BI AND THE PATH TO PREDICTIVE ANALYTICS
Transcription
BI AND THE PATH TO PREDICTIVE ANALYTICS
BI AND THE PATH ANALYTICS TO PREDICTIVE A Review of the Perceivant BI Platform for MidMarket Organizations Robin Bloor, Ph.D. Rebecca Jozwiak WHITE PAPER BI AND THE PATH TO PREDICTIVE ANALYTICS Executive Summary Perceivant’s Platform and Services deliver a comprehensive Business Intelligence (BI) solution that focuses primarily on mid-sized businesses. In this paper we discuss its capability in the context of the business use of BI. Our conclusions are summarized in the following bullet points: • Mid-sized companies grow at a faster rate than large enterprise organizations and experience the same need as large organizations to exploit BI. In recent years, the investment in BI has been increasing at a greater rate than almost all other investments in IT. • Nevertheless, most mid-market organizations have yet to properly exploit the opportunity that BI provides, most of them having gone no further than to implement regular reporting. The investment in dashboards and, to a larger degree advanced analytics, has lagged. • Conceptually, BI can be categorized as providing four distinct capabilities, all of which can be provided by Perceivant: ! – Hindsight: Consisting primarily of regular reporting ! – Oversight: Consisting primarily of dashboards and associated capabilities ! – Insight: Consisting primarily of drilldown capabilities and data mining ! – Foresight: Consisting primarily of predictive analytics • The Perceivant Platform targets the mid-market by offering a relatively low-cost, lowrisk software stack to provide BI to entry level users. It can be tailored for customization or used out of the box. This can be regarded as a BI foundation for the business. • Perceivant Services enhances this foundation through consultation both on the IT side to specialize and enhance the capabilities it delivers, and in the area of analytics where it can assist customers by providing data scientist expertise and consulting to exploit its platform • As a consequence mid-market businesses are equipped to compete head-to-head with larger organizations reaping the many benefits of a comprehensive BI capability, as illustrated by the brief use cases described in this paper. In our view, mid-market companies would be wise to consider the solutions and services Perceivant provides. 1 BI AND THE PATH TO PREDICTIVE ANALYTICS What is a Mid-Market Company to Do? Definitions vary, but a mid-market company can generally be defined as having between 50 and 1,000 employees, with revenues in the $2 million to $100 million range. Research suggests that while mid-sized companies account for just 3.2% of all companies in the U.S., they provide 31% of all revenue and grow – in terms of both revenue and jobs – at a faster rate than larger companies. Although organizations at this level are in the expansion phase, they have established a niche and often have a long-term plan for growth. As such, a recent survey showed that 60% of mid-market executives ranked technology as having the greatest potential to increase business productivity. Large businesses often have extensive BI capabilities established over years, involving a variety of capabilities. In mid-market organizations an extensive focus on BI is less common. Most likely BI functionalities will include regular reporting coupled with the ability to extract data from various sources to be analyzed in MS Excel. The prevalence of media stories about Big Data can sound daunting to a mid-market organization, and while some may think they will never have the need to gather large heaps of data and subject it to any kind of analysis, Big Data can be a relative term. Even if they do recognize the need, it is unlikely that these companies have the resources in the form of IT staff or “data scientists” to explore and leverage it. Nevertheless they gather and process data, and the volume of their data increases year after year. There can be little doubt that such businesses could profitably employ BI technologies that, at the moment, they tend not to use, and perhaps know little about. The rapid adoption of BI technology has been an established trend in the IT market for well over a decade. According to analyst reports, BI spending has exhibited double-digit growth in most years, growing at a faster rate than most other areas of IT spending. Clearly large organizations have found good reason to invest in this technology, but few businesses in the mid-market have capitalized on the business opportunity that BI technology presents. Categories of BI There is a wide variety of BI software that can be deployed in almost any business. For the sake of clarity, we can classify all BI products and services as belonging to one of four categories, according to what they deliver: hindsight, oversight, insight and foresight. • Hindsight: To this category belongs the reporting software that almost all businesses use: regular reports, aggregations, decision support summaries, trend reports and so on, often including graphs and bar charts for clarity. Such reports are historic to some degree, since they are compiled from information that is days, weeks or months old. • Oversight: There are a variety of BI capabilities that help to provide oversight in the sense of providing fairly up-to-date information, often showing thresholds that may directly trigger alerts to some staff when an encouraging or discouraging trend appears. These include dashboards and business process oversight capabilities. They are normally fed by fairly recent or even real-time information. • Insight: BI tools that provide a business with insight range from online analytical processing (OLAP) to data mining. OLAP provides drill-down capabilities that allow the user to investigate the fine details of aggregations of data: sales figures, 2 BI AND THE PATH TO PREDICTIVE ANALYTICS manufacturing defects, price trends and so on. Data mining embodies a variety of sophisticated statistical techniques that reveal important trends and correlations. Such BI capabilities are most effective when they harmonize with business processes to ensure that their results lead to action. • Foresight: In terms of delivering business benefit, predictive analytics can be the most effective aspect of BI. It can transform a business from being reactive to proactive. Rather than simply responding to business events and trends, predictive analytics enables a company to anticipate and exploit such events and trends. There is a rough logical order to the implementation of these