Trust In Analytics

What Creates Trust in Analytics?

As the volume of data grows and more insights are generated from this data, many organizations are asking the question: how can I have better trust in analytics results?  Because without that trust, business teams cannot be fully confident in becoming more data-driven and driving actions based on analytics.

When creating a great culinary dish, it often boils down to two key aspects: am I using the best ingredients and how do I cook the dish?  In analytics the ingredients are data while the cooking is the process the analyst performs to produce the results.  Both aspects are critical in creating trust in analytics.  Let’s explore how.

Data Literacy

In Predicts 2020: Analytics and Business Intelligence Strategy, a key trend and need Gartner discussed was data literacy — “learning both to ‘talk data’ and to apply and interpret it with a business context.” Data literacy is an essential part of building a data-driven culture and organization.

While data literacy has a common definition, it has slightly different variations to individual personas in the decision making process:

  • Data Analysts – to the team doing the analysis, data literacy focuses around understanding the various datasets at their disposal and how to apply them in answering analytic questions.
  • Knowledge workers – to those trying to get a deeper understand of the analytic results, data literacy centers around interpreting data from analytic results to the business situation.
  • Management – to the decision makers, data literacy focuses on if the data tells me what I need to know, does it make sense and does it give clear guidance to a set of actions.

Data literacy is an essential part of trust in analytics.  The trust is built around some essential pillars of data literacy: knowledge, understanding, sharing and transparency.

Creating Trust in Data

Trust in data is seen as a major factor holding back data-driven decisions.  Building trust in the datasets at their disposal helps the analytics teams in two key ways: it speeds the analytic cycle and helps deliver accurate, confident results.  Trust in datasets helps the analyst know where and how to use the data, eliminating time wasted in finding the right data for the problem.  And once the right, trusted data is gathered, the analysts can dig deep and find meaningful, detailed insights.

To an analyst (as well as other personas), trust in datasets is built from:

  • What is the source and where did the data come from?
  • What information is captured in the dataset?
  • How is the dataset related to other datasets?
  • How has the dataset been used for other analytic problems?
  • What does the data tell me?
  • How often is the dataset used by analytics team members?
  • Which analysts created the dataset and which have used it?

Armed with this information, an analyst can make a make a fast, guided decision on whether and how to use the dataset to answer their question, and quickly move into the analysis phase.  This knowledge also helps speed the analysis phase by providing clues as to which areas of the data to explore.

Creating Trust in the Process Through Knowledge Sharing

There is a second type of trust that needs to be built: the business teams trusting the analytic results. This trust is necessary not only for them to feel confident in executing plans based on the results, but is also important for the teams to move away from gut-feel decisions and become more data-driven.

The key to creating trust with the business is allowing the teams to trust the analytics process behind the results.  This requires an transparent, easily explainable process that shows the business teams how the results were generated and explains the meaning behind the results.

To facilitate this, the analytics team needs to supply as much information about the data as possible, which happens through:

  • Capturing plain language (non-technical) and business level descriptions about the data behind the results
  • Sharing analyst knowledge about the data – where it came from, how it was captured, why it was captured
  • Sharing ways the data has been used and other insights generated from it to help business teams put context around it
  • Having a complete, easily readable lineage of how the data was modeled and transformed

A robust knowledge-base and collaboration facility around data and analytics processes helps facilitate the capture and sharing of the information about the data.  This also helps with data literacy in the business, as business teams are continually exposed to this knowledge and can apply it to their decisions.

The Role of Governance

Governance is always a critical aspect of the data and analytics processes.  Policies that are too strong strangle analytics efforts and if too weak it creates security and privacy risks.

The same analogy is true relative to facilitating trust in analytics.  Governance that is too strong can hurt the knowledge sharing that facilitates trust.  Governance that is too weak presents the same risk as above – too much knowledge sharing creates security and privacy risks.

There are ways in which governance policies can walk that fine line.  Policies could be put in place that facilitates knowledge sharing without necessarily exposing the underlying data.  For example, here are different levels of governance that could support this:

  • Data analysts requiring complete modeling and use of the asset would be granted full control on the asset
  • Other data analysts who would need to use the asset would be granted access to use the asset
  • Business analysts and or knowledge workers would be given access to output from the asset (not the underlying data) and the knowledge around it

This facilitates the sharing knowledge of knowledge and proper interpretation by the business, while maintaining the right levels of governance around the raw data.

Wrap Up

Building trust in analytics is incredibly important ranging from the source data to the analytic process to how the results are interpreted.  Building trust in data facilitates faster analytic cycles, and creating strong trust in the analytic process and results enables swift, confident data-driven actions.

The Neebo Virtual Analytics Hub is a cloud-first solution allowing analytics teams to find, create, collaborate and publish trusted analytic data assets in complex hybrid landscapes.  Neebo provides unified access across data silos, increases use of data assets and furthers data knowledge to build trust and rapidly answer new business questions. A key capability Neebo provides is collaboration, knowledge sharing and usage information about data and analytics assets to build trust in the results delivered.  To learn more visit the Neebo website or test drive Neebo by registering for a free 14-day trial.

Sign up to our newsletter

Editors Pick

Analytics Knowledge Management Collaboration analytics knowledge management
Advanced Analytics Banking Customer Experience
Best Practices Collaboration General big-data-analytics

Want to see our Virtual Analytics Hub in action ? Request a demo

Request a Demo