Everything’s better when you can do it yourself, right? That seems to be the way many software companies and startups think about the new economy. Why use a taxi service when you can just catch a ride with someone (Uber)? Why stay at a hotel when you can just crash in someone’s extra room (AirBnB)? And while I applaud the innovation and disruption these startups have created I have to wonder: is self-service always best? Or are there times when it pays to use a professional? In the software world there is a push toward self service analytics — basically giving people access to the data and analytics tools they might need so that that individuals can run their own analyses, interrogate the data, etc. On the surface, democratizing the data and analytics seems like a great idea, especially considering the dearth of qualified analysts to fill a glut of jobs. And as data becomes more central to jobs in every sector and field, it seems to make sense that the HR department should be able to run its own reports as should the sales team, the marketing department, and so on. But what I have now seen many times is that companies have stopped analyzing data and producing reports centrally and instead move to 100% self-service. And I believe that this is a mistake. When no one is taking charge of the interpretation and analysis centrally, it often means that no one is getting the big picture. If companies only offer self-service analytics they run the risk that people miss key insights, misinterpret the data or perform the wrong analysis. The problem is that the people being given access to the data aren’tanalysts, and don’t necessarily have the mindset or skillset to view the data holistically. Certainly a smart HR manager could pull the numbers she thinks she needs to understand about hiring, retention, performance, etc., but what if she is asking the wrong questions? What if she is focusing on the wrong data points? What if her data conflicts with what other department heads are seeing? I believe that identifying key strategic insights from data, packaging these insights into good headlines, graphs and narratives, and explaining it to people is still very important as it ensures the key messages are heard and key corporate stories are created. And I believe a good analyst should perform these functions. Then, that same HR manager receives reports with expert analysis highlighting the key insights from the HR data but then has the ability to perform further analysis by herself. I believe it is absolutely important to give more people access to data and self service analytics to take away the over reliance on central teams. But this has to be balanced with the need for key insights and messages that are centrally coordinated. Self-service analytics should be a complement to a data science team, not a replacement. I want people across all levels of an organization to feel empowered to consume and use the data they need; but I also want someone at the helm, steering the ship, able to answer questions and help the company and its individual departments course correct based on the data they receive.