Over the last few years I have immersed myself into the exciting and at times daunting world of Big Data Analytics and Solutions. My teams and direct reports would probably classify me as a "super-user" of data analysis in assessing business performance as well as research and preparation for advice to clients on business solutions and benefits.
What I have come to learn and appreciate now is the exponential growth in the capacity and speed at which data is now able to be extracted, integrated, visualised and distributed.
Interestingly the concepts around big data management have actually been around for a little while with Doug Laney (Gartner) in 2001 articulating what has come to be known as the core 3 V's of Big Data Management being; Volume, Velocity and Variety. Various additions and extensions such as Veracity, Variability and Value have been added by others since, but the first three remain at the core.
Since 2001 we have seen and experienced rapid growth in the processing power of technology, greatly reduced storage costs and space requirements. This has also meant an observable shift in mindsets and approach toward rapid, agile technology development and delivery. The growth and usage of mobile, social media, digital data technologies now also provides additional and valuable data elements able to be captured, integrated and analysed.
“ Advanced Data Analytics has been happening for some time in the internet / online world with larger players such as Facebook, Twitter, Google, Apple etc; the next wave comes as that thinking, approach and technology are now increasingly applied within other corporate and even smaller organisational enterprises”
Improving practical applications of Big Data
From my research and observation in the world of Big Data analytics I see two main applications of the data in what I call "Reactive" and "Predictive" uses; and the two also go together.
For example, the ability to capture vast amounts of complex data from across multiple countries, sites, equipment and devices on a virtual real time basis and have that customised, analysed and visualised to the needs of different organisational users and functions. This significantly improves important “reactive” decision making as well as better decisions from more timely data and better insights.
The second mainstream application being the "predictive" analytics capability as data is captured and analysed and patterns are observed. Intelligent algorithms then applied to form predictive suggestions on what is potentially going to happen and what may be needed next. As more and more data, patterns and trends are analysed, the predictive algorithm capability becomes even greater.
From a customer service and sales management perspective this becomes very useful and exciting in providing timely suggestions on what the "next conversation" with a customer should perhaps be. Now there also exists the capability to deliver a multi channel approach of marketing and sales through mobile, email, written and personal contact from such insights. The benefit being that targeting customers with a higher propensity to purchase is a far more effective and productive application of resources.
Within Financial Services this has become very attractive as analytics can now be applied on internal customer data such as transactional behaviour, credit card usage and location, assets and liabilities position, life stage position, internal and external website hits, social media interests and commentary.
Capturing, integrating, analysing and distributing all of that data in a secure, reliable, usable format (adaptable to needs of different users) is now a real and urgent challenge for many organisations.
If I earned a dollar for every conversation I have had with many people across many companies, over many years; seeking to deliver a "timely, reliable, holistic, single customer view" I would be very wealthy! Many companies have and continue to spend many millions in the search for such a solution. However as more data is acquired and stored on customers (through traditional database structures) then the challenge just seems to get harder. Perhaps I should just settle for a dollar earned for every duplicate customer entry that exists across systems in most typical corporate organisations.
In addition to the customer sales and service insights opportunities, advances in Big Data management also provide great benefit in both reactive and predictive risks management. Having timely data provided to management (in a form of most benefit) to highlight adverse operational events or trends is of significant value and benefit. Again as these events and trends are learned and applied in the “Big Data” analytical algorithms the reactive and predictive capability is continually improved. This provides better and faster decision making in taking necessary or precautionary actions (reactions) as well as better prediction of potential downstream risk consequences.
The way we have previously and traditionally used and applied data insights is rapidly changing and that requires both an open mind and acceptance of new ways of working with Big Data inputs and outputs. As we are already observing and experiencing, the reactive and predictive capability of technology is in many areas ahead of the reaction time and accuracy that a human is capable of. We see much example of this in the car industry with safety advances eg, automatic braking, external information recognition, GPS navigation leading now to driver-less cars.
Interestingly, in various discussions on Big Data I have also experienced quite a range of reactions. Some see great insights, opportunities and benefits; others I speak to highlight concerns of personal intrusion, privacy and a loss of human judgement and control. I can see and appreciate both points of view.
Big Data - all very interesting, but how is the R.O.I. looking?
My personal career experience has primarily been in the customer facing domains of business management however having also lead a number of major projects I have been engaged in many interesting discussions with technology colleagues in translating business needs and opportunities into technical solution requirements. The best results have of course always come from a shared understanding and appreciation of business and technology perspectives with respect to the ultimate business problem / goal to be solved.
In my exploration of the Big Data world I have now had my “Hadoop” education (check it out if you have no idea what I am talking about) and even found myself getting into the debates on SQL vs NoSQL. It really is quite a shift in how we capture, integrate, analyse and distribute data.
The reality remains that great advances in the data technology capability are still only as useful and effective as the application and utilisation of that information across the organisation. Hence my title for this article that transforming [Big Data] ideas and opportunities into positive beneficial outcomes really is an organisational wide journey. It still requires a collaborative understanding and effort across multiple organisational functions to really apply and capture the opportunities that now exist from the Big Data capabilities.
