Composable Analytics - Creating Actionable Insights

The nature of reporting analytics and bi-insights has changed significantly in the last few years because of the changing nature of how solutions are capturing data.  Today’s enterprises have, in many cases, orders of magnitude more information about how things are working within their environment than just a handful of years ago.  Data that includes customer engagement, system log data, IoT and device-based events, and location and other telemetry data types all contribute to an insight at the moment of need.  Those who can work dynamically and stitch together this data into insights that make a difference in a short period will thrive beyond what their competitors are capable of.

These massive amounts of data collected across multiple sources form the backbone of improving process, design, and delivery along every measurable dimension in the enterprise.  The challenge is not data access or visualization; it’s leveraging all the pieces into a shared data insights platform that everyone in the organization can leverage.  Then, those insights can be integrated into the operational model so that the people who benefit the most can get it in real-time.

Using traditional bi-analytics is a typical reaction; information is cobbled together using a manufacturing assembly line approach that doesn’t scale well when the data is highly fragmented.  Various scenarios can create true insight and understanding; however, they surface long after their usefulness.  Sometimes, it is at the point of regulatory drivers or some other type of forensic need.

Building operational models on top of this can be very risky, especially when other data streams, such as transactional records, process information, and system information, need to be part of the picture.  Having all this data at your disposal, yet being challenging to stitch together, can introduce risk into the operating model and degrade the organization's capabilities.

In the following few sections, we will look at how people can use composable analytics to bring together just-in-time information, forming the backbone of a data-driven organization.

Often, BI platforms, data warehouses, and data lakes are the one place where everything has to come together to create understanding.  There’s nothing inherently wrong with this approach if the need is appropriate.  These approaches serve well in foundational reporting and after-the-fact analytics.

The central question to answer in the race to improve is: Does embedding the insight into a process for it to be used now or within a report that somebody can look at 30 days after it happened?

BI platforms have a legacy in the enterprise that goes back decades and have different levels of capability, from simple databases to OLAP and OLTP-based solutions to highly specialized solutions like ODSs, then to data warehouses, data lakes, etc.  Any of these can be part of an operational scenario where composable analytics are used.  This paper looks at composable analytics as the result of a set of distributed and fragmented data coming together at the moment of need to provide insight.

This traditional architectural or design approach has been valid and does hold a place in the enterprise design model for specific use cases.

How long does it take for an insight to be acted on?

Needs define the time to respond when action is needed. Organizations create hundreds and thousands of signals around a single event every moment.  Signals in the aggregate can create a view of what is happening and then act on it.  The unlock here is not just to surface the insight once but to repeatedly inject it into the point at the right moment and then use those process learnings to improve operations further.

How do we utilize entirely all of the customer data value?

The challenge for many organizations is to get the most out of their customer data.  Combined factors that make this difficult are working with organizationally distributed data and navigating through the organizational dynamics to collaborate on a specific need - spanning operations, technology, and business teams.  As mentioned above, the traditional approach is to have an internal creator-consumer data analysis model.  Organizations looking to become insight-driven must move beyond this approach into a more cohesive process.

Is the complete 360 view of information accessible?

Operational data needed for composable analytics is located across the system landscape - often partially duplicated and then enriched for local needs or embedded within a process or black box system. Many organizations have tried bringing this data into a centralized repository.  However, it doesn’t keep up with the changing nature of queries or doesn’t fit into the data models easily.

Put It Forward created an intelligent automation platform for organizations with the flexibility to work their way.  A single, easy-to-use interface integrates systems, enables process automation, surfaces deep insights, and connects them directly to the user, all through a no-code platform that anyone can use.  These analytics bring teams together, spanning marketing, revenue, operations, finance, HR, and IT, enabling them to scale their great ideas, which make a difference.

Not everyone can be under the same roof all the time - distributed teams, disconnected processes, and data lock-in applications are the limiting factors of many organizations as they try to reach their maximum potential.  Put It Forward combines all of these through a standard user interface and infrastructure.  This data platform infrastructure enables people, groups, and teams to work and learn together quickly while getting the most out of insights as they emerge.

This approach is not a copy of your data into yet another repository or requires some heavy infrastructure footprint to enable.  It works with the data where it is, assembles it into valuable structures, and then delivers it to the ultimate consumer.  This approach controls costs, improves speed, and unlocks the scale of execution.

Any organization looking to scale with composable insights by leveraging their current IT investments has a new solution approach, including delivering insight scenarios from revenue creation process automation to supplier and partner management.

The following are some use cases of how Put It Forward's Siteline can help those looking to leverage better insights through composable analytics.

Next Best Customer:

The flip side of the “next best offer” coin is identifying which customer to focus your resources on.  The configuration-based AI algorithms of Put It Forward help you zero in on which customer you should focus on.  Working with all engagement data across various platforms, it becomes possible to surface predictive insights such as the probability of customer acquisition and help appropriately direct resources.

