Uri Maoz, Anodot VP U.S. Sales & Business Development, offers a variety of uses cases across industries for real-time anomaly detection in business data. Presented at the 2017 Innovation Enterprise Chief Data Officer Summit.

In 1998, NASA lost their $125 million Mars Climate Orbiter when the spacecraft burned up in the Martian atmosphere. While the engineering and software were meticulously built to NASA’s high standards and operated as intended, the data that put the spacecraft on the doomed trajectory was flawed. The navigation team at the Jet Propulsion Laboratory used the metric system for its calculations, while Lockheed Martin Astronautics in Denver, who designed and built the spacecraft, provided crucial acceleration data in the English system of inches, feet and pounds. JPL’s engineers assumed the acceleration data (measured in English units of pound-seconds) was in metric measure of force and sent the spacecraft on a doomed and costly flight. While, most data quality mistakes don't end in the fiery destruction of multi-million dollar spacecraft, misunderstood data is costly for today’s businesses. Data is the lifeblood of every company, helping companies work better, work smarter, and reach their target audiences. According to Gartner, modern business intelligence (BI) and analytics continues to expand more rapidly than the overall market, offsetting declines in traditional BI spending. Data quality is still the biggest challenge.  While many companies are investing in BI visualization tools, they are not necessarily applying the same efforts to the data itself. Then companies face frustration and disappointment when the ‘right’ data is not processed. Data Isn't Treated as a Business Asset Even though data is at the heart business decisions, companies don't always handle their data as an enterprise asset. Data may be handled tactically, with databases and applications created as requested by a business unit. Enterprise-wide data dictionaries are rarely applied to enforce consistency on the meaning of fields, and departmental IT teams address issues in isolation from wider business goals. The overall approach is ad-hoc, leading to a fractured data system, leaving the business to question the reliability of their data. Data Fuels Insights...Unless it’s Wrong Companies are often more focused on simply collecting data, losing sight how to ensure the quality of data. Unreliable data undermines business’ ability to perform meaningful analytics that support smart decision-making and efficient workflows. Quality data is required across the organization; for management, operations, compliance, and interaction with external partners, vendors, and customers. Maintaining Good Data Quality What makes good quality data? Data quality is measured by many factors, including: Even a dataset that seems accurate and consistent can lead to poor results when there is missing fields or outdated data. Maintaining high quality data is a real business challenge. It is further complicated by the dynamic nature of different data generation resources and devices, and the enormous scale of data itself. Companies need to confront their data quality challenges, before eroding trust in their data. When trust in data is lost, this is shared, leading to questions by all levels of the organization. Data Quality Case Study: An E-Commerce Company Misses a Key Event Here's a recent example of how data issues led an e-commerce company to make some costly business decisions. The e-commerce company collects event data through its mobile app, feeding data to a central data repository that drives their analytics and customer strategy. Every page and every click is collected for analysis, including tracking when products are added to or removed from a cart, how a user searches, and other user interactions on the site. There are potentially hundreds of events from each page. When a new version of the app was deployed, it had a bug that failed to collect some event data for certain iOS versions. Because of the large volume of data that's collected, the missing data problem wasn't noticed and went unidentified for several weeks. As a result, the business perceived a drop in purchases (while in reality the opposite occurred) and in reaction they increased the marketing budget for a specific product. Unfortunately,  in reality, there was no need for that increased marketing investment and would have been better spent elsewhere. The Neverending Story: Ensuring Quality Data Artificial Intelligence can be used to rapidly transform vast volumes of big data into trusted business information. You can immediately address problems, saving weeks of inaccurately reported data. Renewed trust in the quality of your data directly impacts business priorities. Anodot's AI-powered analytics solution automatically learns the normal behavior for each data stream, flagging any abnormal behavior. Using Anodot, changes that can impact data quality would be immediately alerted on, so that they can be addressed. preventing wasted time and energy and ensuring that decisions are made based on complete and accurate data.

Wix needed a real-time alert system that would indicate issues without manual threshold settings in the key metrics. Anodot proved to be the system required for providing the necessary insights to the company's analysts.

Anodot Sr. Director of Customer Success Nir Kalish presents how eCommerce organizations can leverage the powerful anomaly detection capabilities of Anodot’s AI analyitcs to proactively address business incidents in real time, protect revenue and keep customers happy.