With lots of available data there is an even greater need to be selective on which data is most important and that which can or should be disregarded. Otherwise a very real risk arises of what I call " information paralysis".
Then there is the consideration of having captured and analysed relevant data, how to get it out to those who will most benefit from it and perhaps even more importantly managing and measuring what people do with it. Some people may prefer to have the information provided more broadly for their own interpretation and application; others may want more targeted analytics and specific suggestions.
The good news is that current technologies provide the ability to meet all of the above requirements. The more important point however, as suggested by many other “Big Data” solution experts:
“Don’t get hung up on the technology capability or complexity, the primary requirement is to be very clear on what the business problem or opportunity is that you are seeking to solve - and keep that front and centre of all decision making”.
Big Data technology is also about better, faster cheaper delivery!
Perhaps not as immediately obvious, one of the key benefits I speak about with clients on Big Data projects is the speed and agility in which new smarter technologies are able to deliver the data solutions.
New fast agile start-up technology companies are now delivering proof of concepts and projects in weeks rather than months (or dare I say years). Given the project delivery cost burn rate that exists for many typical larger organisations, often as a result of legacy data warehouse structures, these new proprietary Big Data platforms, technologies and algorithms have the ability to substantially reduce delivery time and costs on these projects.
The benefits from this are then two fold. The actual project delivery costs are substantially reduced but perhaps more importantly; solutions are able to be delivered to market quicker and end benefits are realised sooner.
This in my experience also requires open consideration to the technology resources that should be utilised from within the internal organisation and increasingly a delivery model of collaboration with external specialist big data technology providers on both thought leadership and proprietary solution capabilities.
Again I would hasten to point out that even if the “Big Data” technology teams now have the ability to accelerate the project delivery, there remains the key requirements for:
Clear definition of business end goals and requirements (and keeping that focus throughout the project),Development of product, service, marketing, operational support materials to support the data insights and resources, processes and procedures to implement the opportunities or actions suggested by the data outputs
Achieving organisational wide understanding and alignment on Big Data projects and opportunities
Considering all the above I would like to suggest and offer a simple but effective assessment framework that can be easily applied across both technology and business teams in considering opportunities and execution of “Big Data” projects.
They are as follows:
S.P.O.T. and E.I.A.V.A.R. (which for simplicity we can call Eva)
Let me now explain what these mean.
S.P.O.T. is an acronym for Sales, Product, Operations & Technology.
These being the major organisational functions and groups that will typically be involved in successful delivery of Big Data solution opportunities.
Sales – covering people / teams in customer facing functions including business development, service management, account management, customer relationship management etc.
Product – covering people / teams that design, manufacture, manage and providing marketing support for physical products and services to be delivered to customers
Operations – covering people / teams that support, process, enable, equip, measure and monitor the delivery of solutions to customers via the "Product" and “Sales” teams.
Technology – covering people / teams that design, develop, manage and distribute via hardware and software solutions the “Big Data” inputs and outputs to the Operations, Product and Sales teams.
E.I.A.V.A.R. is an acronym for Extraction, Integration, Analysis, Visualisation, Application and Results
Extraction - what information and data you would like to see and where does that data currently reside in its source form?
Integration – how and where is information from different sources and systems currently stored and configured and where should it ultimately reside – as a central repository?
Analysis – what is the the ultimate business problem or goal wanting to be solved and what combination of data elements can and should be applied, cross referenced, compared etc to produce the predictive insights?
Visualisation - how is the data to be assembled, displayed and distributed to the locations, needs and preferences of different users of the data?
Ideally in a form in which users can further tailor and customise to suit specific needs and situations.
Application – how will the data provided be utilised and applied through operational processes toward business and customer needs and opportunities?
This for example may need to be supported by sales training, new processes, marketing campaigns and materials or specific management and escalation policies and processes for operational management applications.
Results – what are the specific, tangible end results expected through the investment in Big Data capabilities and solutions and how will those be measured?
This should ideally include a combination of Customer, Staff, Operational, Financial / Shareholder and Stakeholder performance assessment metrics – tied back to the originating business strategy and project delivery business case goals.
..and finally here's how to use it:
Across the top of a page (or whiteboard) write S P O T (giving a column for each). Then down the left hand side of the page write E I A V A R (giving a row for each) - thereby setting up a matrix of boxes in which you can write a few points.
At the commencement of your next Big Data project have representatives from each of the S.P.O.T. teams each consider all of the E.I.A.V.A.R. elements for the project. Having those people in the same room at the same is ideal however this can be undertaken within separate groups. Shared perspectives and the “wisdom of crowds” on any issue or opportunity is a powerful force and a key element in my experience to the success of an organisational project.
Once the key thoughts and comments are captured in the matrix page per above then give a copy to all S.P.O.T. teams to review. Whilst I fully acknowledge that many other elements and factors need to come together for a successful strategy or project delivery, this I believe will set you up well for a continued collaborative discussion and success on your Big Data projects. Getting ideas to outcomes really is an organisational wide journey!
コメント