Revenue Forecasting and LTV:

Put It Forward can tap into past and present customer behavior to help accurately forecast future revenue and LTV of every prospect.  This data can be surfaced in a report or at the individual contact level, enabling the person focused on them to understand what product or service will be the most valuable to offer.  The power of this composable analytics solution is that it allows you to consolidate your reporting into a single location and push out high-value analytics to the end user. 

Churn Risk Identification:

The most expensive customer isn’t the one who overutilizes your resources or takes up time with support - it’s the one who doesn’t use anything and abandons your offer after the sale.  Using the same AI algorithms to identify your next best customer, the Put It Forward customer data platform lets you determine which customers have the highest churn risk or propensity not to purchase another product or service.

Customer Onboarding:

Working with customers efficiently at scale is critical to operational success in every organization.  Within Put It Forward, several advanced analytics options can help you gain insight into the customer journey.  These help identify where process blocks are for a customer, where bottlenecks are occurring, and when you’re dependent upon the customer fulfilling a commitment on their side.  Use these insights to trigger process flows or come from the process flow where there is a blockage.  You can surface metrics in reports and dashboards, which show the efficiency of operations and where opportunities for improvement exist.

Identity Resolution:

Every operational team spends time figuring out who is who and what belongs with what.  Another way is to understand which customer is related to which account, which employee is related to which device or issue, and which support tickets are related to common problems.  Each point has a unique identifier, which is often uncommon across the organization.  Within the platform, identity resolution software helps link and maintain relationships between these entities.  This approach enables operational teams to work efficiently across the organization in separate teams or as part of a global structure.

Issue Management and Automation:

Understanding how to trace issues across an organization is a common requirement with identity resolution.  Several issue management solutions fall under the banner of ITSM.  Put It Forward ties the problems in the various platforms and the processes together to automate the flow of information and events across them.  Further, using the data mining capabilities, you can uncover hidden efficiency gains by examining the processes at the event level and turning them into operational metrics used to manage the organization.

Financial planning and analysis process (FP&A)

Central to any organization's resource allocation process is financial planning and analysis software (FP&A).  Solutions like Put It Forward, tie together the revenue teams' projections and operations efficiency gains into the FP&A process to create an end-to-end view of capital and resources used.  This approach helps finance teams accurately model possible revenue scenarios with actual pipeline data rather than working with assumptions continuously.  This data is then linked with other operational data sources to help create a better understanding of where revenue at risk exists and use AI models to help identify its location.

Procure to Pay:

Many analytical data points need capturing through the procure to pay process.  People start with supplier information, payment terms, location, and availability as foundations.  By tying this information with operational data insights, procurement analytics enable category management to identify process and supply risk efficiencies and improve supplier relationships.  Users can use the Put It Forward Process Designer to work with these composable insights in dashboards, reports, or process flows.

Fraud Detection and Control:

one of the lesser-known and essential functions of finance teams is identifying and controlling fraud in the system.  Fraud detection isn’t limited to point-of-order credit card issues but can run through an organization's processes.  Leveraging the machine learning and AI functions on the “Put It Forward” platform, teams can work with massive amounts of seemingly disconnected events to identify possible fraud in their processes and then send alerts and notifications to the people who need to act.

The Put It Forward intelligent automation platform enables the IT team to make the most out of their strategy for composable analytics by giving them a single solution to work with.  You can reuse this solution across multiple departments and parts of the organization over and over.  Which in turn helps surface its insights and values, such as:

• No code solutions for composable analytics that the business user can leverage
• Centralized management of the systems with decentralized use
• Cost reduction of highly specialized skill sets
• Operationalization of business processes that can scale across the organization
• Architectural fit and ability to leverage existing IT infrastructure without new overhead

Unifying the scattered fragments of information across an organization can help scale great ideas and avoid the risks of lost or unused data. Consolidating this information can improve efficiency, inform decision-making, and increase profitability. Combining all the scattered information is essential to maximize potential and drive success.

BI and composable analytics are two different and complementary approaches to understanding what is going on in the organization.

BI serves the purpose of after-event reporting, and composable analytics enables the automation of decision support.

• The central question to answer in the race to improvement is: does the insight need to be embedded in a process for it to be useful or within a report that can be looked at 30 days after it happened?
• Look at where the insight needs to be placed in the process for it to have the greatest effect and how it is going to get to where it’s needed at the right moment in time.
• Think about the integration process and access to the data that is needed. Is this going to be enabled by highly centralized or hard-to-find technical skill sets, or is it best served by operational teams that are close to the data and process?
• Get alignment on all business initiatives today and potentially around the corner to ask providers how their systems will deal with the multiple data sources, data quality, and AI models to generate insight and how the insights will be shared.