Hadoop-based DBMS for Machine-Guided Analytics

Xurmo is a Big Data analytics platform that determines the natural structure of raw data and stores it in a patented format. The platform then allows instant Search-guided Query so that complex analysis can be performed without any schema design and use-case specific data models.

HERE’S WHAT OUR CUSTOMERS
HAVE TO SAY
It's Instant
It's Flexible

It's Intelligent
It's Scalable

CASE STUDIES

  • Insurance

  • A leading global Insurance company wanted to identify the customer segments most likely to churn and take steps to mitigate it.

    Read how Xurmo solved the problem.

    A leading global Insurance company wanted to identify the customer segments most likely to churn and take steps to mitigate it.

    The client wanted to reduce the number of customers that were churning out of their programs. To this end, they hoped to cluster customers into unique segments; and then identify the customers most likely to churn within those segments. Our solution involved the Clustering, Classification and Visualization modules of Xurmo. We created a clustering model that allowed our customer to segregate their clientele into eight distinct segments. Classification was used to creates a model that could predict, with a certain estimate of probability, whether a customer would churn or not. Once the models were created, they were tied to Visualizations and presented as an end user application (Dashboard). The Client’s analysts used this to help the business take pre-emptive measures to reduce customer churn. The entire project including integrating various data sources, building the churn model and an application interface for the business user was concluded in 12 days.

    A large Oil & Gas multinational faced a problem of employee attrition and needed to take steps to mitigate it.

    Read on to learn more about Xurmo’s solution.

    An Oil & Gas major was looking to make their production more efficient by analysing information from various industry sources.

    Read how Xurmo provided the perfect solution.

    A large Oil & Gas multinational faced a problem of employee attrition and needed to take steps to mitigate it.

    For this client, employee attrition was a major problem. They were unable to retain trained employees for long and pre-emptive efforts to retain them were failing. Using Xurmo, the client was able to build a model, which could predict which employees had a high probability of churning and which retention program could work for that employee. The model was built in under a week using data points that were extracted from a multitude of the client’s data systems, including emails, forums, HRMS systems, etc.

    An Oil & Gas major was looking to make their production more efficient by analysing information from various industry sources.

    The client wanted to derive intelligence about their oil production. They also wanted to track the various business deals occurring in the domain. Xurmo was deployed to integrate data from various sources including company databases and sources in the public domain. Xurmo connectors pulled in relevant structured and unstructured data into the platform and various text analytics modules were employed to determine relevant data sets for analysis. The output was rendered in a custom built application interface which gave near real-time information to business users.

    A leading Consumer Product Company in India wanted real-time analytics on data from a multitude of sources, with different periodicities, formats and structures.

    Here is how Xurmo addressed the challenge.

    A leading soft drinks bottler in the US needed demand prediction at the retailer level.

    Read more about the Xurmo solution.

    A major Consumer Product Company in India needed to catalogue the huge bank of images scattered across various media, databases & formats.

    Read on to know how Xurmo enabled this.

    A leading Consumer Product Company wanted to create a unified system that stored the advertising production and vendor costs scattered across the marketing department employees’ files.

    See how Xurmo solved the problem.

    A Consumer Product Company wanted to standardize quality-check data from various factories and automate Business Intelligence requirements.

    Learn how Xurmo implemented the solution.

    A Consumer Products Company needed to track competitor prices of merchandise across regions.

    Here’s how Xurmo offered the solution.

    A CPG client wanted a solution that crunched time needed for their planning & forecasting from several months to a matter of weeks.

    Here’s how Xurmo did it.

    A leading Consumer Product Company in India wanted real-time analytics on data from a multitude of sources, with different periodicities, formats and structures.

    The client’s data pertaining to Sales, Trade and Merchandising, Brand Activation, Media, Logistics, Household Panel, Market Research, etc. was in excel files which were stored in a hierarchical format in a file system. The excel files had different table layouts, multiple tables per sheet, merged cells, discrepancies in data conventions, and information stored as meta data (sheet names, file names etc.). It took up to three months to prepare a batch of data for analysis, and with data being updated periodically (in some cases weekly), the analytics teams were constantly working with outdated data. Using Xurmo, they were able to ingest all excel data and create stacks of transformation operations which could be applied on any new data. This allowed the client to reduce the effort of preparing the data from about three months to a few minutes.

    A leading soft drinks bottler in the US needed demand prediction at the retailer level.

    The client was able to predict demand at the distributor level with some level of success, but failed entirely to forecast demand at a retailer level. Additionally, they wanted to include data points from social media for the analysis. Using Xurmo, they were able to extract data from their existing systems, from Twitter and from news, sports and weather related websites. Using the Machine Learning algorithms available in Xurmo, various models were created. The client was then able to use data points from these different sources to build a demand prediction model that could foretell the demand at a SKU level for each retailer much more accurately. The process of integrating data from various public and private sources, creating a demand model at retailer/ SKU level and rendering it as an application functioning in real-time was concluded in 3 weeks.

