Courtesy of B-eye Network

Mario, let's begin by having you tell us why you founded Metric Insights.

Marius Moscovici: Prior to founding Metric Insights, I was a data warehousing practitioner for more than 20 years, first through a consulting company that I founded and then as the director of data warehousing at Linden Lab. During this time, I found myself continuously frustrated with the business intelligence tools I was using. I saw huge sums of money being spent on collecting and organizing valuable data in data warehouses. Yet time and time again, I would see these initiatives fail in the last mile in delivering meaningful insights to users. So I started Metric Insights to build the BI tool that I wished I had all those years as a BI practitioner, a tool that supports rapid implementation while at the same time providing information through a user interface that businesspeople love to use.

Well, Marius, your website indicates that Metric Insights turns data into actionable intelligence. Can you explain to our audience what you mean by actionable intelligence?

Marius Moscovici: Actionable intelligence is information presented with sufficient context to allow a business user to take meaningful action. So let me give you an example. If I tell you what your sales were yesterday, that just gives you data. To make the information actionable, you'd want to know not only how sales are trending but also whether external events - for example, a currently running promotion - are impacting those sales. You'd probably want to know how the current sales trend will translate into an end of month number, and how it's trending against your sales goal. Actionable intelligence is really about enriching data with context and meaning. It's about transforming data into insight that businesspeople can use to take timely and impactful actions.

How does Metric Insights provide the deeper context necessary to make information actionable?

Marius Moscovici: Unlike most reporting tools that simply visualize the data, we collect and store the aggregate metric data in our repository. We then perform statistical analysis on all the metric data to deliver tailored alerts to users. So users simply indicate which metrics they care about, and the system alerts them when there's a significant change in those metrics. Users can also compare any metric to a target, see projections for that metric, and view any external events that correlate with a particular change in the metric.

Finally, we enable a robust collaboration layer by making it easy for people to capture and share their insights. There's usually someone in your organization who knows exactly what has caused a given metric to change. If that knowledge can be captured and associated with the relevant data point in the chart, it becomes incredibly valuable to others who are trying to understand the data. What we're able to accomplish is take these conversations about metrics that typically happen in email and tie those insights directly to the relevant data point so users have the necessary information right at their fingertips and can take action.

Marius, do you have the ability to alert the end user to a specific time to take action?

Marius Moscovici: Alerting is an integral part of a solution. If you think about it. in most organizations, 99% of the metrics that are out there probably have not changed in a significant way from yesterday to today. But there's that 1%, often less than 1%, that has some kind of a significant change that should be acted upon. Our application allows users to indicate broadly what they're interested in. Then it monitors those areas of interest. As soon as there's a metric in those areas that has changed in a significant way, that user is automatically notified that there's something going on. An email is pushed to that user to pull him into the application where he can do deeper analysis and understand what's going on.

Excellent. Now organizations are doing a lot with what we call "big data" systems like Hadoop alongside their traditional data warehousing environments. How is Metric Insights addressing customers' big data analytic challenges?

Marius Moscovici: Well, one of the greatest challenges that's related to getting value from investments and big data is how to make sense of the information that's collected. These systems often use a different technology stack from traditional data warehousing systems. And as a result, this creates a data silo problem where it becomes difficult to combine analytics across big data and traditional data warehousing environments. So we address this problem by first allowing our customers to source from their big data systems using native data access methods. For example, in the case of Hadoop, they can use Hive or Pig to collect the data. Then, they can easily combine this information with metrics pulled from other sources. They can, for example, combine website traffic data that they've collected from their Hadoop environment with sales data that's collected from an ERP system, and they can do this in a couple different ways. They can look at this information on the same charts so they can understand how trends correlate between data from these different sources. Also, they can create metrics that are formulas based on these different individual source metrics. So for instance, they can create a new ratio that computes your sales per website visitor by just building a formula that divides the results from the two metrics.

Now, let's look at another major trend - mobile BI. How does Metric Insights address the needs of mobile users?

Marius Moscovici: We've found that mobile BI means something completely different to users of a tablet versus users of a smartphone. On a tablet, users typically expect a fully featured BI application so we deliver an experience that's essentially identical to what they see on their desktop. On the smartphone, on the other hand, we've implemented an application that is tailored towards on-the-go information consumption. Users can quickly see alerts and browse newly updated charts. Customers don't want to maintain a completely different set of mobile analytics from their desktop reporting. That just creates a huge maintenance nightmare so we've designed Metric Insights so that the same chart can be deployed to the desktop, Web browser, tablet and smartphone without any modifications. We believe that using a build once and deploy everywhere approach and providing users content on every form factor based on their specific interest is the best approach. It makes mobile BI a critical capability that's seamlessly integrated within the overall BI user experience.

Marius, from an implementation perspective, how long does it take to implement Metric Insights at an enterprise?

Marius Moscovici: Well, we have created a very agile environment where it's very fast to build - you can build incrementally. With the traditional business intelligence approach, you often need to define a complex metadata layer. Only after you've fully defined the metadata layer can you build reports, and you're looking at days or weeks to be able to start with analytics using that approach. With our approach, the reports and the metrics are themselves the metadata, which means that you can be up and running within a few minutes after implementing the application. In fact, typically within the first half hour after you've implemented the Metric Insights application, you can start developing metrics. You can have a handful of metrics and reports up on the dashboard, and you can then start expanding and evolving that over time. It allows the ability to deploy quickly when a new data source comes online. You can then analyze that data and provide broad-based consumption of that information. Then you can iterate over time and refine that infrastructure without having to incur a huge upfront investment of time.

Marius, it's  been great discussing Metric Insights very interesting, agile solution. Many people talk about their frustrations with BI tools, but few do anything about it. I am glad your experiences led you to develop a solution that will be a great benefit to all BI practitioners.