Get a clear view of business performance
According to a 2019 Harvard Business Review (HBR) Pulse Survey paper, the transformative insights powered by cloud, data, and AI “are enabling organizations to fundamentally re-imagine how they engage customers, transform products, optimize operations, and enable employees.” Elsewhere, Gartner talks about moving from “data, to insight, to action, to impact”.
This sounds good in theory… but how can you achieve it in practice? Most organisations have a complex patchwork of on-premises and cloud systems, in which the same information is represented multiple times from different angles. To get a single version of the truth, you need:
- a centralised repository of high-quality, well-governed data delivered in real time
- data that is secure and complies with privacy standards
- analytics and reporting that are easily accessible to business users.
This repository is a data warehouse – a central store of data, extracted and transformed from multiple data models, definitions and formats in siloed operational systems.
A data warehouse enables you to leave your operational systems doing what they do best, and maintain a separate store of data for consistent analysis and reporting.
Again, that’s the theory. In practice, data warehousing projects have a legendary reputation for missed deadlines and busted budgets.
Why data warehouses are hard to do
- Enterprises change: insight is a moving target, so you can’t set a fixed end goal
- Unknown future goals translate into major risk
- Enterprises tend to lack internal expertise in data warehousing
- Data-quality issues often rear their heads, causing costly delays and disappointing outcomes
- Project creep further increases costs, delays delivery and moves the goalposts.
Six golden rules
Northdoor believes that – when done right – data warehousing can deliver exceptional business benefits at high speed. Here are our six steps to success:
1. Lay the right foundations before you start: discover and analyse your existing data resources; classify your data; identify potential data-quality issues; map out the target transformations.
2. But don’t spend too long on the groundwork: things are going to change, so think about achieving a “just enough” position that will enable quick wins rather than trying to solve every challenge at the outset.
3. Work with experienced consultants who have a track record of successful delivery of large-scale data warehouse projects.
4. Let software do the heavy lifting: take advantage of best-of-breed solutions for data discovery, classification, validation and reconciliation.
5. Plan and execute an agile approach, ensuring that your project adapts seamlessly and cost-effectively to inevitable changes in scope and target outcomes.
6. Design organisational structures and processes to maintain and develop the data warehouse as an ongoing source of business value.