How to ensure the success of your company’s data programme

| Updated on October 11, 2019 Published on October 11, 2019

The complexity of a data platform mainly arises due to the multiplicity in collection and handling of data by different departments of an enterprise   -  iStockphoto

It is important to identify the non-technical factors, such as involvement of all stakeholders and flexibility in design and function, which go into the building of an enterprise-wide data platform

The enterprise data platform continues to be the holy grail for many organisations. Let’s face it, data projects, that too enterprise-wide, are complex — and with complexity, the risk of failure is higher when unmanaged. There are several technology-related factors that one should consider to mitigate the risks. Equally important are the non-technical elements, that strongly influence the outcome of data projects. Let us look at those a bit more closely:

Selling the agenda, continuously and consistently within the organisation, is imperative. Whether the primary intent is to create a platform that provides a single version of the truth or that helps the various business units to extract value out of the data corpus — or both — it is important to understand that such a transformation will take time. Therefore, the stakeholders must stay consistently true to the agenda of the programme and its intents. To that extent, the sponsor needs to ensure that there is no dilution to the goals (which normally happens during long-duration projects).

The involvement of business stakeholders until the completion of the project (and beyond) is non-negotiable. At the end of the day, establishing a solid data platform directly benefits businesses. It allows for more informed decision-making, better insights into market situations and opens up plenty of possibilities to reimagine the business models, leading to innovation. If the project is left to IT alone, the complex dimensions of the business will not be factored and will be constrained within the purviews of data engineers and scientists — whether it is about sourcing the right data, creating the right models or algorithms, or making the right assumptions. Enterprises that stay true to reinforcing alignment between the business needs and technology functions around data, tend to succeed in creating value differentiators that matter.

‘Start small, grow big’ has almost become a cliché that applies to all complex large-scale transformations. Data transformation is not an exception to that rule. Starting with a few high-impact digital processes (if there are digital user journeys established, then target those) and fortifying them with insights derived through the data platform is the easiest way to establish and build on success. Within the contours of the journey, one can define which dataset is needed, what transformation it has to go through, how to organise the data/insights for easy consumption — and then build the necessary AI/ML models to derive insights. Such a small, step-based approach not only validates the design, but also brings tremendous confidence on the programme itself.

Get past the “We have it already” syndrome. In many instances, the common use cases (customer 360, for example) have already been identified and implemented, however rudimentary they may be. Chances are that every company and its peers may be following applied thinking as well. Settling for what is already there, without continuous review and improvement, may soon render from data models results that are stale and ineffective. Data models need to continuously evolve, with strong feedback that indicates the accuracy levels of the insights. An organisation that is open to a relook at existing practices and assets, and improves upon the programme by critically analysing the data dimensions to bring far better accuracy and currency to insights, will continue to drive the value differentiation among its peers.

Be agile — even in the design of the platform. While establishing an enterprise-wide data platform, data transformation touches pretty much all departments within an organisation. The complexity of the platform mainly arises due to the multiplicity of the way different departments collect, store and handle data. It is important to have all of the details studied well to arrive at a technically feasible design. Many a time, despite a thorough, low-level design, there are bound to be surprises along the way. Bringing modularity in the design will allow us to accommodate necessary changes, yet preserve the architectural principles put in place to achieve the desired business results. During implementation (generally, this is where one discovers the surprises), expect and be ready to introduce alternatives in design, tools, methods, and processes that will fit everyone’s needs better.

Almost all of the points mentioned appear to be of common sense: they are. Yet, there are scores of companies that embark on a data platform project and get frustrated along the way because of the lack of sufficient benefit realisation. Apart from ensuring that the organisation has the right skills and key technology choices, the above factors must be seriously considered, if not adopted. This will make successful data programmes a reality.

The writer is Chief Technology Officer, IBM India/South Asia

Follow us on Telegram, Facebook, Twitter, Instagram, YouTube and Linkedin. You can also download our Android App or IOS App.

Published on October 11, 2019
This article is closed for comments.
Please Email the Editor