Minimum viable data warehousing?

Martyn Jones, Santoña, 10th October 2024.

Narrator: Do minimum viable and acceptable products exist in data warehousing? That’s the question.

Dud: Did we ever do minimum viable and acceptable products in data warehousing?

Pete: If you listened to the cretinous pundits, Dud, then you could imagine MVP and DW have nothing in common.

Dud: That’s not what I heard, Pete.

Pete: What?

Dud: Well, this bloke at the mall said that to me.

Pete: Oh dear, oh dear, oh dear! How often have I warned you about engaging with random idiots at the mall?

Dud: Sorry, Pete.

Pete: Okay! Come on, pony up! Who let these twits in? Had he been at the blow? At the Mall, indeed.

Dud: Funny you should ask, Pete.

Pete: Sounds like more bullshit bingo, Dud. You should pay more attention to the rappers. Data Warehousing epitomised the notion of business demand-driven minimum viable products. The data marts were your very first data and analytics data products and the data warehouse. That’s also a data product.

The problem with concepts such as minimum viable product is that every fool and their dog will have their own definition of what it means.

Dud: What does a minimum viable product mean to you, Pete?

Pete: I work in IT, so I use the term. I’m not even sure if it’s used elsewhere. There are a lot of synergies and stupidities around these days. For me, and in the most formal sense, an MVP is a product development strategy focusing on creating the most basic version of a new product that offers enough value to attract early adopters and validate a business idea. The goal of a minimum viable product is to get to market quickly with minimal effort and investment, gather user feedback and iteratively improve the product based on that feedback.

Dud: And MVPs in data warehousing?

Pete: Yes, Dud. You can have MVPs in data warehousing, AI, advanced analytics, data mesh and business intelligence. But beware, Dud, the form and nature of MVPs will vary widely and on a case-by-case basis. Not all MVPs are the same, nor do they always deliver the required outcomes.

Dud: What do you mean, Pete.

Pete: I’ll give you an example. A bank creates a smartphone MVP to allow clients to ask for a loan in just three clicks. The MVP is developed, tested and delivered to the market. Initially, it was a success. Then, the main competitor of the bank, sensing that the ripe is ripe for this type of functionality, copies what the first bank launched. Still, instead of a three-click process, they provide a one-click service. So, the first bank offers an advantage immediately outclassed by the second. The lesson here, Dud, is that if your MVP is small and straightforward, it can easily be copied and outclassed by the competition. Bigger and more complex MVPs are much more resilient to short-term competitive expediency. But they all come with risks. If a small MVP fails, then it’s a slight loss. If a big MVP fails, a more significant blood loss will occur. However, screwing up a small but potentially highly monetizable MVP also is a massive loss of opportunities.

Dud: So what’s the advantage to the customer, Pete.

Pete: These IT organisations are doing nothing to benefit the customer, Dud. They are using money, time and other resources to create activities, initiatives and launches that they think will endear their senior management to them.

Dud: So, it’s a bit of a con then,  Pete?

Pete: Not always, Dud. Buy, yes, quite often. And who are the most significant con merchants behind these cons, Dud? Senior IT managers, ignorant project managers, bent business managers, corrupt IT service suppliers and degenerate industry analysts.

Dud: Seriously, Pete?

Pete: Seriously, Dud!

Dud: Blimey! There goes the neighbourhood!

Pete: MVPs have their place in analytics, data warehousing and AI, Dud. But, they should be defined and developed within their proper context and not within a generalised, amorphous and ill-defined whole.

Narrator: MVP initiatives can become unhinged and unservable for various reasons. We can fail to properly understand customer needs and priorities. MVPs can be so small as to be insignificant. They can fail to deliver any noticeable advantage. The process can forget to include the proper management of expectations and feedback. MVPs can miss windows of opportunity. It can be built on a mountain of technical debt, poor marketing and introduction into the digital sphere, and lousy team alignment. Like any other approach, MVPs are as good as their worst examples.