Information Availability and the case for a Analytics PMO
Many a times when engaging in analytics assignments we are called upon to manage the vagaries of data. This has nothing to do with the competencies of the analytics team, but for certain upstream decisions that may have been taken in a certain time shift that lead to the quality, granularity, recency or frequency of the data that were made available to us.
When long term Strategic Information Management plans are made it is key to ensure that the plan for future data capture as well based on the need, growth and direction of the business. There are seldom situations where server space and analytical software required for doing analysis are not available; however the same can’t be always said about the availability of data. In the analytical world the simple truth is that for more accurate and precise insights you need large volume of data and interesting variable set that can be causally linked to outcome based on initial hypothesis and validation thereof.
However, many a times as you understand an analytics business problem and try to frame a solution you get strapped with data that is either full of holes, lacks historical context or misses out on variables that could lead to interesting interactions. Not too long ago we were called upon to analyze drivers for inbound call volumes to understand whether the client could do anything internally to reduce the inbound calls. What we really needed was data on multi-channel customer interaction e.g. billing, payment, customer service and product- service’s enrollment at a granular level with about six months of history which we would then use to correlate and find events that lead to higher concentrations of call. The data was simply not there, what we got was smattering of interaction information padded before and after calls using some kind of a “time between event” logic which was neither accurate nor extensive. Some exploration on the data situation revealed that creating new downloads from back end system is a expensive process and the data layout that they were giving us had been there for considerable period of time and this was not the first time that the business had their “aw!! Oh!!” moments.
As someone who has been working in the information management field , I feel that these problems can be resolved easy enough if certain amount of planning is done as per the operating rhythm of the business. As CIO organizations come under considerable cost pressure they are relegated to stay focused on the enterprise system requiring upgrades and execution efficiency. The data and analytics systems are typically business prerogative even though there might be a large reliance on the technology organization in terms of infrastructure. Since infrastructure doesn’t usually run out high level planning around business information demand is not obviated, however a deeper level of planning and preparedness goes missing from both parties.
Based on my previous experience of successful organizations the following management practices have been helpful:
1. As part of long term, yearly and quarterly business planning, identify areas in business that will come under stress or will need to support strategic/ near term goals
2. Evaluate at a high level what nature of analytics may need to be used to optimize the business areas for success:
a. Use experience from previous situations
b. Look at new opportunities e.g. Big Data
3. Drive action plan to align correct resources within the business- Human Resources or any other kind
4. Identify data need and engage with the Information Management Groups and sensitize around the need and availability of new data
a. Plan, set time lines and agree on when new data facets need to become available
c. Point person to work on the data
5. Formalize a matrix organization to drive through the imperatives- Business/ IT vendors, Business Leaders, IT managers, Business Managers et al
6. Set up a execution rhythm with regular report out
In order to make the analytics adoption process more proactive than reactive it is helpful to set up a Analytics PMO who would be responsible for driving out new plans, watch out for emerging trends and advise management on the next dos in terms of short and long term opportunities.
Some planning and a optimized organization may not necessarily drastically increase operations cost, but help in recuperating the ROI of the analytics and information management asset. Until and unless the operating model is that of a finely tuned e-business where data foot print is captured as a default or a highly regulated business where information capture is mandated, it is easy to get blind sighted by information need.
In order to get around this planning is very important and in order to establish the rhythm in analytics and information syndication setting up an Analytics PMO is helpful.
