Is your organization ready for Data Science ???
Data science(DS) can add immense value to business by contributing actionable insights across workflow, be it hiring new candidates or helping senior staff make better and informed decisions. The ability of DS to add value to an organization is industry agnostic.
“In 2016, McKinsey estimated that big data initiatives in the US healthcare system could account for $300 billion to $450 billion in reduced health-care spending, which roughly estimates to 12–17 percent of the $2.6 trillion baseline in US health-care costs.”
How does Data Science add value to business?
- Empower management and officers to make better decisions with quantifiable data driven evidence
- Direct the actions of senior management based on trends which in turn help in defining goals
- Challenge the staff to adopt best practices
- Identify exciting new business opportunities
- Analyse and target relevant audience
- Quantify the effect of decisions taken
“Capital One is making an impact on nearly 60 million customers by using data science algorithms to develop next generation of financial products and services. Analytics is highly influential at Capital One bank as the bank has significantly grown its earnings per share by more than 20%. Capital One is now the third largest provider of credit cards in US.”
Even if Data Science expands the horizon of our understanding of data and making data backed decisions, it still lacks notable absorption in organizations. At this point, the need to discuss some road blocks for adoption of Data Science in organizations becomes necessary.
Roadblocks to Data Science
- Culture
There has been a tectonic shift in the culture of how decisions were made just a decade from now and the way they are made today. A culture that used to largely depends on gut feeling based decision making is getting slowly transformed into a culture that is much more objective and data driven and embraces the power of data and technology, and this tectonic change in culture is not seen very welcoming by some organizations.
- Uncomforted Managers
Data Science analyses revolve around data and inevitably expose data Irregularities and faults in the process, this may be caused during data collection and poor data management process, which might cause unease in managers handling these departments.
- Data Sharing Issues
Data Sharing issue is one of the most crucial stage in any data science project. Not every company is open for data sharing which causes major road blocks to the data science implementation and extracting meaningful insights.
The Data Sharing issues- may arise from either the company’s policy or due to the presence of sensitive and personal information of clients.
- Lack of Analytics and Data Science relevant man power/resources
A Survey conducted by UK based data science consulting and resource management firm Jaywing stated that:
“Despite nine in 10 marketers (92%) agreeing that data management is a key priority for their business, two-fifths of marketers (40%) believe a lack of analytics and data science skills is preventing them from delivering effective customer relationship management (CRM) strategies”
There is indeed a shortage in Supply for skilled data science professionals, which ultimately hinders the intent of firms going for data science projects.
- Non realization of momentousness of analytics
Even if its well known that predictive analytics largely benefit an organization, it is true that most organizations can function reasonably well even without it. At least until their competitors start driving optimized decisions based on deep analytics. Which ultimately leads to a state in which companies wait for an impetus to actually drive them towards data science implementation in their organization.
Apart from the listed Roadblocks, we have also seen the following reasons for the poor absorption of data science in some of the organizations.
- Data Science is expensive
A common misconception is that data science is really expensive but just by looking at how much money was being spent during the days of the early dotcoms, ‘CRM fever’ or even now on ‘No SQL’ solutions and Hadoop-like platforms, the data science looks really frugal and not to forget the long term benefits the insights will provide, it will not only help save companies billions of dollars but also helps in transforming into data centric companies.
- Presence of non usable data
Some organizations have a misconception that their might not be enough information contained in their data or their data might be in really bad shape to be used up for any data science projects. This is a clear case of misconception as data science projects rarely have the opportunity to work with pristine datasets as almost all the cases of data science projects have some data quality and data quantity issues that data science team finds a workaround.
“Major insurers in European union have made fraud detection and analytics part of their core business rules and development, They combined all (business) rules in the company and put mathematical modelling on this data and got the necessary accuracy to find fraudulent cases with a 72% success ratio for 20% of all claims.”
Pointers to consider for assessing the preparedness of organisations
- A Business question to solve
Presence of a business problem to be solved is the basis for any data science project, if suitable business problem is already not formulated at client side then data science team can help client formulate business problem.
- Data Volume
As the name suggests the basic building block for any “data” science project is the presence of data. The data preferably needs to be cleaned from any anomalies and should be representative of the project’s deliverable. “The more the data the better” is the mantra.
- Proper data storage
Organisation should manage the processes needed to ensure its timeliness, accuracy, completeness and reliability. Data -sanctity should be kept at the highest point of cardinality and all measures needs to be taken while storing data, specially during data migration when most cases of data loss is seen. Data should be considered as an asset and therefore it should be managed accordingly.
- Suitable Manpower
Resources having proper knowledge about the entire business process along with the understanding of flow of data is required at the stakeholder side to better understand the business problem.
- Readiness in undergoing an iterative process for model refinement
DS projects are iterative in nature which makes the engagement of business experts very crucial hence client side must be ready for frequent engagement with DS team after deployment of services, so as to better understand the process and provide feedback, which will help in model refinement and better processes.
- Proper data collection at source and management
Data Science involves dealing with computers which by themselves are dumb machines and would thus require quality data collection to be fed for analysis. Data should be treated as a core asset for the entire organization and proper measures needs to be taken so as to maintain its sanctity.
- Data Dictionary and computational resources
Readily available data related documents such as data dictionary and servers that can take computational requirement of that machine learning process.
- An open mindset for insights and faults
Employing data science project will unearth a number of insights and faults present in the current process of decision making and organization must be ready to take it in a positive direction to better improve their processes.
- Adoption and usability
The data science project is a futile exercise if after completion of the project, the predictions, recommendations and faults unearthed during the project are not considered while taking business decisions by senior management.
“Published by Netflix executives, the on-demand video streaming service claims its AI assisted recommendation system saves the company $1 billion per year. This means Netflix can confidently spend huge sums ($6 billion a year) on new content, knowing viewers will consume enough over time to give them healthy returns on the investment.”
Checklist for DS adoption in organisation
“Dextro analytics, an analytical company claims that they can successfully predict 98% of fraudulent claims through new system and the fraud model also reduced the false positive by virtually 85%, providing superior service to legitimate claims.”