We often get asked by our clients to differentiate the Kinaesis DataOps methodology to a standard Data Management Methodology. Many organisations are implementing standard methodologies and are not seeing the benefits. The key differentiator to me between the methodologies is that DataOps is based around practical actions that over time add up to deliver results greater than the sum of the parts. If delivered correctly across the 6 pillars of DataOps they help you to transform the organisation to be data driven. The key to this is the integration of people and process that deliver real business outcomes avoiding paper exercises.
On my consulting travels I find that many industry data management methodologies layout the theory around implementing a Data Dictionary for example and this is taken as a mandate to deliver a dictionary as a business outcome. Within DataOps a dictionary is not a business outcome, it is part of the deliverables that are part of the methodology and an accelerator of delivering a business outcome. This is a subtle difference, but one which leads to the effort of the Data Dictionary being part of the business process and not an additional tax on the strained budgets. Within the methodology it is produced as an asset within the project and for the future to make subsequent projects easier.
Another difference is that in standard data management approaches the methods are quite prescribed and consistent across all different use cases. The nature of the DataOps methodology is that it fits the approach to the problem being solved. For example some data management problems are highly model driven like credit scoring, customer propensity, capital calculations, etc. Other problems can be more reporting and analytics. Each of these require a different focus and a different emphasis and sometimes a different operating model. Through the iterative approach then there is freedom within the methodology to achieve this.
Many data management approaches prescribe an approach that tries to encapsulate all of the data within an organisation. This is a noble cause, but it is a large impediment to making progress. Firstly in many cases we have found that only 15-20% of the legacy data is ever required to meet existing business cases. Secondly we find that the shape of data is highly dependent on the use case being implemented and because you do not know all the use cases and future use cases it is not pragmatic to do this. By being able to measure this usage through instrumentation and driving them through use cases then the data management problems can be simplified to achievable outcomes in short periods that can form the foundations of the data management strategy for a business area to be leveraged.
There are many other differentiators within the DataOps methodology however all of them start with the principles that anything you do needs to be implemented and pervasive. The methodology builds its strength from tying business benefit to the process and builds from this. The goal is to deliver value early and often and to leverage the benefits to build momentum to deliver more benefits over time. Once integrated into the operating model then the approach builds and transforms the culture from the ground up. This delivers great benefits to the organisation where many data management methodologies start with great promise and then struggle to gain support when the size and the complexity of the task start to become apparent.
If you would like to find out more about the Differences of DataOps then please do not hesitate to contact me.