In all Head of Data and Data Analysts' job offers, there is always the mention of "build trust in our data".
What does that mean? Why is it important? And how to actually build trust?
Trust is the belief in the reliability and truth of someone or something.
When focusing on data, trust means that when one uses data to back a decision it won't come back to bite them later. The worst thing you want is one of your business user proudly backing a decision with data found in their brand new dashboard and you coming a few weeks later saying "I am really sorry but the revenue data you used for your presentation to the CEO is wrong. You forgot to subtract the refunds."
Data is generated at every steps of a Customer's lifecycle from early prospecting to your customer finally churning from your product/service.
Over the years, the data space innovated and iterated quite a lot to focus on what we now call the ELT pipeline. Data is Extracted from all the different sources, and then Loaded in a single data warehouse (or data lake, but as it's not yet Friday I won't fight right now on the terminology even though I have my own preference). Great. But while gathering all data sources in a Single Place brings reliability, it is however not enough to become the Single Source of Truth - SSoT for those in the known.
The truth follows and comes from the Transform layer.
The Transform layer will define the domains that are validated by the Data Team and therefore will highlight where the truth can be found. In particular, the Transform layer will define the various metrics and the few KPIs needed for the Business — read more on the differentiation between metrics and KPIs. Those definitions will bring uniqueness and consistency.
Properly setting up a data pipeline requires actual technical skills, often called hard skills but trust is actually a Human feeling and nothing will happen without your data team's soft skills.
Soft skills are interpersonal, social skills that boost communication between colleagues or teammates. More generally those are non-technical and develop over time based on your experience. While there are many different soft skills,
I would like to focus on 2 characteristics that are particularly important for bringing trust around data and that you can start to implement (or improve) starting this week: consistency and communication.
Your stakeholders should see the Data team as a consistent whole, on any topic from prioritization to field naming to processes. Consistency often starts by writing down a process and then making sure you follow it over time. For example, when you prioritize the topics you'll work on make sure you always follow the same prioritization rules so stakeholders know when and how to ask you for something.
Consistency will build habits both for your team and your stakeholders. Psychology research shows that habits bring happiness. And happiness is followed by trust. But this research also shows that you should start small and iterate if you want your habits to succeed.
I repeat start small and iterate.
Very often, analysts I interview complain about the lack of data literacy of their business counterparts. It's frustrating for them but it is very often due to a lack of communication.
People are smart and communication helps them understand the whys hidden between what initially felt like an arbitrary decision. Communication, and written ones particularly, also helps create connections in people brain that will be triggered when facing similar circumstances later down the road. "oh! I can remember reading an article about churn and our levers. I'll find it!" said a Business Users when faced again with a raising churn.
Again, consistency is key even for communication. Do not promise a bi-weekly newsletter if you are not able to send it consistently (I can say that now that I built our bi-weekly Huspress habit, and you can subscribe too!).
Do not forget, trust is a process. It needs reliability and truth but your soft skills will consolidate this trust. It won't appear overnight but be careful as it can break fairly quickly.