Collecting data is definitely the trend these days, even more so that we are now taking about big data, data mining, machine learning, deep learning, etc. But not all data are made equal. Your data is only as good as the use you make of it ... and it comes at a cost.
But what makes data valuable? Understanding the organizational data, how it is used in decision making or to draw conclusions is paramount to implement a good data management strategy and bank on your investments.
Collecting and processing quality data will enhance the value of the information you gain from it.
Keep in mind the following characteristics when looking at your data and its potential to generate value:
How does the data relate to the context in which you are using it? Before investing in collecting data, it should be clear on how it will support your research or organization’s goals.
If you cannot explain how the data will be used, you should question yourself why you invest money in collecting it.
Is the data collected free of errors and from a reliable source? Quality data that does not contain mistakes, such as typos, redundancies, etc., are undoubtfully critical in a relational world.
Unreliable data you inject into your process will directly impact the quality of the information and the decisions you will make based on it. Inaccurate data lead to no or even negative return on your investment.
Does the data include enough information to be useful? When investing in collecting and transforming data, take a hard look at how comprehensive the information is. When processing it, also think about including all of the data you have available to provide for more thorough analysis.
Incomplete data lead to mistakes, false assumptions and at the end might simply be unusable, and costly.
Can the data be interpreted the same way each time it is used? Consistency of data means that it exists only one version of the same data and that it is maintained.
Bias injected in the analysis process by using inconsistent data may leads to costly mistakes.
Is the data timely available and accessible when needed? How good is the money spent on collecting data if it is not readily available to the users that to extract value from it.
Timing is everything and time is money. Your investment will be worthless if the data cannot be used promptly..
Can changes made to the data be traced back in time?
Investing in keeping high quality data includes implementing the adequate controls to preserve its integrity and assess its relevance over time.
Not understanding how the data evolves within the organization may impact the value of the information derived from it.
To reflect on the real cost of data, let’s take a look at its lifecycle and the different costs associated to it.
This first stage relates to moment when the data enters your organization. Whether you generate it, acquire it from other systems or applications, receive it from various devices, somehow data is created and requires to be stored somewhere.
Collecting data is usually labor intensive. Data collected also require significant time to cleanse and restructure. Making it usable by the various systems can also be complex and prone to error.
Besides the investment in human resources and expertise, the data collection phase also includes costs related to storage, such as disk space, and access, in the form of applications and systems.
Once the data collected is prepared and saved, there are many processes that may be undertaken to prepare raw data to be used by various processes in order to convert it into valuable information.
While the order of operations may vary, data preparation typically includes integrating data from multiple sources, validating data, and applying various transformations. During those process, the data generally are reformatted, summarized, subset, standardized, and enriched as part of the data processing workflow.
The costs associated to this stage is often associated to investments in computer hardware, memory, CPU capacity and of course into the talent of developers that can program the various treatments required.
With data ready to be consumed, the analysis/reporting stage is where the most critical treatments are taking place. Regardless the degree or the method of exploring and interpreting the data, investments in this stage are the ones for which the higher Return on Investment (ROI) is expected.
The investments made at this level are mainly focused around the acquisition of applications or expertise to extract the knowledge that will generate value for the organization or research projects. With the volume of raw data that are now readily available and/or generated, gaining efficiency at this stage is critical and therefore directly correlated with the latest advancement in data modelling and Artificial intelligence methods, the human brain being no longer suited to make sense of such volume of data.
Backup and Archiving
Once the analysis is performed and the information extracted is shared internally or externally, as required, the data is typically stored for future reference. Also, throughout its lifecycle, safeguards need to be implemented to ensure its integrity is preserved and access is provided. This is true not only for the raw data, but also for the information that was extracted from it that must also be stored and preserved.
The investments at this stage include disk space, development of data warehouses, business intelligence and data visualization tools, etc.
Quality data from which valuable information can be retrieved is key to provide organizations with competitive edge and credibility to generate growth.
As mentioned earlier, the return on investment in your data will materialize with the output of the analysis and sharing of the information extracted from it.
Maximum value is extracted from your data only when you can make the outcome your analysis (or information extraction) fully explainable, therefore reliable.
Extracting knowledge from your data must be a controlled process. As such, you must understand and effectively manage each step of the process to ensure the integrity of the information is preserved.
Sound data modelling can provide the structure you require to better visualize your data and detect error and oversight early in the process when it cost the least. Ensuring that reliable and interpretable methods are used to transform and analyze your data will increase the reliability and value of the information you obtain.
Timely decisions made from reliable and well-structured data undergoing processing using justifiable methods providing results that can be explained will maximize return on your investment.
Once you have paid to collected, processed, analyzed, and store your data, what else can be done to extract more value from it? Several strategies are possible, and often overlooked to extract more value. Let’s explore a few:
Maintain it One obvious strategy to take to extract more value from your data seems paradoxical as it requires more investments. However, keeping your data clean and relevant will indeed create savings in the long run by minimizing costly mistakes that may originate from unreliable data.
Reuse it Unfortunately, data is often used once and then tabled or discarded. In a lot of cases, the same data could be reused by other parts of the organization or by other researchers, if only they would be aware of its existence. Each time data is reused, it generates value without requiring much additional investment.
Enrich it The data you have available can be combined and enriched to increase its capacity to generate additional value. Enriching data can take many forms ranging from obtaining more attributes or by injecting additional data in your analysis to obtain more valuable information of conclusions.
Share it Sharing data within your organization, between peers, within a community with similar interest can bonify your investments. If you can combine or regroup investments to target the collection of complementary data, the synergy will be worth the effort.
Choking on data while starving for information? Step back and take a closer look for yourself …