Daybreak Analytic Database

Most mid-market companies make one mistake or another when investing in a data analytics platform, not understanding the many intricacies associated with preparing their data to get the best results. Some of the most common mistakes include:

• Not realizing they need to build pipelines to get the data from their multiple data sources to the analytics platform
• Tasking IT with implementing a data analytics solution, when the IT department does not have data science skillsets
• Basing analytics on data that is riddled with errors, incomplete, or stale, which compromises quality of decision-making due to the inaccuracy and tardiness of the underlying data.
• Relying on the reporting function of one data source and not taking into account data beyond that source for decision-making
• Using dashboards that provide insights into the past only, and not the future – a gap that needs to be bridged to compete with larger enterprises

It should also be noted that analytics requires massive storage and compute to mine data for actionable insights. Even in a cloud environment, which is less costly to maintain than on-premise servers, data analytics takes up a huge amount of compute to mine transactional data for AI-driven insights. Most data warehouses used by mid-market companies are not built for analytics, and their contracts with public cloud vendors for data storage often incur huge overage charges for compute spikes as millions of calculations are being completed for algorithms to converge for analytics results.  Data analytics needs a cloud built for analytics. The mid-market should demand a built-in analytics cloud from an analytics solution, without a third-party public cloud contract to make it work (or attempt to host it in their regular institution data storage).

Industry Knowledge is Key

One of the most important aspects to understand is that there is no one-size-fits-all when it comes to data analytics. The value lies in industry specific data models, which must be built with algorithms using salient data points for a specific industry and appropriately weighted for that industry. For example, to build customer revenue in financial services, mining transactional banking data is important to reveal if a customer is doing business with competing financial institutions so that action can be taken to win this business over. In manufacturing, comparing product inventory at various channels and channel or retail sales locations is important for discerning sales performance and growth opportunities. In healthcare, mining insurance reimbursement claims for underpayments to recapture lost revenue requires comparisons of contracted amounts and fee schedules for multiple private insurance companies, plan coverages, and more. The true business value of analytics lies in industry specific data models and the data enrichment made possible by deep learning and the generation of actionable insights.

The Importance of Having the Right Expertise

This brings us to the requirement for data scientists and business analysts, which assist with achieving powerful and current actionable insights that lead to AI-driven decision-making and better business outcomes. Data scientists build algorithms to detect trends, patterns and predictions based upon the data, to position an enterprise for the future. Business analysts are industry specific and connect the dots between the data points relevant for answering business questions in that particular industry. They also help to design dashboards and other data visualizations to ensure that the insights generated answer questions important to that industry using industry-specific terminology, and analyze analytics results in the context of industry knowledge to reveal growth drivers and other opportunities for driving revenue. Typical IT departments do not have these skill sets.

Data Analytics Platform Requirements

Consider these questions when evaluating a data analytics platform:

• Which data sources are forming the basis for the insights: Is the analysis based upon only some of the data, leaving out important data sources? Do the analytics use the most important data points? Is too much data or the wrong data being used?
• How is the data cleansed for accuracy: To judge the accuracy of the insights, know what is being done to eliminate errors in the underlying data being analyzed. Garbage in leads to exponential garbage out when data is turned over to AI. Can the data be trusted?
• Which algorithms are being used to find insights: Is it a specialized or generic deep learning model? Is it optimized for the type of inquiry or result being sought? Is it tailored to a specific industry and weighted appropriately for that industry and the question posed? Can the results be trusted?
• Are the analytics results providing actionable insights such as how to grow revenue, improve efficiency or achieve other business outcomes? Are the results tied to solving business challenges, and not just AI for the sake of being cool?

Side-by-Side Approach

Most mid-market companies are not in the business of IT or data science. IT is a necessary administrative function for operating a business, but should not become the main focus of most non-tech mid-market companies. However, given the expertise needed to use and maintain most analytics solutions, and the cost of data expert professionals as FTEs, the cost of data management and analytics can quickly overtake mainline business COGS expenses. Given all the complexities and challenges associated with unlocking the true value of data analytics, a new approach is needed that mid-market businesses can afford, enabling them to leverage AI-driven analytics and more effectively compete.

A side-by-side service model offers an alternative that goes beyond most tools and platforms on the market by providing a data platform with built-in data management and analytics, as well as access to human intelligence in data engineering, machine learning, and business analytics. Ideally it should operate as a cloud-based platform with a subscription service that places the burden of the data engineering expertise, technical tooling, and building and maintaining the infrastructure on the data management provider.

While many companies offer tools, and many consulting firms can provide guidance in choosing and implementing the tools, integration of all the tools and expertise in one end-to-end solution built for non-technical business users is key for digital transformation success for midmarket businesses.