data Analytics at Dialpad.

When Bedard joined Dialpad, a cloud-based business phone company, he carried out a current state assessment, looking at the company’s areas of opportunity and areas for improvement in data insights and analytics.

He grouped his to-do list into four categories, generally following psychologist Abraham Maslow’s famous hierarchy of needs:

  1. Identify the most critical areas for business data needs
  2. Generate factors and insights that would further improve decisions
  3. Suggest ways to increase efficiency
  4. Conduct experiments that can add value

It didn’t take long before the business realized that the analytics team needed to be embedded in the foundation of all business decisions. Also, as requirements from the start, the team needed to produce accurate metrics consistently and ensure stakeholders felt confident in basic reporting.

How a modern data stack helps teams meet goals

A large part of building stakeholder trust in the data is laying the right foundation. In a presentation during the Modern Data Stack Conference 2020 hosted by Fivetran, Bedard gave his best practical advice for setting up a modern data shop — and keeping it modern.

First, let’s define the components of a modern data stack, which consists of (for a more detailed breakdown of the modern data stack, see our detailed overview).

  • An automated data pipeline tool such as Fivetran
  • A cloud data destination such as Snowflake, BigQuery or AWS Redshift
  • A post-load transformation tool such as dbt (also known as data build tool, by Fishtown Analytics)
  • A business intelligence engine such as Looker, Chartio or Tableau

With the foundation for analysis set and stakeholders confident in the data, the business felt comfortable moving towards data science and machine learning to take the business to the next level. Below are a few lessons Bedard learned along the way.

Focus on earning and keeping trust

During the assessment, Bedard learned his number-one priority needed to be gaining and keeping stakeholders’ trust. “We can’t stop right in the middle of an important business conversation and ask, ‘Are we sure we trust the data?’ or ‘How did this get put together?’ or ‘I think I saw something different last month,’” Bedard said.

Earn the company’s confidence by consistently demonstrating high data integrity and promptly responding to issues that arise, especially when it comes to accuracy or availability.

Prioritize the ‘nightmare scenarios’

Make a list of the worst possible situations your team may find itself in and determine how you can either prevent or swiftly react to them. Bedard thinks about “the things that would wake me up in a cold sweat in the middle of the night if I had a nightmare about them. And then I ask, am I actually ready for that, and will it go smoothly? Because these things, unfortunately, do happen.”

Whether your nightmare is miss-stated revenue figures, broken data pipelines, or reporting on obsolete data, plan now for when the worst happens.

Be nice to your future self

Consider how scalable your tools and processes are. Your systems should be built for future success. Metrics, input data, and processes will change over time, so make sure your data infrastructure is flexible enough to evolve, too.

This is where a modern data stack is crucial, Bedard said, since you can find the best tool for each specific job. For example, Dialpad uses Google BigQuery as its enterprise data warehouse and central data store, with Fivetran synching CRM data from Salesforce (among other data sources) into it.

Embrace change

Lastly, get comfortable with change. Your company and your customer base will grow, new leaders who join your company will swap out software and processes and you will have to phase out and replace tools to stay modern. Instead of being afraid of progress, embrace it, Bedard encouraged, as he left the group with a final message: “Things change fast.You’re probably going to have to pick up a new language over the course of your career. Just get ready for it.”

Watch all the Modern Data Stack Conference sessions

VIEW THE SESSIONS NOW
Topics
No items found.
Share

Related blog posts

No items found.
No items found.
No items found.

Start for free

Join the thousands of companies using Fivetran to centralize and transform their data.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.