The following post represents the second installment of my data strategy presentation.
Once a company has recognized the tangible value of its data, the CEO will assign the role to a direct report making this initiative his or her sole responsibility. As noted in my first post, managing data as an asset is so important that it requires direct senior management responsibility and should not be delegated to the head of marketing & sales or IT (in each or any case, both bias and conflict with their ‘real’ responsibilities will prevent the program's benefits accruing to the entire business). In addition, the ability to break down silos, confront the assumed internal politics and impose a solution will be greatly diminished if the executive in charge already has other functional responsibilities. In my view, if this is the approach of the CEO, the initiative will fail. By way of example, the Chief Data Officer at Dun & Bradstreet sits at the highest levels of the company in recognition of the importance of data to that organization and is evidence of the strategic importance of that data.
The first task of the “Chief Data Officer” (CDO) and their team would be to conduct a business-wide data audit to determine where data is being collected, both through internal efforts as well as externally by customers, vendors and other partners. This is an important first step and should be as expansive and specific as possible. Things to look for include not only the data sources themselves but their quality (perceived or explicit), frequency of update or creation, source/owner, format, whether the data conforms to any particular standard and whether the data is ever validated against this standard. Reviewing standards might also include not just the data’s technical framework but also whether it is subject to normalization – to a specific categorization method such as BISAC, for example. I once undertook an effort like this at a large educational publisher and the process took about six weeks and required site visits, interviews and the reviews of policies and procedures.
During this exercise, it is more than likely that some amount of ‘visioning’ will take place and this feedback should be captured during the analysis; however, at this early stage, the only really important aspect of this effort would be where someone might say ‘we currently make use of this particular data source, but we would really like to use this other one and can’t because our systems can’t handle it’. Questioning and challenging the current state should be the focus during this evaluation stage. There will be ample opportunities to focus on the ‘to be’ environment as the project evolves.
Out of this ‘baseline’ analysis will come several recommendations concerning how the data strategy initiative will roll out. The baseline analysis will define all existing data sources, the resources required to manage each effort, the tools deployed in support and perhaps an estimate of what would be required to centrally manage the data operations. In turn, this analysis will form the basis of the business operating plan for the data strategy program and, in common with any business plan, it will include a set of objectives, suggested approach or tactical plan and the requirements or measurements for success.
It is likely the initial plan will be progressive in terms of scope and complexity for several reasons: Firstly, this initiative is likely to face significant internal skepticism and, therefore, early observed success will be important in retaining support. Remember, the CEO has also placed his or her ‘reputation’ behind this initiative, and for this reason, the initial roll-out should be modest. Secondly, taking a full-scale approach is unlikely to result in benefits that anyone will support, and finding and assigning an appropriate level of resources to an initiative of that breadth would be cost prohibitive. As often happens, trying to ‘boil the ocean’ will fail. Thirdly, starting the project on a relatively small scale will enable the project team to select a project scope they feel they can comfortably deliver on.
There are likely to be varying approaches to initiating a corporate data strategy but, for my money, I would start with product meta-data, which I will discuss in my next post.
The above is the second of several posts (1) on data strategy which I have completed for a presentation later this year.
Once a company has recognized the tangible value of its data, the CEO will assign the role to a direct report making this initiative his or her sole responsibility. As noted in my first post, managing data as an asset is so important that it requires direct senior management responsibility and should not be delegated to the head of marketing & sales or IT (in each or any case, both bias and conflict with their ‘real’ responsibilities will prevent the program's benefits accruing to the entire business). In addition, the ability to break down silos, confront the assumed internal politics and impose a solution will be greatly diminished if the executive in charge already has other functional responsibilities. In my view, if this is the approach of the CEO, the initiative will fail. By way of example, the Chief Data Officer at Dun & Bradstreet sits at the highest levels of the company in recognition of the importance of data to that organization and is evidence of the strategic importance of that data.
The first task of the “Chief Data Officer” (CDO) and their team would be to conduct a business-wide data audit to determine where data is being collected, both through internal efforts as well as externally by customers, vendors and other partners. This is an important first step and should be as expansive and specific as possible. Things to look for include not only the data sources themselves but their quality (perceived or explicit), frequency of update or creation, source/owner, format, whether the data conforms to any particular standard and whether the data is ever validated against this standard. Reviewing standards might also include not just the data’s technical framework but also whether it is subject to normalization – to a specific categorization method such as BISAC, for example. I once undertook an effort like this at a large educational publisher and the process took about six weeks and required site visits, interviews and the reviews of policies and procedures.
During this exercise, it is more than likely that some amount of ‘visioning’ will take place and this feedback should be captured during the analysis; however, at this early stage, the only really important aspect of this effort would be where someone might say ‘we currently make use of this particular data source, but we would really like to use this other one and can’t because our systems can’t handle it’. Questioning and challenging the current state should be the focus during this evaluation stage. There will be ample opportunities to focus on the ‘to be’ environment as the project evolves.
Out of this ‘baseline’ analysis will come several recommendations concerning how the data strategy initiative will roll out. The baseline analysis will define all existing data sources, the resources required to manage each effort, the tools deployed in support and perhaps an estimate of what would be required to centrally manage the data operations. In turn, this analysis will form the basis of the business operating plan for the data strategy program and, in common with any business plan, it will include a set of objectives, suggested approach or tactical plan and the requirements or measurements for success.
It is likely the initial plan will be progressive in terms of scope and complexity for several reasons: Firstly, this initiative is likely to face significant internal skepticism and, therefore, early observed success will be important in retaining support. Remember, the CEO has also placed his or her ‘reputation’ behind this initiative, and for this reason, the initial roll-out should be modest. Secondly, taking a full-scale approach is unlikely to result in benefits that anyone will support, and finding and assigning an appropriate level of resources to an initiative of that breadth would be cost prohibitive. As often happens, trying to ‘boil the ocean’ will fail. Thirdly, starting the project on a relatively small scale will enable the project team to select a project scope they feel they can comfortably deliver on.
There are likely to be varying approaches to initiating a corporate data strategy but, for my money, I would start with product meta-data, which I will discuss in my next post.
The above is the second of several posts (1) on data strategy which I have completed for a presentation later this year.
0 comments:
Post a Comment