Skip to Main Content
Sisense
Menu Button
Demo
Home > Glossary > Data Maturity

Data Maturity

May 9, 2019
Table of Contents

What is Data Maturity?

Data maturity is a measurement of how advanced a company’s data analysis is. A high level of data maturity is the stage reached when data has woven its way deeply into the fabric of an organization and when data has become incorporated in every decision that an organization makes.

Some great examples of advanced data maturity are early data innovators that are already proving to be dominant in their industries. Companies like AirBnb, Uber and Netflix take data so seriously that it’s more accurate to call them data companies than to consider them traditional competitors in the hospitality, transportation or entertainment industries.

The development of data maturity

Answering questions with data is not a new idea, but modern techniques have evolved rapidly and spread widely. Companies across the world are at different stages of experimenting with data and realizing what kinds of decisions can be made via deep analysis. Since these trials and realizations are happening at differing speeds, and often start from scratch within each organization, there are significant disparities in maturity between otherwise similar organizations.

As the developer of a data platform, every day we deal with companies that are in all stages of their data journey. Some are just realizing that they need to combine all of their data sources into one place, while others are running advanced queries in Python or preparing data for predictive machine learning models. At Sisense, we have a unique perspective from which to observe the progress that data teams worldwide are collectively making.

As organizations develop their use of data, and improve their adoption of advanced analytics, they develop characteristics that we have identified as five stages of data maturity.

These five stages show a way for companies to look at their overall data processes, ask some difficult questions and start a discussion about how to do more with data. They can also be used as a roadmap for organizations looking to make long-term plans to develop their data team into a more powerful resource. The stages are:

1. Business Reporting

  • Siloed system of record reporting
  • Canned SFDC / Google Analytics reports
  • Merge data in offline spreadsheets rather than a data platform
  • On-demand reports only, no automated reporting
  • Can report on siloed data (Salesforce, Google Analytics, Facebook, Marketo, etc)
  • Business rules are consistent inside a single silo (Salesforce, for example) but not across silos
  • Exclusively deals with backward-looking, descriptive data

This stage is the beginning of any company’s data journey. SMEs and early-stage startups are here. They’ve recognized the need to collect data for their records, but haven’t built any kind of structure to do serious analysis of that data, likely because they don’t need to. These companies export their Salesforce or Marketo data and keep it siloed in a series of spreadsheets on a local device rather than blend it together for cross-functional analysis in a data platform.

2. Business Intelligence

  • Standardized datasets
  • Reporting in one place
  • Refresh cadence
  • Sales and Marketing data sources connected, aggregated
  • Business logic lives on the report
  • ETL and warehousing based on data volume
  • Exclusively deals with backward-looking, descriptive data

By this stage companies have blended data together into a single warehouse. They have a more holistic view of their data and they see a bigger picture emerge. They can now go beyond asking questions like, “what were sales last quarter?” to others like “how did the marketing campaigns from last quarter affect sales?” These companies don’t have one source for their Salesforce data and another source for the Google Analytics data, etc. They have one source for all of their data.

3. Ad Hoc Analysis/Insights

  • Reporting is sophisticated enough for investigative analysis and rapid model development
  • Data connected across sources, analytics environment abstracted from origin tables
  • Centralized business definitions in a warehouse / model
  • Retain ability to query across modeled and unmodeled data
  • Warehousing is essential at this phase, data lakes are useful
  • Beginning to do diagnostic analysis, moving past recording a spike in data into pinpointing the cause of that spike

At this stage, companies are gaining more autonomy in the analytical question-and-answer process. Before, they could get answers from their data sources, but couldn’t ask their own unique questions. Now they can. It’s really important at this point that a company has an independent data team with personnel sophisticated enough to use SQL, Python or R to create their own data models.

At this stage of maturity, it’s natural that companies start to have conversations about data democratization as a way to expand the reach of existing insights.

4. Hybrid Centralized Data Teams

  • Data landscape is holistic — business rules are versioned and managed
  • Central teams define systems and methods, embedded analysts provide specific value
  • Business model relies on harmonized data across product, sales, success, marketing and ops
  • Data has C-level representation and visibility, business-critical
  • Starts to analyze predictive data

By now, data analysis is sophisticated enough that it is a regular part of every team’s routine operations at a company. The demand for data is high and it’s vital to find a way to prioritize requests to make the most of the data team’s resources. At this stage, there’s a shift in organizational structure to a hybrid model. The centralized data team still exists as a means of collecting information into a single source of truth and building sound data models, but we start to see individual analysts embedded in different business functions who are in charge of answering questions specific to that line of business.

While maturation to this stage is marked by personnel changes, it also requires a lot more from tools and technology. There are now requirements around governance and engineering that didn’t previously exist. There are now several additional steps in the overall data analysis process, but the result is a flexible, scalable data function that can answer several pressing business questions at once.

5. Predictive Analytics and Machine Learning

  • Business forecasting and planning operates on projected data
  • Online models have product and business ops impacts
  • Offline models used to manage and mitigate negative business dynamics
  • Data lakes are required
  • Regularly using predictive and prescriptive data to make decisions

This final stage is for the most cutting-edge data operations. They have the technology and the tools to answer questions that other companies aren’t even considering yet. They’re analyzing information that they’re seeing right now as a way to make decisions about future products, markets, customers, staff, etc. At this stage, companies are investing in more than just how they run their business well. They’re looking at how to make fundamental improvements to the company based on sophisticated data models.

Reaching each stage of data maturity takes increasing amounts of time and resources. Becoming more data mature requires a heavy investment in technology and people. There’s a big payoff, but the investment takes a longer time to recover. As you become data mature, it’s important to manage your own expectations and be realistic about the value that you’re getting from these investments and when you’ll reap the rewards.

Back to Glossary

Want the latest in analytics?