DataOps: ‘Build it’ or ‘Buy it’?

by Maxx Silver
February 15, 2022
Arrow Pointing R

“When you need capability that you don’t have, there are two options: you can build it yourself, or you can partner to get it.”

 

This was a comment made by a Senior IT Leader in a meeting of his organisation’s executives with BizCubed CEO, Zachary Zeus. We’ve thought a lot about that comment, largely because it reflects what we’ve seen in the market.

Most organisations that have recognised the need for (and value of) a data capability have jumped straight to building it themselves. You can sort of imagine how this plays out:

  • Business teams invest in data analytics
  • Analysts are limited by data quality
  • Technical teams propose investing in data capability
  • Executives allocate budget
  • IT teams deploy data engineers, data architects, and (potentially) data scientists

For some organisations this approach might make sense. These organisations are likely already invested in robust technical teams, often have mature digital capabilities, and might already be responding to threats from digitally native disruptors. But this approach won’t make sense for every business.

So what’s the alternative?

First, we need to dispel a common myth – that your organisation’s data is somehow unique. I know, this statement amounts to blasphemy in some organisations. When I was an analyst I would have, and in fact did, argue that there was simply no way that any organisation’s data needs compared to ours (and this was true, so far as it went).

Looking back with what I know now, I recognise that I correctly understood that the intricacies of our data ecosystem were akin to an organisational fingerprint and there would never be another like it. Further, the ways in which this complex network of interconnected systems and processes (what we at BizCubed call “business logic”) served our organisation were entirely unique.

But I was wrong in the sense that regardless of which systems, databases, and applications we used, or the ways in which those components were uniquely combined, the underlying structure was fundamentally consistent. Put another way, there was a finite number of data sources, a finite number of data types, and an infinite number of use cases.

So, if my former organisation’s executives came to me today and asked how they might get the best return on their data investment, here’s what I would tell them:

  • START by partnering with a data team that prioritises Data Operations (as opposed to consulting or other project-based approaches)
  • STOP investing time and money in skills that are fundamentally repeatable and scalable (e.g., data engineering)
  • CONTINUE investing your organisation’s best resources (read: people) in value-added work such as building business logic, creating data visualisations, and performing advanced analysis

Ultimately, whether organisations decide to build their own data capabilities or partner to get them will depend on the costs and benefits specific to that business. It’s important, though, that business leaders recognise that there IS an alternative to building it themselves, and that the alternative may ultimately be faster, cheaper, and more effective in the long run.

Want to see how BizCubed’s Data Operations as a Service offering has made a difference for Bing Lee?

Click here to read the case study

More blog posts

Building a New Habit is HARD!

 “Building business logic is a low priority, from my perspective.” This was what my CEO said to me as I shared with him my priorities for the week. The funny part? I work for a data engineering firm...

read more

Data Engineering and Infrastructure

This week we continue our exploration of The BizCubed Method and the Data Maturity Model. See this post for an overview of the model and this post for a deep dive on Security. This week we tackle...

read more