BI categories, from the perspectives of both IT and business. From the IT perspective, many reporting capabilities (hindsight) can be fed directly by data from source systems. Where this is not feasible, a data warehouse that organizes company data can be built and used to feed such capabilities. FORESIGHT (Predictive analytics) Business Benefit Also IT complexity INSIGHT (OLAP, data mining, etc.) OVERSIGHT (dashboards, alerts, etc.) HINDSIGHT (reporting) Oversight can be more challenging depending upon how current the data needs to be. Dashboards, for Time example, require data feeds to be more up-to-date than is sometimes Figure 1. The Evolution of BI feasible to achieve with a data warehouse. Nevertheless it is usually possible to establish data feeds that are as current as such BI applications need them to be. Insight is a far more specialized area, since it involves extracting specific subsets of data and applying a whole series of analytical techniques until useful knowledge is discovered within the data. This is an ongoing data exploration activity which can fan out in many different directions. Finally, the foresight of predictive analytics usually requires the integration of data from many sources including sources external to the business. From the IT perspective, it can be a significant challenge, but from the business perspective, it can be essential for growth. As we illustrate in Figure 1, while the implementation of the different categories of BI becomes increasingly challenging, the business benefits escalate with each new capability. It is important to note that these business benefits will not be delivered unless the business itself adapts its processes to accommodate the intelligence that they provide. This is particularly important when it comes to insight and foresight. Mid-sized businesses are unlikely to be able to attract fully experienced data scientists, since data analytics is a highly skilled activity that requires a knowledge of statistics, and predictive analytics tools need to be presided over by such a skilled individual. In addition, it is important that the business adjusts its activities to swiftly and surely leverage the knowledge that such activities deliver. This is not so easy to achieve. 3 BI AND THE PATH TO PREDICTIVE ANALYTICS Savvy mid-market companies want to expand, and their challenges are as complex and relevant as those of a large enterprise, if not more so. They have the advantage, however, of being nimble enough to adjust business processes more quickly than their monolithic counterparts. Adopting advanced BI tools that cater to mid-market operations and resources can undoubtedly put a mid-sized business at the leading edge of the competition. One company that specializes in delivering and implementing advanced BI capabilities to mid-range businesses is Perceivant. We now describe the solutions the company offers. 4 BI AND THE PATH TO PREDICTIVE ANALYTICS The Perceivant Platform and Services A frequent barrier to mid-market adoption of BI technology is cost, both in terms of hard and soft resources. Growing companies simply cannot justify the risk and expense of the hardware, software, training and licensing that usually accompany a BI implementation. Perceivant offers a scalable software-as-a-service (SaaS) platform which eliminates the need to purchase additional machines and reduces the human capital required to maintain the environment, in turn providing enterprise-grade software at a fraction of the cost. What’s more, Perceivant lets customers begin with a low-risk, short-term license which can be extended after adoption. As data grows, it tends to disperse in ways that make it difficult, if not impossible, to access. This is especially true for organizations that don’t have individuals whose specific job is to manage the data. Perceivant built its platform to negate this disparity, keeping the business user in mind. It collects and joins all the organization’s data and stores it in a custom, purpose-built data warehouse in the cloud. This gives mid-sized businesses the opportunity to perform complex analysis on much larger data sets than they ever had before. Users immediately have access to real-time data and the advanced analytical tools Perceivant provides, and if desired, they can export the results to MS Excel or the interface to which they are accustomed. It is worth noting that this takes place without any requests to IT. Perceivant does this by helping customers leverage the power of familiar technologies: Hadoop, Elastic Search and NoSQL databases. It also integrates with other public services such as Google Prediction and Google BigQuery. But it has developed its own proprietary and unique integration tools to make access to data and analytics both painless and easy to use. In a recent performance benchmark using web traffic data, Perceivant returned a query 5x faster than Hadoop’s Hive, and it did so with only one server, compared to Hive’s six. It has properly architected a platform solution that merges the power of commodity software with a specialized, purpose-designed user interface, which can lead to faster insights and better business process. As with any hosted service, users do not have to learn the inner workings of several different applications, but instead can operate within their existing environment. Out of the box, the Perceivant Platform is natively suitable for the typical needs of the customer. Additionally, from the initial implementation to the end-user experience, Perceivant Services can provide custom solutions where required. Mid-market businesses thrive because they know their bread and butter: the customer. Perceivant follows this model and offers tailored solutions to meet the needs of the organization through direct consultancy and a deep knowledge of analytics. Through its software and services, Perceivant can guide new-toBI users from basic reporting (hindsight) to the proactive realm of predictive analytics (foresight). The Perceivant message is clear: get the data in and give access, fast. Mid-range businesses do not have the luxury of waiting six months or more for a BI installation. Time-to-value is critical. The Perceivant Platform provides customers with an affordable entry point to BI via its integrated software stack and can lead them down the path to achieve hindsight, oversight, insight and foresight. A mid-sized business can transform its operations and processes literally overnight, gaining advantage over competitors and closing the distance between itself and better armed companies who use these tools every day. 5 BI AND THE PATH TO PREDICTIVE ANALYTICS In the corporate environment, those who analyze, win. It is our view that Perceivant delivers exactly the kind of business value that end users need: a unified view of information assets, the ability to perform complex analytics on those assets, and an easy-to-use interface. Taken together, the solutions that it provides can catapult a mid-sized company into unprecedented growth. Perceivant has removed the barriers that often hinder BI adoption, and because of this, mid-sized companies now have the opportunity to leverage an enterprise-class solution with enterprise-class results. Organizations that fall into the mid-market category and wish to reap the benefits of BI would do well to consider Perceivant. 