New Year’s Resolutions all Data Management Professionals Should Make
The Bulb: December 2020
Written by: Mike Lampa and Julie Burroughs, GDM Advisors
As we reflect on the trials and tribulations of 2020 and look towards the prospects and promise of 2021, leadership has a new opportunity to differentiate their enterprise through an effective and efficient analytics-driven transformation journey. Here are our thoughts for your consideration as your data and analytics 2021 resolutions…
Level up my delivery teams to be true lean agile.
If you want to innovate and deliver at scale, you need to adopt a lean agile product management operating mindset for your Analytics program. Yet most organizations tell us they are lean agile and after assessing their operating model against the seven core competencies of Business Agility (see November Bulb article) we see evidence of the Team & Technical Agility core competency, but limited or no adoption of Lean Portfolio Management, Agile Product Delivery and Organization Agility. These four core competencies are key to continuously delivering quality innovative analytic products at scale and more quickly than ever before. Adoption of the continuous innovation in cloud technologies further speeds time to market with value generating analytic products. The new game in town is co-development with these technology vendors while taking advantage of their IP and resources. If you aren’t exploring scaling your agile practice, you will likely continue to not meet the needs of your enterprise.
Augment standardized dashboards with data science augmented operational insights.
In the age of self-service analytics, the business needs more than standardized dashboards and visualizations. Business knowledge workers need to answer different questions and discover new insights without preconceived constraints yet, the average adoption rate of those dashboards is around 21% according to Gartner. This is a lot of wasted time and money. The latest breed of analytics reporting tools offer “Google like” search capabilities to ask a question, get the right answer AND receive machine learning enabled insights related to the question could very well show the cause for the answer results. We say “analytics for all” and tools like ThoughtSpot take that challenge on. This is a new breed combining search and AI-driven insights which needs to be considered as part of differentiating through analytics.
Stop delivering monolithic data governance.
Data Governance programs have a long and disappointing history of costing millions and taking years to implement and generate ZERO value. Let’s face it governance considerations are simply additional requirements that define a quality Data Product the business needs in order to build Analytic Products. Stop the madness. Adopt the Universal Principles of Data Ethics. Take a lean-agile product management mindset when implementing governance as you build data products. Focus on critical and common data elements that make up those data products and which will further the achievement of strategic objectives. Include governance, quality and regulator team members on your Agile Teams and start generating a continuous flow of value through improved data security controls, regulatory compliance and quality. This , taking years It’s OK, you can let go. You can do it. Sometimes you just have to do a reboot. Sunk cost fallacy comes into play here but as you take a hard look at these programs you see where and their annual cost to continue to support them, is not giving the results. Modern technologies and approaches to governing data will contribute significantly to the value produced by your data products.
Help my organization become more data literate.
Get them collaborating, ideating, discovering even arguing for a point of view armed with data. In an Accenture study, only 32% of companies reported they are able to realize tangible and measurable value from their data, partly due to poor Data Literacy. Data Literacy is a foundational organizational competency that will lead to increased data democratization and monetization. Perform Data Literacy Health Radar on a regular cadence to continuously gauge your current level and desired level of mastery and proficiency across the eighteen key traits of Data Literacy. With those gaps in hand, design and execute on a Data Literacy implementation journey. This isn’t a daunting activity. Consider an automated data catalog to springboard your literacy program. The benefits are never ending.
Automate my data and machine learning operations – keep my head above water.
Come on man. The data tsunami isn’t coming, it has arrived. And the increasing demand for the data, NOW, is off the charts. DataOps & Machine Learning Ops (MLOps), are the analytics domain adaptation of DevOps, a proven software delivery discipline. The Continuous Delivery Pipeline is the foundation that enables enterprises to release Analytic and Data Product value, in whole or in part, at the right time and when it’s needed. DataOps has a primary focus on delivering data pipelines that produce the Data Products which in turn enable the creation and delivery of Analytic Products. This is the Analytic Product demand-driven model for Data Products. Data that is fit for purpose for analysis and the supporting services to easily access the data for analysis. MLOps has a primary focus on delivering additional data pipelines and persisted data that is fit for purpose for Machine Learning (ML) analytic models, as well as the ML model code and execution telemetry metadata to enable ML model efficacy monitoring and alert triggering to refine those ML analytic models. Modern technologies for DataOps and MLOps automates the build of the pipeline code, the testing of the code, the management of the code, the bundling of the code into deployable builds, the enablement of provisioning infrastructure as code and the monitoring of the telemetry of the deployed pipeline’s execution cycles.
So raise your glass, toast the work you did in 2020 and make the resolution to modernize your analytics program.