Data architecture and the rise of modelling

I wrote a blog 6 years ago (https://www.linkedin.com/pulse/data-architecture-poor-younger-brother-now-rising-star-ben-joyce) where I referred to data architecture being the poor younger brother to other architecture domains. I had assumed we were about to see a big up-tick in demand, and I think that has been true, but it’s primarily (at least until now) been more on the data technology side of things.

Rightly or wrongly, when I am engaging with clients re data architecture, I split data into two:

• Data Governance (Modelling, Strategy, MDM)

• Data Technology (Platforms, BI, Analytics)

The primary reason I have split it, is due to capability in NZ. I’ve now been back in the country for 6 years and know most data architects. What I have found is that there are only a handful of people that go across both governance and technology. The strong majority of data architects are technology/platform focussed. For those in that space, the market has been great, there seems to be constantly growing demand for those sorts of skills.

The big trend I have notice in the last 8 months has been the growing demand for data modelling skills. This has traditionally been the un-sexy side of data. It’s a lot easier to get people excited about technology and what it can do, compared to good governance and modelling processes. However, from clients I have been speaking to, who are investing in AI, they have come to a realisation that good enterprise data modelling allows them to take better advantage of AI and just implementing technology without looking at the overall structure and governance is now very limiting.

Unfortunately for NZ because the demand just hasn’t been there, we do not have that many data modellers out there, everyone jumped into the sexy platforms and left modelling behind. We’ve now got multiple organisations (ANZ, NZ AA, Southern Cross Travel Insurance to name a few) all actively looking to build up their capability and another bank hire 4 data modellers at the back end of last year. There also (at least to my knowledge) aren’t many data consultancies that can be called upon to do this work, because, like I highlighted above, they all got attracted to the sexy platform programmes.

There have been some larger consulting orgs investing in establishing modelling practices, but they are primarily hiring grads and flogging them off as consultants, hoping to make a sweet margin, assuming that their clients don’t know what good modelling looks like.

If you are someone who is in the data modelling space, I do now, for the first time think there is going to be a lot of work there for you. I still think you will come up against blank stares and glossed over eyes as you talk about what you do, but there seems to be more understanding that, although not the most exciting topic, it’s becoming necessary.

Also as an attempt to ensure this did alright in the LinkedIn algorithms I thought I’d add a photo, after searching on google, I thought, lets use AI.. Turns out putting the above overview and asking it to make an imagine about data modelling becoming sexy was not the right move..