Do Your Stakeholders Trust Your Data Insights?

I conducted a survey to understand the biggest challenges Data Scientists face when it comes to Data Storytelling and collaborating with business stakeholders to successfully execute data science projects. 

In this survey, an anonymous respondent additionally shared this comment:

“I've come to realise that stakeholders do not really trust your analysis sometimes when they realise there may be errors from your data source. Some organisations don't really have a structured source of data or a reliable method of recording data across teams which always causes problems when it comes to analysing the data for decision-making.”

A lack of stakeholder confidence or trust in an analysis, according to the survey I conducted, was the 2nd biggest challenge Data Scientists face when collaborating with stakeholders. This lack of confidence or trust can stem from different factors - including ineffective data storytelling - but since the respondent specifically cited data quality challenges as the factor, I will address it from that angle.

Tackling a lack of stakeholder trust due to data quality challenges

What are some strategies you can employ to build stakeholder confidence in your analysis when there are data quality issues?

1. Be transparent with your stakeholders about the data quality challenges. Don’t hide it from them because eventually it will become obvious. Transparency builds trust.

   
   

2. When sharing the data quality challenges with your stakeholders, don’t just share the problems but also present solutions if possible. For example, if there is an issue of duplicate records, show the volume of data with the duplicates present, and also after the duplicates have been removed. Afterwards, advise your stakeholders on whether the volume of data is still enough to draw trustworthy conclusions. By presenting them with solutions, and not just data quality challenges, a level of trust will be maintained.

   
   

3. Assess if there are opportunities to solve some of your stakeholders’ problems with any available subsets of clean data, and deliver insights if so. This will help your stakeholders gain confidence that they can be supported with reliable data driven findings.

   
   

4. Report any data errors you encounter to the appropriate team in your organisation so corrections can be made for robust capturing of data at the root level. And if applicable, recommend that your organisation invests in hiring a Data Engineer – in other words, a data professional who is skilled in developing and maintaining data sources - since generally speaking, Data Scientists or Analysts are not trained to do so.

Conclusion

As I mentioned earlier, a lack of stakeholder confidence or trust in an analysis can stem from several others factors including ineffective data storytelling and weak stakeholder collaboration. If you're looking to grow in these areas and take your data science career to the next level, keep an eye out for my upcoming online Data Science Masterclass for Junior Data Professionals and Students in Training.