When working with Clients, the typical BI consultant will come across a vast array of different levels of maturity of the data landscape. From the most basic reporting solutions such as Excel reports, to the most advanced artificial intelligence solutions. Being able to identify at what level a company currently is operating makes it easier to talk about how to make things better.
While there are some very interesting models available by consultancy companies on these maturity levels, they often work with one single dimension of improvement. However, some companies can get by operating at a lower level of maturity in one area, while needing the best of the best at another. Think of financial companies with strict government regulation or highly automated industries.
These points are what I think are 3 key areas any BI specialist should be aware of:
- Data Governance
- Data Driven Operations
- Level of IT Infrastructure
It seems counter intuitive at first, but only at the third point are we actually looking at the technological level of the company. This is to further underline that the technology used is always a tool for delivering the business implementation that a company wants to perform. While we of course like to work in complex high-tech environments, for some companies this would be overkill. For us tech geeks it can be a painful confrontation when you realize a company only needs a manually updated Excel workbook, rather than a Data Warehouse.
In the next paragraphs I will take you through these 3 points. We will look at the lows and how you change them into highs. This can be a great tool to judge how mature your BI situation really is, and how you can make steps to improve it!
When looking at any company with the perspective of an IT specialist, you must realize that using computers on such a scale is something of the last 30 years at best. Working with data to this extent is relatively new.
At the lowest level of Data Governance companies have no clear understanding what data really is. They might see it as a language that applications use to talk to each other, but like any spoken language you need some agreements on what is what! Imagine the confusion if you would have an international department all talking in their native language to each other!
A first step would always be to introduce some form of business rules or data descriptions. When discussing different data concepts between departments, everyone needs to agree about what the concept is they are actually discussing. The great part about this is that a first step in any direction is a good first step. If we get this done, we can start integrating data way more easily and make our processes a lot more efficient!
When integrating data, the question often arises who changed what and when? To make sure we track all of our data changes we need to implement some kind of audit trail. If we leave this behind, a user can only look at what the company looks like today. We don’t just want today, we want to be able to look at our company through its entire timeline!
With our audit trail in place, we have prepared the company for one of the biggest challenges. Legal compliance. By describing our data, having it fully integrated and even having a conclusive audit trail we are fully prepared to tackle tasks like GDPR! This will make our board of directors at the very least 50% more relaxed, and a relaxed board is a happy workplace!
The good news is that even the weakest company in the Data Governance aspect can make quick and cheap steps at becoming better. We provide value every step of the way, and this is a great project to make business and IT work on one common goal.
Data Driven operations
One of the buzz terms now is “Data Driven Decision making”. No decision should be made on a hunch of a director but should be clearly supported by facts and data. What this often leads to however is bombarding management with dashboards, of which they only use one or two and ignore the rest. We can do better!
In many small companies, managers are so tightly involved with operations that they don’t need anything fancy. They know what is going on. However, the human brain only has a limited amount of space so if the business gets bigger we need a place to store this information.
An easy first step for any company to get better at this is by setting up KPI. What do we want to achieve? What is our goal for this period? By not setting goals, we are setting ourselves up to drifting and this makes us paralyzed to any form of decision making. When we have defined our KPI though, we must capture the relevant data to know if we’re going the right way. This means that we need to implement some basic data integration.
We asked management some basic questions, and we immediately have our buy-in to start working on a BI solution!
We can do even better. By capturing data in an easy to understand data model and teaching people some basic querying and reporting skills, we’ve set up a support organization that can easily identify trends in the data and make data driven management decisions. Why stop there though? If we can make sure our data models stay user friendly and push these new data experts to teach others, we can get proficient people everywhere in the company creating dashboards to maximize their daily effectiveness.
Data driven operations isn’t about just making management reports. By simplifying the IT landscape for smart users and supplying them with easy to use tools, we create a culture where everyone is comfortable using data to their advantage!
Level of IT Infrastructure
To have data that we can use perform our analysis, we need some form of IT landscape to get data from. While we can extract information from separated excel workbooks, it is considerably easier to work with more mature infrastructures.
Let’s set up a prerequisite here though. The company in question should be mainly working with computers and should have some form of centralized or standardized IT solution available. Convincing a client that having computers is an added value in [current year] is maybe a topic for another time. What we can focus on now though is how we can raise the level of the IT landscape.
When trying to improve the level of IT within a company we need to think of what we want to achieve, and that is automation. Our primary goal is to make workflows easier for people on the work floor. To make a workflow easier, we first need to describe one. To make life easiest for us, let’s pick the process that contains the least amount of branching and describe it in a tool like Bizagi Modeler.
With the preferred process documented and validated we can start on our adventure. While it is difficult to give any be all end all solution for how to optimize any solution, we can talk about some steps you can take to be of use during this process.
- Find representatives of the departments associated with the process
- Showcase what BI can do in terms of Data Movement, Integration and Visualization
- Identify where the process is currently bleeding efficiency
- Together define what would ideally need to happen to stop the bleeding
- Define a pragmatic solution
- Develop and test the solution together with your representatives
- Integrate working with the solution in the standard workflow
- Make any non-technical data accessible through either a seperate database (an Operational Data Store) or through data dumps
While this progress is overly simplified, the main point to take from this is that you sit down with people on the work floor to make their lives just a bit easier. While creating this solution on its own does not impact the total IT landscape as much, we have gained a platform for ourselves to redo this process in the other parts of the company. We can even reuse some of the tools we have built!
Why should we want to do this process over and over again? The more of these solutions we implement, the more valuable an integrated solution will become. By combining the data of all of these solutions we have created our first Enterprise Data Warehouse! It might not look very impressive right now, but it will grow over time.
Our options of improving the IT landscape are limitless now, and we are able to drive change much more efficiently. At this point it becomes important to get some people from the business up to speed on their technical skills so they can assist in the overall BI development. This gives people even more feeling of “joint ownership” over the data solutions that you wouldn’t get when only hiring external resources.
Most of the what we have discussed in this blog boils down to the power of effective documentation and working together with your stakeholders. By creating descriptions of what the current business is doing, we create more clarity in where we can help the them improve.
We create data descriptions to get our Data Governance policy up and running.
By defining operational goals, we make our first steps into Data Driven Operations.
Through documenting workflows, we start improving the companies Level of IT infrastructure.