6 BI AND THE PATH TO PREDICTIVE ANALYTICS How Customers Use Perceivant College Correlates Data Points to Increase Retention and Plan for the Future Student recruitment and retention are the primary business goals for any college and are ultimately driven by many factors: quality of teachers, availability of courses, classroom size and demographic measures outside of the college’s control. One of the largest community colleges in the United States had all of its data about teachers, courses, and full and part time students in a traditional ERP system. Its data analytics platform could provide baseline reporting on these areas but lacked the ability to correlate data points between the silos of information. In addition, these reports took over 20 hours to render, at a cost of $250,000 per year for the software license alone, with additional support and hardware costs on top of that. The system also provided no historical view of data. The college chose to move to the Perceivant Platform for its knowledge of predictive analytics and its ability to offer services that would help leverage data at a market-competitive price. The real-time nature of the reporting, the ability for these reports to be accessed by individuals across the organization (rather than only by the IT staff) and Perceivant’s native ability to cleanse data were also deciding factors. Using the Perceivant Platform, the college will be able to connect data points around enrollment, attrition, teachers and class assignments, student prospects and facility management. They will use the data to predict future enrollment, manage class sizes, decide course offerings and more accurately plan classroom usage in an effort to improve retention for future semesters. They’ll utilize historical data to decipher what contributed to drop-out rates and combine these with external economic data to create demographic-specific retention programs. In short, the Perceivant Platform is aiding the college in a path to cost savings, better utilization of resources and increased student achievement. Software Company Increases Efficiencies and Real-Time Response to Customer Requests A software company serving some of the world’s largest online retailers was tasked with analyzing 5TB of website traffic and millions of product SKUs per client. With 50 employees and $6 million in revenue, human and capital resources were scarce. Often asked to find correlating product and shopper attributes in the midst of these enormous amounts of traffic and SKUs, the company’s existing solution was a combination of Hadoop/ Hive and MongoDB, with updates entering the system nightly via batch. The limitations of this solution for the company were two-fold. First, many internal, cross-team and technical resources were needed to prepare and export the data, with even more resources needed to analyze it. Productivity suffered with the inefficiencies of creating a customer support ticket, handing it off to another department, waiting for the data query to occur, then waiting for the results. Additionally, if insufficient data was available, it would not be known until the entire query/send process was complete, 7 BI AND THE PATH TO PREDICTIVE ANALYTICS forcing analysts to restart the process. Attaining a full and complete data set would often take an entire day, something the company could neither afford nor manage to meet the real-time demands of its clients. The second challenge was cost versus performance. Trying to run analytics against Hive was still too complex, and it was slow, even for basic queries. After attempting many changes within its own architecture, the company chose the Perceivant Platform. Users found the interface familiar, resembling an Excel PivotTable, and on-boarding new users to the platform was a simple task. Rather than data access and analysis being a cross-functional process within the organization, data analysts had quick access to the actual data and reporting functions, making them more efficient and responsive to customer requests. The company has found Perceivant to be considerably cheaper and faster. Queries that took 36 seconds in the Hive setup are taking 7 seconds on the Perceivant Platform, at a lower cost. Healthcare Company Opens New Revenue Streams On average, companies pay almost $3.00 an hour to provide health benefits to their workers. Decreasing these costs can therefore have significant bottom-line impact. A healthcare company serving self-insured businesses received 7 million health insurance claims a day for processing. The company wanted the capability to turn these claims into usable information their clients could leverage to better know their employees’ health challenges (in aggregate) and reduce healthcare costs. While the company investigated developing its own in-house solution, the Perceivant Platform’s competitive pricing and hosted solution, with no required internal IT resources, changed its mind. The company used the Perceivant Platform to integrate its BI tools with its client portal to provide high-level access to claims data. This initial implementation came with few set-up costs and launched quickly, with very little risk. In Phase Two of the project, the company will utilize Perceivant to provide additional BI features, creating a new revenue stream by providing predictive analytical data based on claim information. About The Bloor Group The Bloor Group is a consulting, research and technology analysis firm that focuses on open research and the use of modern media to gather knowledge and disseminate it to IT users. Visit both www.TheBloorGroup.com and www.InsideAnalysis.com for more information. The Bloor Group is the sole copyright holder of this publication. ❏ PO Box 200638 ❏ Austin, T X 78 7 20 ❏ Tel: 512–524–3689 ❏ www.InsideAnalysis.com www.BloorGroup.com 8