BI and analytics are a bit like having a baby. It is incredibly painful and exhausting, but the end result is so beautiful, we quickly forget how bad it was. In BI, everyone focuses on the pretty charts and graphs, and quickly forget the laborious and resource-intensive process it took to get there.
There are a few painful steps that everyone needs to experience in order to cuddle that first table or chart. You need to determine:
Where your data is located
What data you need to move
How to move that data
How to assemble the data so it answers your question
How to write a SQL query
Typically, as I explain these steps to people, they start to nod in painful remembrance of what it took to actually answer their business question.
It doesn't have to be like this. BI and analytics exist for one reason: to answer questions. That’s it. So why not start with a question instead of looking for a bunch of data? Wouldn't it be great if you could simply ask a question and that long, laborious process of finding the data from multiple different databases, data warehouses and data lakes, figuring out how to move it, how to join it, how to model it, how to put it together, how to write a query and then querying it, just went away? Imagine if that process could be reduced to under 3 minutes. It can.
In almost every conversation I have with potential clients, the top 4-5 questions that they ask me are almost always the same, and usually asked with a fair degree of skepticism:
How are you going to move all of my data?
We don’t. We never touch your data. We only move the metadata, which is the data about the data, so your data is never compromised. We pull your metadata directly into a metadata index, allowing us to search for data as if we have the entire file. You can search by name, type, vendor, IP address, context, or even by asking a simple question. Lastly, we only move the tables related to answering your specific question. No more. No less, resulting in a highly efficient, fully automated process.
Does it only work for relational databases?
No. We actually work with any data source, automatically discovering where the relevant data is located across your infrastructure.
I need to first pre-define the data and join all the data ahead of time, right?
No. The whole point of being able to do ad-hoc analysis is that you want to answer questions that haven’t been asked or use data that has not been discovered. Promethium discovers the data and figures out how to assemble the data all in real-time AFTER the question has been asked. This avoids the painful task of identifying ALL of the data and relationships BEFORE the question has been asked.
But you can only do the query if the data is all on one table, right?
No. We do it across multiple tables and across multiple vendors as well, eliminating the time-consuming and costly step of manually searching for your data’s location.
At this point in the conversation, there is a long pause, followed by either an almost imperceptible head nod or grunt. The light is coming on, but the skepticism is not completely gone.
I still have to write the SQL query, right?
No. We automatically do that for you, providing step-by-step instructions on how to assemble the query, and with Promethium, you don’t have to know SQL code. What’s more, the directions can be easily attached to an email and sent to the relevant person in your enterprise to run the query.
These aren’t the only questions I am asked, but they are the top ones. There are other valid concerns that people have before the lights go completely on, usually starting with this one:
In order to locate my data, you would have to know the exact names of the tables and the files and nobody knows that level of detail.
If we can’t discover those names directly against the data source, we actually integrate with all of your data catalogs, so we can see all of the tags your business users have identified as the real business names against the actual technical names. If we can't find that across the source and we can't find it across your data catalogs, then you have a different problem.
What if the data is dirty and the results are wrong?
If the results are wrong, you want to know specifically what to clean and where. Many solutions only tell you that in your five petabyte data warehouse, there is some bad data. That’s not very helpful. But since we are only answering a specific question, and know that to do that involves these four tables, and these six rows we can quickly and easily clean that data, which greatly increases the accuracy of the answer you are getting.
So that is how we take the pain out of BI and analytics. The long, complex and disjointed process that you are currently using is gone. A thing of the past. The dozens of tools that don’t integrate well with each other: gone. The teams of people and months of time it takes to get through the process: gone. With Promethium, you simply start by asking a question. Promethium finds the correct data, moves only the data needed, assembles the data and runs the query. It can be this simple, and with Promethium, it is.