Most businesses are alive to the possibilities that data exploration and analysis can offer their organisations. Too often however, big data projects carried out in large organisations can become mired in internal wrangling about technology and data management, diminishing any real chances of a speedy return on investment.
As with many other areas of business, true success requires a range of skills and input from within the organisation, along with the specialist skills only a multi-disciplined professional services team can bring based on their industry experience.
Rather than simply buying technology tools and expecting IT to generate value from it, specialist teams work consultatively with organisations to ensure they understand what their wider industry or unique business challenges and opportunities are, before bringing a tried-and-tested methodology which can apply to each project in order to guarantee a successful outcome.
In the early phase of this approach, dialogue will focus on the nature of the business problem and how it can be addressed, rather than putting the emphasis on the technical demands of the project.
This will include a statement of the business problem, followed by identification of the analytics required, the data needed for it and the action that is required to close that loop and deliver results.
> See also: The era of big data won't materialise without fast data
Having worked across a broad range of industry and enterprise-specific challenges and opportunities, good professional services partners will recognise the fact that, while data visualisation and analysis are interesting in their own right, unless an organisation can take action on the information, it will not generate any return on investment.
Building a business impact model
To define the areas of most potential value to the business, the first step in a big data project should comprise a business-focused workshop of between one and three days. This should involve specialists from each appropriate industry sector and representatives from both IT and business within the client organisation.
The objective of these essential sessions should be used to tease out problems and challenges and to set agreed business priorities. Part of this – perhaps a day – could be devoted to taking a detailed look on a more technical basis to understand data and resource constraints.
In the case of an online retailer, for example, abandoned baskets may be a big problem. Here, the team would focus on what data is required and what is available to the organisation that could help identify the primary reasons for abandoned sales and enable them to create an appropriate solution.
In this case, it could be establishing the value of abandoned baskets themselves before working out the value to the business of improving the percentage moving to the checkout.
As a reference, this information can be pulled together and presented in a business improvement matrix, which establishes the challenge addressed and the effects that will flow from solving a problem.
To focus hearts and minds in the organisation, in addition to setting out the challenge and solution, a clear ROI statement would explicitly quantify the value that improving checkouts by a set percentage would actually be worth.
Again, this is not about using technology for technology’s stake; In exploring this and other challenges and opportunities with a clear and quantifiable emphasis on value and ROI, organisations will find themselves far better equipped to prioritise strategy and focus of big data projects and achieve true business value.
Agile, fail-fast exploration
After bringing in data engineers to analyse and structure the data for analytics via a programme of business-focused projects, the next logical phase of the ROI-driven big data project involves testing the idea against the data.
This is likely to include running a hackathon or datathon to test out these ideas with staff over three-to-five days and will typically involve business analysts who are familiar with the tools, weeding out ideas that don’t work or where the data has poor quality.
From a technology perspective, a powerful, ready-to-run platform can play an important role at this stage in the project by allowing data to be stored in its raw form ready for use with analytical tools. Ideas can then be tested in an agile way, using a fail-fast approach to solving problems and capitalising on potential opportunities.
When testing these various concepts, big data appliances and analytical platforms which allow businesses to carry out analytics at fast speeds can also provide insightful visualisations and contribute to important conclusions.
As the project nears completion, the analytics phase will gather pace and be conducted in agile bite-sized chunks known as sprints, using strong implementation methodologies to take the solutions reached to market in the best and fastest possible way.
Clear focus and expert support
Rather than deciding which IT boxes and tools to install, smart organisations are increasingly realising that getting fast ROI on big data projects is not about simply buying-in to the latest technologies or giving specialists an open mandate but rather, having a clear focus on solving business problems.
Bringing their understanding of common industry challenges and listening to business-specific opportunities, a multi-skilled team will help steer the project towards the right data and analytics, use developed methodologies and cultivate an interactive, agile approach.
> See also: Forget big data: what businesses really want is simple and easy
In a field where there are many new technologies – not all of which are mature nor easy to use – making the most of the skills and experience big data service partners and industry experts bring is what really helps to separate nuggets from the fool’s gold.
Flying high above their competitors, these organisations will quickly hone in on the best approach to deliver value fast, reacting to insight and using their new-found agility to change direction swiftly and achieve the right business objectives.
Sourced from Vic Winch, director, Big Data Centre of Excellence, Teradata UK