For many companies, the decision in the future will be whether they need to redesign their systems so that real-time or near real-time analytics run alongside their transaction processing.
A high transaction industry like financial services, where nearly 24,000 transactions per second can be processed, is a prime example. And if you’re a retailer or an airline reservation company, you will want to know where consumers are spending money with you so you can optimize your revenue opportunities. Since high volumes of transactions are part of your business, the need to see where the money is going–so you can spontaneously present promotions to prospects and customers–is a critical ingredient for success.
SEE: Prescriptive analytics: A cheat sheet (TechRepublic)
Here’s how it works.
Say that you’re a worldwide hospitality chain. You process thousands of room reservation transactions per second for all of your worldwide properties, but you want to optimize your revenue opportunities and you still have empty rooms at some locations. Consequently, you want real-time or near real time analytics to run alongside your transaction processing so you can find out where you have a surplus of empty rooms around the world that you want to fill.
If the room surplus is in Beijing, you might present an instantaneous promotion to customers who live near the area or who have stayed at your Beijing properties in the past. If the room surplus is in Minneapolis, you will want to do the same.
When you can present compelling offers on the spot, you can be first to market for harder-to-sell inventory. Real-time or near-real-time analytics can tell you where your empty rooms are and who, based on demographics and past usage patterns, are the most likely customers to purchase these rooms.
EnterpriseIoT calculator: TCO and ROI (Tech Pro Research)
Commercial software providers understand these revenue optimization and real-time analytics needs. They also know that as other use cases emerge, such as Industry 4.0 with its real-time and IoT-driven manufacturing operations and maintenance, similar needs for running analytics alongside transactions will emerge.
One company that is answering this challenge is SAP, which now offers a SAP HANA (high performance analytic appliance) system as an alternative to its standard ERP. SAP HANA is more expensive–but it can run analytics alongside transactions, enabling companies to respond to markets, equipment maintenance needs, or anything else they feel requires an instantaneous or near-instantaneous response or intervention.
All of this sound great–but the decision point for IT leaders dealing with transactional and big data is this: Do you actually need real-time performance from both?
Here are several best practices for those who must make that decision.
1. Assess your business
If you’re the CIO, meet with other C-level executives and the CEO to assess your company’s business and determine whether you need real-time analytics.
If you’re a high transaction volume business and you want to maximize revenues (e.g., airline reservations, hotel reservations, online e-commerce), real-time analytics can provide you with insights that allow you to respond on the fly to market opportunities with promotions to the most likely customers. If you’re a manufacturer and see no need this for real-time analytics now, but foresee that you will need it as you convert to IoT, automated operations, hands-free maintenance, etc., real-time analytics might be on your IT roadmap.
In all cases, the business value proposition has to pencil out in terms of revenue gains because real-time analytics is expensive.
SEE: 60 ways to get the most value from your big data initiatives (free TechRepublic PDF)
2. Tell key stakeholders if you don’t see a need for real-time analytics
Many companies won’t require real-time analytics to capably run their businesses. Instead, they’ll need effective analytics reports that guide decision-makers to innovative ways to improve revenue and reduce costs. If you’re a CIO in this category, you should clearly present your view to the CEO, the board, and other key stakeholders who have all read articles or seen ads that promote real-time analytics. You should be prepared to discuss with them the pros and cons and why you don’t see a present or near-term need for real-time analytics–as well as what type of business transformation would need to occur for you to reassess your position.
3. Expect challenges in the implementation
Your marketing staff and data scientists will need to develop relevant questions that can unearth the data that lets the company optimize revenues. And if you’re implementing real-time analytics alongside a transaction processing system that already exists, both systems will need to be integrated and your staff must be adequately trained to run and support them.
4. Build your metrics and arrange for a trial
Real-time analytics is a relatively new technology. Most vendors know that to sell it, they have to demonstrate that it works. Therefore, if you’re a CIO, work together with other C-level peers to construct a set of metrics you want to measure against to ensure that you are getting the value you want from this real-time analytics system. Then, arrange with the vendor for a proof-of-concept trial so you and your stakeholders can see if those results are delivered.
Has your organization been assessing the need for real-time analytics? What criteria have pushed you toward a decision? Share your thoughts and opinions with fellow TechRepublic members.