Home / Analytics  / Complete Reporting Automation: Are We There Yet?
Abhiram Bhagwat

Abhiram Bhagwat is the senior technical lead at The Search Agency, where he has worked since 2007.

1 Comment
  • Michael Pratt

    Abhiram – I really appreciate the title and direction of the post. I’d like to offer a few contextual statements. One big question to ask is who is the intended audience of the report. That is critical because the software tools you cited (Domo, Tableau, etc) are built primarily as tools for operators, I like to say. In our world (marketing) an “operator” is the person managing the campaign. Those tools, awesome as they are, help visualize data for analysis and operators. They are really good at that. That’s why they create dashboards. Juts like a car’s dashboard is for the operator of the car.

    Reports are much different and they really don’t produce them. We have obtained so many customers for that very reason.

    A report is a narrative. Dashboards tell you what is happening right now. In marketing, a report tells you 1. What you did with the client’s money (actions) 2. What happened as a result of the money you spent 3. How the client performed (results) and 4. What to do next. A dashboard is a round peg and a square hole in that equation.

    So, to automate report generation must start from the point of view of the narrative being assembled. Much of what you so eloquently pointed out contributes to that but I’d argue only after the narrative has been established.

    With dashboards, you start with the data and represent it according to the function/role of the operator. With reporting, it is the exact opposite. You start with the question being asked (by the client) and then retrieve the data necessary to answer the question and THEN represent the data visually in a way that bests help convey the answer.


    BTW – We’ve figured out a way to automate that process using symbolic AI

    March 5, 2018 at 11:08 pm Reply
Post a Comment

Subscribe to Newsletter