Mikkel Kvist is Head of the Group Business Intelligence Competency Centre at MT Højgaard.
As a major construction firm, MT Højgaard needs to be economical, efficient and accountable in every single aspect of its business. For this reason, we have a Business Intelligence Competency Centre to promote rapid, data-driven decision-making processes across the entire organisation. In my previous blog (4 signs our BI Architecture might need automation), I outlined some of the reasons for taking an automated BI approach. This time I’m going to zoom in on why many still resist change, even though it's so invaluable...
1. Understand the reasons you’re changing
If it ain’t broke, don’t fix it – right? While that’s often true, it rarely applies to digital systems, where the rapid pace of change ties interoperability to the newest technological iteration.
MT Højgaard’s existing data warehouse was built on previous versions of Microsoft SQL server but because we wanted to use the improved functionality of MSSQL 2016 and Power BI, we were faced with a migration task. Since we knew we had a tough job ahead in migrating the data platform, we took that moment to also make some wise decisions about what we wanted from any future framework.
Heading into our period of change, we already knew what we wanted coming out of it. For example, we didn’t want to be dependent on external IT technicians. So we hadn’t started but already, we knew why we were changing.
2. Explain the benefits to EVERYONE
With our old system, different departments accessed business data in a very traditional way. They pulled filed reports, accounts statements and so on and used this static information resulting in out-of-date decisions about the future. Another issue was a binary authorisation model to this data – either you had full access or no access at all.
We mitigated this with a more granular authorisation model and by using TimeXtender as our data warehouse was only the starting point of our change. We then had to educate people about its benefits. Business data is no longer filed away and backward looking. By accessing up-to-date information that’s relevant to them, departments can take a far more analytical approach to data and, in doing so, make better informed, smarter decisions. But they need to know that.
3. Make sure your new system is an improvement
We want data democracy and we aim for as much self-service as possible but at the same time, we also want MT Højgaard’s valuable business data to be highly governed, since some data just isn’t meant to be consumed by everyone. Giving the correct access to the right people and keeping a clear picture of who is creating and using what requires both insights into DWH production and data access.
Our goal is to offer everyone the data access they need for their decision making without either flooding them with a confusing amount of information or handing out all data to everyone. So a governed access to data, with a focus on data quality from a ‘single source of the truth’ is a clear improvement.
By taking this approach, we’re starting to offer departments direct access to only the data they need. In this way, they can build their own reports and do their own analysis while we maintain a company-wide standardised reporting package. We are not there yet but the direction is set and we are moving towards our goal.
4. Set achievable goals that prove your case for change
My team of BI specialists was initially skeptical about the decision to use TimeXtender’s Discovery Hub®. Would they lose their freedom to develop a solution the way they thought was best? What about flexibility? Could it handle complex ETL or other programming languages? Would it be able to quickly adapt to new Microsoft technologies.
Convincing a skilled BI specialist who has spent the last ten years coding SQL or SSIS or whatnot that they should do something else is a difficult sell. My approach was to be very insistent on trying it out on a case-by-case basis, looking at the decision from a business perspective. We started with one use case to explore both the tool and its framework and when that worked and everybody had a chance to look at the possibilities, we moved on to the next equally challenging use case.
Little by little, the team realised that Discovery Hub® is not a black box tool but that yes, you can add customised SSIS packages, views, user defined functions, scripts and stored procedures to meet our unique requirements.
Keeping the focus on the big picture benefits has been a good way to go. By proving automation, documentation and flexibility, by showing there’s no need to be entirely dependent on a specific person to solve a problem, our company is realising that this is a framework with business benefits.
5. Change is hard – stick with it
Introducing a DWH automation platform isn't just an IT issue, it's a change that affects the way the whole organisation works. Instead of referring to static reports, produced periodically by IT, individual departments can now work in real time by looking up the data they need, analysing it and then building their own reports. This is a completely different way of doing things, so it’s no surprise that it takes some time to get used to. I think being used to the old way is 90% of why adoption has been slower than it could have been.
However, the more valuable knowledge each part of the business gets, the more they start to ask about what else they could do. We’re here to make them curious about these possibilities and while we’re not there yet, we can certainly see a cultural shift occurring.
We’re approaching data warehousing as a journey rather than a destination, never being content to stop, always moving forwards. Part of this journey is combining our old way of doing things – our existing data warehouse – with our new way. And that’s what I’ll be talking about next time.
Mikkel Kvist is MT Højgaard’s Head of the Group Business Intelligence Competency Centre. The MT Højgaard Group a major construction and civil engineering player in the Nordic countries, employing 4,200 staff and generating an annual revenue of DKK 6.8bn.