As a data analyst, one of the satisfaction I get from my work is to see the analytics results get utilized and that decisions are made based on my work. From talking with other data scientists, this however is not always the case. As also pointed by a recent study with data scientists at a software company, one of the biggest challenges of the job is to actually ensure that the results get utilized (Riungu-Kalliosaari, 2017). Reasons such as business stakeholders being not fully convinced or confident about the results, are among the contributors for this.
In a survey conducted by Forrester Consulting for KPM for their Building Trust in Analytics report, it was identified that about 34% of organisations have a high level of trust in their own organisation’s use of different types of analytics. This is a disappointingly low number, considering that many companies want or are on the path of being data- or analytics-driven and the need for data scientists in companies continues to grow.
Trust in analytics and the results are important for its effective use. There are several reasons as to why there might a trust problem. Ramaloo in his post on Hidden Insights, acknowledged that in most businesses, there is a disconnect between the data scientist teams and business units. “Business units often don’t understand the predictive models and how they can really add value to the organization. This can be hugely frustrating, especially for the data scientists, who have often invested a lot of time and effort into developing these models”.
There was one webinar I attend which tried to provide reasons why there is an analytics ‘trust gap’. The facilitators pointed out the following top reasons why this happens:
- Incomplete data,
- Data is not clean,
- No single version of truth,
- Understanding output of analytics: This goes to how the output of analytics is presented. If there is failure to talk of the results in terms of business, there is a low chance of adoption.
- The mindset of decision makers – experience vs data: Since sometimes analytics and data do not give the complete picture – some high management might feel that their experience is better.
There are several strategies proposed for instance in the KPM report and in the webinar about how to bridge this trust gap but I will not detail them here. From my experience so far I believe a lot of the trust problems can be addressed by having data scientists be not too far removed from the decision making process and end beneficiaries of the results. Data scientists should work early on with the business units making the decisions as a way for the business stakeholders to gain understanding and confidence in the data, the process, and the people doing the analysis and hence the results – as well as for data scientists to better develop more effective, consumable, actionable and domain-guided analytics.
What is your take? How is the trust between analytics and decision making in your organisation and how have you managed to this trust gap be bridged?
Also published on Medium.