With marketplaces rapidly evolving, the data required to support decision making processes are changing at an equally breakneck pace. Is your technology delivery pipeline equipped to support the data demands of your business stakeholders? Often, there is a disconnect between an organization’s data strategy and the tactical delivery capabilities that support it. Central themes of usability, consistency, and accessibility are usually at the core of a data strategy’s value proposition to its customers. Frustration erupts when big, bold ideas aligned to these concepts cannot be translated into tactical and tangible delivery mechanics. As your organization’s analytical capabilities mature over time, and your data strategy evolves, your technology delivery strategy needs to evolve in lock-step.
If you find that your organization’s analytical appetite is outpacing its ability to serve data up for consumption, then you are not alone. Does this mean that your organization is doomed to lag behind in a global marketplace that’s moving at light speed? Of course not! What this means is that change is inevitable. Your organization can adapt by embracing a continuous delivery, just-in-time model that places emphasis on speed to market, builds tight guard rails around quality expectations and focuses on flexibility in order to meet today’s business needs. Pairing a Kanban work management process with a continuous delivery model provides business stakeholders with the means to iteratively evaluate solutions and pivot if needed, ultimately ensuring that the solution delivered is both effective and sustainable.
You might ask, can this really work in the complex world of BI/DW delivery? The short answer is: yes. The age of multi-month BI project delivery timelines is going the way of the flip phone. Today’s business stakeholders are dealing with more disruption and ambiguity than ever before, and as a result they need the right data, at the right time, in order to navigate through these challenges. SEI clients have undergone similar transformations, and as such, there are many lessons learned that our consultants leverage to ensure client engagement success. While this is by no means comprehensive, it has proven to be an effective primer for the technology, business, and organizational governance changes required. Here’s how to begin adapting your technology delivery lifecycle to better align with this shift in customer need and expectation.
First, take stock of where you are today.
- Evaluate your existing technology organizational alignment, challenges and risks
- Who are your primary business stakeholders, and what are their primary concerns regarding technology delivery capabilities?
- How is work funded and prioritized in your work queue today?
- How often do you meet with your business stakeholders to survey their needs?
- What’s the lead time for new data requests, from inception to production release? What should it be?
- What dependencies exist in the data supply chain that create bottlenecks and unnecessary delivery risk?
- How does your organization support data exploration tools and strategies?
- What level of QA automation exists?
- Evaluate your code change management processes
- What controls exist in order to ensure that your code base maintenance is efficient and effective?
- Do code dependencies create risk in your release management activities?
- Evaluate your release management practices
- How do you communicate release content to your business stakeholders today?
- What criteria are utilized to evaluate release readiness?
- How is the release pipeline transparent to the team?
Now, go grab some low hanging fruit to get this transformation off the ground!
- Find a small, well organized technology team to use as a proof of concept. Ideally this team is already working within an agile framework, like Scrum, making the change more of an evolution than a revolution.
- Schedule, at a minimum, weekly work queue discussions with your key business stakeholders. Focus on business value and transparency, let them drive the priority of work and negotiate the order of operation based on capacity, complexity, and risk. Find the quick wins!
- Embrace the ‘Citizen Data Scientist.’ These folks are your analytical soldiers; armed with a traditional BI skill set. They may not be able to write deep learning algorithms, but they can use traditional BI, ETL, and data visualization tools to delivery phenomenal insights that will drive the business forward.
- Begin a weekly release cadence for completed code. Implement a release ‘QA done’ threshold, usually the week before the release, for all content. If the work hasn’t been signed off on by QA by this day, then it gets deferred to a later release.
- Create small pieces of work with an iterative release schedule in mind. Said another way: don’t try to boil the ocean. Provide business stakeholders with a continuous flow of content, enabling them to rapidly evaluate and pivot requirements as needed.
- Establish baseline data quality expectations with your business stakeholders. Make sure that pristine data doesn’t become the enemy of good data, when good data is perfectly suitable for the task at hand.
- Get your application environments in order. In order for a continuous delivery model to be effective, include a UAT environment with as close to production data as possible, as well as a golden environment for staging ready-for-release code.
While this process isn’t foolproof, and in some cases the learning curve can be steep, the improved transparency and focus on business value creation almost always results in stronger technology/business partnership and improved tactical collaboration. The key is to shift the technology delivery strategy away from doing things the right way and towards doing the right thing. The end result is a culture shift, and your technology organization will never look back.