Good People need Good Process

Before we discuss Data Governance, I thought it useful to take a side look at Processes. It sets a framework to embed future topics within.

“Process” is not a dirty word.

The word “Process” seems to have a long-standing relationship with clipboard and peaked-capped jobsworths. It has a bad reputation as stifling creativity and innovation, but it is exactly the opposite, it creates a safe space for those to thrive.

Many think that process doesn’t benefit them or makes their life seem harder. Some see a process as a barrier to their personal freedoms. But if you look closely, good process happens invisibly without people being aware of it and when done well, it becomes the glue that allows your people to thrive and succeed.

As always, this article will be data centric, but it applies to everything. For my friends in the creative industry- Yes! it applies to you as well. You have your own practices and processes, that are part of your creative flow. You just call it something else.

Process is part of the P-P-T circle

It’s well understood, that for any organisation to succeed, it needs a blend of People, Process and Technology. This blend defines an organisation’s culture.

Everybody likes a good piece of Technology and everybody values great people. I have never known anyone come and praise a specific process though.

Process has an unfair reputation and good people rely on and create good process, even if they don’t realise this. ?

Why? Because they are not being bogged down in doing pointless tasks or wasting energy in debating the right way to perform a particular task.

They know that following a well-defined path and following best practice is always easier and more efficient.

What is the point of process?

One of the popular arguments is that if you’ve got good technology and you’ve got good people, is to just let them get on with it?

“Don’t you trust them to do the right thing?”

Well…

What is the right thing? How do you know what you are trying to do? Ultimately, you need a goal, a target, a deliverable.

The point of a process is that you are not starting from scratch each time. You are not re-inventing the wheel. You are using the experience of the people who came before you. You are using the collective experience of the team and embedding that within the process.

You can see previous results and use that to modify your approach as appropriate.

Clearly, just like anything, you can create bad process and it can feel more jobsworth and debilitating, but process development needs to be considered in the same fashion as any other item. It needs consideration, development, validation, implementation and then monitoring.

A process is repeatable. If you follow the same steps, with the same inputs, then you will get the same output.

How does it relate to data?

Data is uncompromising. It must follow a process. Without a process, you will end up with a pile of unstructured and possibly random data.

A simple system has three steps.

1.????? Data input

2.????? Execution

3.????? Output

These are then broken down into more complex steps/sub-processes or related processes. I’ve deliberately called step 2, “Execution”, avoiding the term "Data Processing" for now.

Output comes in different forms. It could easily be another data set, created by the execution phase. Equally, it could be a report, an analysis or a combination of all three.

Data needs Process and Process needs Data.

Data naturally requires a set of processes and most data packages will provide tools to allow these processes to be built and executed. These processes will be automated following some form of schedule.

However, stepping back from the technology and back to the people side of things, it is essential to ensure that other processes exist and that these are given the necessary gravity within an organisation.

These include

• Data Quality – How are you making sure that your data is accurate and up to date?
• Data Management – How are you making sure that your data is being organised correctly to avoid confusion or duplication?
• Algorithms – How are you ensuring that your models and scorecards are aligned to the data available?
• Outputs – What are you doing with the final results and possible most important do you understand what your end-users or customers are doing with it?

The Processes need to generate data as part of execution and this data can be used to monitor that the processes are working. If you don’t know they are working properly, then you will have uncertainly everywhere.

What does “Good” look like?

In many situations, the best case is a process that executes without any drama, is dependable and generates trusted results.

This can be as simple as a data input process that is generating 100% accurate data or a model that runs regularly and generates trusted results.

I’ve been in environments where we generated hundreds of overnight reports that were accurate and are available to the management teams, when they started work the next day. The object being to generate accurate information to a specific deadline and to do that every single time. (and to have processes that reacted to any failure.)

Processes will always have Service Level Agreements (SLAs). These are the measurements of success for processes and provide a view of how stable the process is. ?

Processes sit at the heart of all data operations. Regardless of the technology and the strategy, a set of processes are required. Some come out of the box with the technology, but they still need to adhere to the principles as described above.

Processes need to be adaptable, and they need to evolve as the organisation evolves. Again, it needs to be simple and straightforward to do. The last thing, a management team want to hear is that their processes are preventing them making any change.

Conclusion

Processes do have a bad reputation, but it’s all about how you relate to them. In many cases, they allow us to describe how a thing actually operates.

• We all agree don’t we, that we want to measure success?
• We all agree that we want to learn from our mistakes?
• We all agree that knowing what works and what doesn’t work is important?
• Where there is room for improvement, we should identify it and learn from it?

Processes are essential to this and they are particularly relevant to all forms of data processing/analytics.

To finish, I have some questions to discuss.

• What do you do when people are not following the process?
• How much process documentation do you need to do?
• Does process lead to a house style?
• If everyone is following the process does that make it easier for people to pick up and handover?