Every industry has real-time analysis use cases. Yet on any average day opportunities are lost and unnecessary expenses are made because organizations don’t analyze data properly, or they don’t do it at the right time. In this blog post I want to share some thoughts about real-time analysis scenarios that may or may not apply to your industry. My goal with this post is to provoke thoughts for scenarios in your organization where real-time analysis would have an impact by increasing revenue, lowering cost, or both.

Let’s assume that almost everything gets recorded one way or the other, on paper, in log files, in a relational database, etc. Real-time analytics may be achieved by querying the records of events/changes directly and that would be perfect. In many cases, however, source systems are not necessarily optimized for data retrieval and/or heavily loaded by just processing transactions, and it makes sense to set up a separate reporting/analytics environment on different hardware or in the Cloud. In other cases some level of data integration is required to combine data from different systems. In these scenarios real-time replication can be the solution that enables real-time analytics.

1. The first example is straightforward and one that already has a fair level of sophisticated analysis: the web store e.g. Amazon.com, eBay.com, etc. If I express my intend to purchase a product then I may be presented with competitive offerings and related products in an effort to convince me to buy a different, possibly higher margin, product, or to buy more than I had originally planned. Of course in this example both the web store and the actual seller could have their own (sometimes conflicting) intentions to influence my purchase decision. The suggestion for alternatives or related offerings makes most sense when I am shopping because I may consider alternatives until I buy and as I decide to make a purchase I may be able to save shipping cost by combining multiple products in a single order.

2. Following on from the first example is the physical store. This is where location-based analysis can be useful. Maybe the right promotion at the right time can convince me to enter a store that I would have walked by otherwise. If a store knows what aisle I am in and what products I generally buy then maybe I can be guided there. The store may be interested in knowing what shelf I am focusing on (maybe that is what Google Glass was intended to do), etc. Of course various services like Amazon Fresh, Google Express and others are trying hard to change the physical store visit into a web store purchase which leads back to the first example.

3. Fraud detection is another obvious example that is much easier to resolve as it happens rather than some time after the fact. For example, credit card fraud is still all too common and both consumers and vendors go through a lot of hassle and expenses to recover from fraudulent transactions. But there are similar examples in the gambling industry, in sports, on stock markets, etc.

4. Hospitals make critical decisions that can make the difference between life and death. Decisions are based on information, and the better the information the more-informed the decision. It is good to have an experienced doctor take care of you, but it is even better if that doctor has all the relevant information and diagnostics at hand to make the best decision.

5. Mining and exploration requires massive up-front investments. A one-day outage can easily result in millions of dollars of equipment and personnel expenses lost, and real-time analysis of equipment diagnostics may help prevent outages. Likewise real-time analysis of findings may also lead to quicker findings and hence save millions of dollars by shortening exploration projects.

6. Transportation companies such as FedEx and UPS may benefit from real-time adjustments to routes based on up-to-date traffic conditions. A lot more than just the traffic conditions weighs into a rerouting decision including – if it relates to a delivery truck – optimum delivery order (and with that how the truck was packed), delivery time commitments, etc.

7. The last example for this post is another obvious one: stock market trading. Whoever responds to major news the quickest can make most money. Imagine how much money could have been made in just a couple of days if you were the first one to learn that Volkswagen had been cheating on the emission tests (provided you are not an insider, because in that case the fraud example would be relevant). There are also less direct ripple effects e.g. Volkswagen may (temporarily) sell fewer cars that could affect Volkswagen’s suppliers, etc.

So as you can see there are lots of examples in many industries. What is your use case? We’d be happy to discuss how HVR could be applied in your industry. Contact us or request a consultation to get started!

About Mark

Mark Van de Wiel is the CTO for HVR. He has a strong background in data replication as well as real-time Business Intelligence and analytics.

© 2019 HVR

Try now Contact us