Driving business value – What to do – and not to do – In Data Vault

The journey from learning about the theory of Data Vault, to becoming a qualified practitioner has a lot in common with the way in which you learn to drive a car – regardless of which country you live in.

First there’s the theory, then you get to practice behind the wheel with a qualified driver sat beside you, not holding your hand – or the wheel for that matter – but ready to step in to prevent something going drastically wrong.

After a number of hours learning and passing the theory test, you venture out for the final road test with an examiner, and hopefully rip up those L-plates when you have finished.

Over the last 40 years, the driving test – certainly in the UK – has changed.

In reality, most of us come to realise that all passing your driving test does, is make it legal for you to begin the real process of learning how to drive.

That driving test analogy was Veronika Durgin’s reference point for her talk on ‘What to do… or not to do… when Implementing a Data Vault,’ at the latest Data Vault User Group online meeting.

Veronika, who is Head of Data at a luxury fashion and designer clothing store, can call upon more than 20 years’ experience working in data, first as a database administrator, before becoming a data architect and data engineer, working for the likes of VistaPrint, Hubspot, Sonos and Indigo.

As a qualified Data Vault 2.0 practitioner, Veronika obviously believes a system of business intelligence (BI) based around a data warehouse that offers consistent and repeatable pattern-based methodology, coupled with multi-tiered, scalable architecture, is the best way forward in data analytics.

The Data Vault model is both flexible and automatable and Veronika strongly recommends – especially if you are coming to the method completely cold – that you have to understand “the why.”

Data Vault is agile – short sprints are very achievable with a good design. It is also scalable and auditable, which answers many of the questions about what comes further down the line.

Winning the boardroom battle and the hearts and minds of business-users is critically important, Veronika stressed.

So having great relationships with those users is key, and in making your case it pays to target areas of the enterprise with the highest business value – to ensure the most impact.

The scope of that project will stand more chance of success by concentrating on one or two source systems and ensuring you take all the applicable data.

One of the biggest causes of project failure, is designing and mapping a model that mimics the source system too closely.

Data Vault founder Dan LInstedt has consistently warned that is a total no-no, and as one of his disciples, Veronika would never deviate from that simple truth.

Returning to the ability to be agile, Veronika further warned that everyone wants a delivery date. So her advice is focus on that throughout, delivering small steps, but as often as possible to ensure those deadlines are met.

Investing in automation will also save time and reduce mistakes – there is no need to build a Data Vault manually – Veronika assured everyone in the room it was a “horrible experience of copy and pasting.”

Once the big day has arrived and you have passed that test, you can then keep practising to perfect your Data Vault – you can even go and acquire the Data Vault Alliance practitioner qualification.

Before reversing out of the room in fine style, Veronika produced a Data Vault demo, which also revealed her admiration for Superheroes.

You can watch a video of both her presentation and the demo above.