    A major Consumer Product Company in India needed to catalogue the huge bank of images scattered across various media, databases & formats

    Xurmo’s custom application helps in cross leveraging brand packaging ideas across teams, educate employees on any given brand or agency history, as well as easily review agencies based on past engagements. This App helps the branding team to quickly search relevant images from various documents and PDFs based on either a keyword based search or through pre-defined filters. To enable this, a custom connector for Xurmo was built in two days that automatically mapped images with relevant text from the file, and the file’s meta-data.

    A leading Consumer Product Company wanted to create a unified system that stored the advertising production and vendor costs scattered across the marketing department employees’ files.

    The client’s marketing department was having trouble identifying, reviewing and updating records pertaining to film costs, production houses, directors and ad agencies costs as this information were in excel files kept amongst various people. They wanted to move to a system that would allow them to store all the information, and update it periodically from one place. Using Xurmo, not only were they able to quickly build an application that enabled this but also use the analytical capabilities of Xurmo to analyse the costs of marketing campaigns on a multitude of parameters.

    A Consumer Product Company wanted to standardize quality-check data from various factories and automate Business Intelligence requirements.

    The data pertaining to quality checks at various stages of the manufacturing process of each product was stored in a series of Excel files, maintained individually by each factory. This meant that data sanctity was non-existent, manufacturing specifications weren’t being met. Also, the analysis process (while not complex) was tedious; because it meant that these files would have to be transferred into a predefined format, collated, cleaned and then processed for various insights. The client wanted to setup a system that would simplify, standardize and enforce specifications for data entry & automate the BI requirements of the management teams. Using Xurmo they were able to implement a system to do this resulting in better quality, reduced effort and increased accountability.

    A Consumer Products Company needed to track competitor prices of merchandise across regions.

    Xurmo’s App helps the merchandising teams to track a product’s price premium with respect to prevailing market rates in various geographies and the top competitors in a particular geography. Xurmo was configured to extract data from a set of webpages and the Application automatically linked it with the correct product, margin calculation etc. This helped the merchandising team to take decisions on price adjustments on a daily basis instead of the then prevailing monthly frequency. The application also allowed users to perform cost projections based on minute changes to the prices at an SKU level.

    A CPG client wanted a solution that crunched time needed for their planning & forecasting from several months to a matter of weeks.

    Typically, this client required three months of the financial year (FY) to conduct their planning and forecasting activities for the next FY. The process was iterative, with forty people from various teams giving inputs for over twenty sources of data. This data was spread across ten plus product categories, and thousands of SKUs. Tracking changes of data on this scale was a nightmare. Different teams made changes to their data in isolation, and any correction was made only after the forecasting engine threw up errors upon working on the entire data. This process would have to be repeated even for the smallest of changes to forecasting. With the Xurmo Platform, the client was able to implement a system that reduced this effort to a couple of weeks by automatically ingesting the various sources of data, finding errors and notifying the responsible users before consolidating the data. This meant that the users could now focus on optimizing their planning process with a larger number of parameters in consideration to maximise profit.

    A leading Indian bank wanted to reduce processing time for product recommendations from over 3 weeks to less than a day.

    Here’s how Xurmo achieved it.

    A leading Indian bank wanted to reduce processing time for product recommendations from over 3 weeks to less than a day.

    The requirement for this Bank was to provide a single view of all product recommendations to each of the 28 million customers based on structured (existing EDW) and unstructured data (CRM and Emails). The target was to reduce the processing time for product recommendations from over 3 weeks to less than 1 day. The lengthy SAS Scorecards were easily translated using the Xurmo interactive query language. The translated Scorecards now run efficiently on a distributed environment. The various Scorecards are tied to a simple application interface built on the Platform to generate the Single Customer View at the click of a button. Xurmo achieved the performance target of less than a day. The application itself was developed in under 2 weeks. It also provided a highly scalable solution that runs on commodity servers rather than requiring expensive custom hardware.

    A Financial Consultancy firm required a solution to leverage industry news for stock price movement prediction.

    Read how Xurmo helped them predict stock price movements with higher accuracy.

    A Financial Consultancy firm required a solution to leverage industry news for stock price movement prediction.

    This client wanted to use industry news to tweak their stock prediction models. Using Xurmo’s Connectors, they were able to extract relevant news articles, RSS feeds and stock price information on various stocks. Using Xurmo’s analytical libraries a variety of trading models were built to offer positional, intra-day and inter-day recommendations to traders.

    A leading KPO firm required agile discovery of globally scattered data.

    Learn how Xurmo significantly reduced both costs and turn around times for the client.

    A leading KPO firm required agile discovery of globally scattered data.

    The client was having trouble managing the information that was being generated by their analysts. Their existing systems were not able to capture insights gleaned by analysts via their experiences. Documents pertaining to projects were stored on individual analysts systems, and there was loss of information due to employee churn. Global operations of the client only amplified the problems caused by these information silos. This meant that analysts were wasting their time re-discovering information on related projects. Using INSIGHT, a Knowledge Discovery application built on Xurmo, the client was able to remove the barriers to the information. Insight was able to parse emails, chats and documents – both on local machines and CMS systems, and other Meta data to be able to recommend the People, Documents and Conversations that could answer a user’s query. This allowed analysts to work more efficiently, thus reducing costs and turnaround times for projects.