As one of the biggest names in technology right now, anything related to AWS is a big deal. We just got back from the sold-out 2017 AWS re:INVENT conference, held this year in Las Vegas, NV. With a record-breaking attendance of more than 40,000 people, it was filled with representatives from tech luminaries such as Accenture, Intel, Deloitte, Salesforce, VMware, and more. Re:INVENT has always been an event for learning. It provides the opportunity for attendees to familiarize themselves with flagship Amazon products such as EC2, S3, Redshift, and more. The hands-on training, networking opportunities, and breakout discussions have always been an invaluable opportunity for professional development – plus there’s always the excitement of being privy to new AWS announcements. As such, we were thrilled to be among the first to know about what, for us, is among the most exciting new developments with AWS – the AWS Machine Learning Competency. The AWS Machine Learning Competency: A Partner Ecosystem focused on AI and ML The market for artificial intelligence (AI) technologies is flourishing. While a broad set of important technologies is emerging, some of these technologies are still in their early stages. The AWS Machine Learning Competency showcases the industry-leading AWS Partners that provide proven technology for a variety of use cases that will help companies deliver ML at scale. Rigorous Testing and Tough Admissions Process AWS set a high bar for admission to the AWS Machine Learning Competency program. To join, enterprises undergo a strict validation of their capabilities, demonstrating technical proficiency and proven customer success. APN Partners must also complete a technical audit of their ML solution. As the AWS Machine Learning Competency datasheet notes, the machine learning competency is meant for partners whose bread and butter is providing machine learning services to their customers, rather than partners who simply use machine learning in the background of their solutions. Out of 150 companies invited to apply for the competency, only 17 were selected in three categories: Data Services, Platform Solutions, and SaaS/API Solutions. And yes, in case you’re wondering why we’ve spent so long talking about this, Anodot was one of just six companies selected under the SaaS/API Solutions category. “Not every machine learning problem requires starting from scratch and building a custom solution, and not all of our customers have access to a dedicated data science team with the time and expertise to build a production workflow for large scale predictions,” said Joseph Spisak, Global Lead for Artificial Intelligence and Machine Learning Partnerships, Amazon Web Services, Inc. “We are delighted to welcome Anodot to the Artificial Intelligence and Machine Learning Competency Program to provide off-the-shelf machine learning solutions that can help speed time to market and bring intelligence to any application.” Unforgettable Event of Amazing proportions We  got to hear VP of Amazon AI at Amazon Web Services, Swami Sivasubramanian, share his insights: "Our goal is to put machine learning capabilities in the hands of all developers and data scientists. The other thing we are excited about is the API services, where people who don't want to know anything about machine learning but want to benefit from the analytics capabilities. I am really excited about all the applications that we can use to improve our everyday life using machine learning." Finally we rounded out our visit with a round a virtual golf and hung out with this great guy, our main contact at AWS, who has a custom tailored Pac Man suit.

Explore how this ride-share leader uses Anodot to identify business risks in real time.

Imagine lining up before sunrise on a cold November morning the day after thanksgiving, waiting in line to buy up to hundreds of dollars in discounted goods – and then being informed that the retailer couldn’t accept your credit card. Would you shop there again? This nightmare scenario is what happened to Macy’s shoppers on November 24th, 2017, aka Black Friday. Around noon, overcapacity issues shut down the retail giant’s payment processing systems, turning them into a cash-only enterprise for up to six hours on one of the busiest shopping days of the year. The blow to company revenue may be calculable, but the damage to the company’s reputation is not. How can companies avoid these expensive and embarrassing disasters? Glitches Are More Common Than You Think Retailers are plagued by glitches, big and small. Even Amazon, probably the most technologically sophisticated ecommerce platform on Earth, is not immune to the occasional error (Free Echo Dot’s anyone?). So how are smaller and less tech-savvy retailers expected to cope? The glitches they encounter could include: High-volume traffic overloading payment processing servers Orders in online shopping carts failing to complete Incorrect pricing for goods sold online or in stores Mistargeted online advertising Missing opportunities to upsell or cross-sell customers based on faulty analytics Did you know that customers are up to 3x as likely to post a negative review of your company after a bad experience? And that 80% of potential prospects will desert your company if they read negative reviews of your products and services? Though it’s hard to put a price on the loss of reputation, the lost sales do stack up, averaging $250,000 per incident during Black Friday and Cyber Monday, and $40,000 per incident during the rest of the year. Smart Error Handling Demands a Proactive Approach Until recently, retailers had few options when it comes to preventing failures in their ecommerce platforms or their brick-and-mortar stores. The traditional approach was to wait for something to break, and then fix it as quickly as possible. In Macy’s case, “quickly” was about six hours. That’s not an acceptable speed. Another option is to try to track everything that is happening with dashboards and alerts, which quickly grows out of hand – how do you know where to look on your hundreds of beautiful visualizations? Which alert is meaningful when you get hundreds every day? One of the difficulties in these scenarios is that similar-looking errors might have diverse causes. Your cash registers might not be able to process credit cards, but that could be due to a failure in any number of separate applications. We refer to these errors as “micro-glitches” – multiple, nearly imperceptible failures in multiple locations which accumulate to cause spectacular outages. Going back to Macy’s, according to news reports they suffered an overcapacity outage of their credit card systems. Even this kind of outage comes with its own subsidiary failures and forewarnings, however. Systems as theoretically robust as the payment processing system for a major retailer don’t fail all at once. There are cumulative warning signs, such as an increasing trickle of transaction failures, or additional latency during card transactions. The ability to detect and interpret these warning signs could have meant averting a system failure at 10:00 AM, as opposed to trying to recover from a crash at noon. Maybe I’m stating the obvious here, but it’s much better and easier to prevent a failure that you identify early than to try to recuperate after the fact. How Anodot Can Help Where does Anodot fit in? Our AI-powered analytics gives companies the ability to collect and interpret real-time time-series metrics from across their payment processing systems and every other internal and external system, letting them detect and mitigate failures before they begin to drain their revenue and reputations. This can protect companies from the types of disastrous events that Macy’s experienced on Black Friday. To learn more about Anodot, and how our technology can give early warning capabilities to your ecommerce platform, sign up for a free demo today.

Find out how Anodot’s business incident detection system automatically alerts the mobile gaming company to any changes in their business data streams.

Explore this white paper prepared by EMA analysts on five use cases for applying AI analytics in telco services.