    A large publishing house required customising emails with user-specific product offering and relevant subject lines for each customer.

    To know how Xurmo helped, read on.

    A large publishing house required customising emails with user-specific product offering and relevant subject lines for each customer.

    A global media house was getting extremely poor responses to their email marketing campaigns. To overcome this, the client wanted to personalize the email subject lines in order to be more relevant to the user. They wanted to be able to predict which product a customer would be more likely to respond to, and use a product-cum-user specific subject line for the emails. To do this, Xurmo was implemented to integrate data (both structured and unstructured) from various sources including transactional, CRM and email systems. Xurmo’s Machine Learning libraries were used to build a supervised learning model, which analysed historical behaviour of customers to recommend email subject lines customised to product and the user. It took only few days to build and deploy the model in an end-user application.

    A leading Pharmaceutical Company wanted to use publicly available information to understand user acceptance of specific drugs.

    Know how Xurmo helped realising this.

    A leading Pharmaceutical Company wanted to use publicly available information to understand user acceptance of specific drugs.

    The client wanted to analyse publically available information to gain insights about how their drugs were being accepted. Using Xurmo’s HTML connector they were able to extract comments about a particular drug from different Internet forums. As a comment could possibly belong to more than one category of information, they split the comments into sentences. They then built a model that predicted whether a sentence belonged to one of the four defined categories using self-learning algorithms of Xurmo. They also used Xurmo’s proprietary Sentiment Model to classify each sentence into a specific sentiment. This enhanced data set was then fed into an application that could show comments broken down by drug, category and sentiment. This allowed the client to gain insights as to what their customers were saying about each product.

    A leading Pizza brand wished to gauge customer sentiment on competitive products.

    Read on to know how Xurmo made it happen.

    A leading Pizza brand wished to gauge customer sentiment on competitive products.

    The fast-food eatery wanted to assess sentiments of customer with respect to their products and their competition’s products. To do this, reviews and comments sourced from multiple social forums were extracted using Xurmo’s connectors and then classified into specific “sentiments”. Once this was done, the top words and phrases for each sentiment were extracted and presented in an application. This allowed the client to gauge what customers’ opinion and attitude about their own and competitors’ products. This application was built in under a week.

    A market research company took on the challenge of predicting the outcome of 2012 US Presidential elections.

    See how Xurmo helped them get it right.

    A market research company took on the challenge of predicting the outcome of 2012 US Presidential elections.

    Real Politix was an analysis that used social media data to predict the outcome of the 2012 US Presidential elections. Using Xurmo’s Twitter Connector, tweets relating to both incumbents were pulled into the platform, and classified as positive, negative or neutral in real time. The classified tweets were then fed into an end-user application which then aggregated them to predict the candidate most likely to win.

    One of the world’s leading Software Companies needed a solution that automatically categorised user comments in their support forum.

    Click to know how Xurmo accomplished this.

    A Business Intelligence firm needed a scalable solution to deliver advertising click-through performance reports to handle fast-growing data of their clients.

    Xurmo’s platform helped them achieve it. Here’s how.

    One of the world’s leading Software Companies needed a solution that automatically categorised user comments in their support forum.

    The client wanted the ability to categorize comments made by users in their support forum into one of over 6,000 categories across five levels automatically. To automate the process of classifying comments, unstructured content had to be made analysable quickly. Then text processing was needed to break down the text into individual tokens; to apply Feature Reduction and to build a taxonomy. Along with this, machine learning algorithms were required to detect patterns in the volume of data. Xurmo was employed to make unstructured data instantly analysable. Then, using Xurmo’s Text Analytics and Machine Learning libraries, the client was able to build models for various hypotheses, and quickly test them. This fail-fast approach to solving such a complex problem enabled the client to build a model that could accurately classify the comments within 2 weeks.

    A Business Intelligence firm needed a scalable solution to deliver advertising click-through performance reports to handle fast-growing data of their clients.

    The client was using an MS SQL Server based OLAP solution to deliver click-through performance reports across numerous dimensions of ad campaigns. The solution was not able to deliver the same performance on scale to meet the growing data of their clients. The client was looking for a solution that could not only exceed the performance of their existing solution, but one that could scale as their clients’ data grew. Implementing the Xurmo Platform allowed them to build solutions that delivered reports and analytics in less than one-tenth of the time. Also because Xurmo’s platform is inherently scalable, the solution was geared to consistently deliver the same level of performance, no matter what the data size is.

    LEADERSHIP

    A simple query can make a big difference to your big data.


    Xurmo Technologies Pvt. Ltd
    Silver Software Tech Park,Plot No. 23 & 24, II Floor,EPIP 1st Phase, KIADB, Whitefield, Bangalore, Karnataka. INDIA - 560 066
    Call: +91-80-4031 4000
    Xurmo Inc.
    #6, Ballentine Court, Flemington, New Jersey 08822, USA
    Call: +1-908-968-9396

    Mail: contact@xurmo.com