Categories
Banking Big Data and Analytics Digital Strategy Insurance

Making Money with Your Data

A practical approach to data management for financial institutions

Much virtual ink has been spilled in last few years about data and analytics. Big data, data science, machine learning, artificial intelligence, neural networks, … – new capabilities keep emerging and dozens of articles on the imperative of data and analytics appear every week.

Unfortunately, many organizations that have invested in data management and analytics are failing when it comes to actually using data to increase profitability. In fact, according to Infoweek, companies are only getting a $0.55 return for every dollar they invest in big data.

This echoes our experience at RedPort, as we’ve been called in to work with organizations that have made significant data capability bets – only to struggle to achieve ROI goals. When we dig deeper we generally find that the organization is suffering from one or more of the following four problems:

1. They haven’t committed to managing with data – and to a culture that supports data driven decision-making

2. They haven’t articulated precisely what they are trying to achieve with their data and data analysis – and how data will make them money

3. When they do find a data driven insight, they don’t effectively instigate a timely ACTION to capitalize on it

4. They didn’t systematically and continuously LEARN from that action so that they can adjust their algorithms and be smarter the next time

Wasted resources

Typical of the early days of emerging technologies, it appears that many if not most financial institutions are wasting money on their data initiatives. This is no surprise to us as over the years we have worked with financial institutions on data related challenges, we have seen failed data warehouse projects, incomplete datasets required for analysis, lack of execution on insights and no ability to continuously learn. Here are some examples:

    • A US banking institution that spent several hundred thousand dollars on a sophisticated data system but never committed to the culture of data management around it – resulting in the write-off of a system and a new quixotic search for something “better”
    • A large emerging market insurance company, falling for the “promise of new technology” wasted several years and several million dollars on a core system replacement that promised better data management – but in the end created an environment that was too complex for their internal team to manage – and too expensive to maintain
    • A super-regional bank that bought hundreds of expensive CRM seats and only implemented a few before leaving the system for a new technology that was more focused on their immediate needs
    • An insurance company that completed a major data driven segmentation effort – only to see it fade into obscurity when its sponsor left the company
    • A US lender that purchased a suite of tools but that doesn’t have the analytical firepower to get the most out of them – or the capacity to execute on the insights they do find

These failures are not new, and similar stories have been documented by others for some time. Recently, experts have cited several reasons for the failure of Big Data initiatives including:

    • Lack of skilled big data experts – “The data scientist shortage is a well- chronicled phenomenon and one that might persist for some time.”
    • Immature technology – “Big-data tools are in their infancy. They require refinement for use by a wider range of business workers — not just highly trained data scientists”
    • Lack of a compelling strategy – “Enterprises often invest in big-data projects without tying these efforts to specific and measurable business applications”

At RedPort we have a different view

While we agree with the experts cited above that the technology needs to evolve and that we need to train more data analysts, at RedPort we believe that often as not – failure to make money stems from internal problems that can be remedied with proper focus and strategy.

In our experience, while some tools are better than others, they are almost never the primary limiter of data and analytics. Instead, internal problems – often solvable – are more often than not the culprit.

Company culture and organizational problems

The root of many problems we run into, cultural and organizational challenges can derail data initiatives before they are even started. Common themes we run into include:

    • Lack of (or inconsistent) senior management commitment to implementing a new technology and data-driven work style
    • Lack of management competency in making data-driven decisions
    • Lack of ability to manage across the functions required to capitalize on data initiatives
    • Lack of communication down the management pyramid of the required change investment

Business case problems

The next set of problems we have experienced focus on the business case. While Big Data is interesting and provides many with a stimulating intellectual exercise – unless management is forced to articulate exactly what they will do with the data (how they will make money) all of the effort will be for naught. Often we find that organizations have:

    • No clear articulation of how data will impact the organization strategically
    • Not developed the business or use-cases that define the problems that will need be solved – and how much money will be made by solving them
    • No clear vision on how ROI will be generated overall and what the priorities are to achieve the desired returns

Skillset problems

Managing data is a difficult task and takes many interlocking skillsets to get right. Often we find that organizations – especially midsized organizations – find it almost impossible to attract and fund the people required to get data right. When we look at this problem we usually find:

    • Nobody able to manage the complexities of data management across disparate systems
    • Nobody able to analyze the data in sufficient depth to draw out the required insights, which requires a deep understanding of the business and how the analysis will drive financial results
    • Deficient ability to attract the many skillsets (analytics, IT, business, marketing, finance, etc.) required to effectively aggregate, analyze and execute on insights – and to knit them into a closely coordinated team

Execution problems

Finally, we find that mid-sized organizations have difficulty executing data projects – and even if they do get the infrastructure together to develop unique insights, they have difficulty taking effective action against their business objectives. Often we see:

    • An inability to coordinate IT projects across multiple systems and capabilities Poorly coordinated implementation across multiple business functions (especially marketing, sales and service)
    • Inability to capture results and report effectively across product, channel and customer segment dimensions
    • Lack of the discipline to systematically learn from the results and improve the actions in the next cycle

Sober and clear-eyed management of data initiatives can achieve real results

Despite the problems we’ve identified, we have also worked with several institutions that have mastered the challenges and found ways to use data – big and small – effectively. Most of these financial institutions have followed a process similar to the following, and over a number of years systematically built the infrastructure, processes, and the teams required to drive returns.

First – they commit to managing with data

Making the commitment to manage with data is always the first step. Only with CEO/C- Suite level engagement to data-driven management – promoting people with analytical skill sets, and encouraging a culture based on data insights – will ROI be achievable. The important take-aways are:

    • This is a CEO/ C-Suite job and should be a key priority of the organization
    • Do not take the cultural change required lightly
    • Expect that some of your managers will not make the “data-transition” and will either actively resist the change or fade away over time
    • Be mentally prepared to maintain focus on data for several years to get it “right”

Second – They articulate exactly what they are seeking to achieve and how much money they expect to make

Before spending a dime on IT or algorithm development, the first step is to build the business or use case for data ROI and specifically, to build out and prioritize the use cases that data and insights will be applied against. Be brutal and force the team to show how much money is actually available if you are successful.

Common use cases include campaign optimization, channel/branch optimization, product optimization, attrition management, channel migration, etc.

Useful questions to ask include:

    • What business problems are you trying to solve? What is the potential monetary return of solving these problems?
    • What priority will you give to solving them?
    • How much data will need to be amassed? How many systems will need to be accessed?

Caution: If one of your managers tells you they will “develop unique insights only once the data is put together” expect to be disappointed! An analysis strategy is required.

Third – They use business needs to drive what data they analyze and how they store it

Once the use cases are articulated, the next step is to identify the data required and to organize it in a flexible way. This is where many organizations falter – as most data stores (including many offered by core vendors) are organized around optimization of products, instead of being designed around customers and structured to enable understanding of customer segments, channels, and product dimensions.

    • Start with the data you need for your use cases. Resist the tendency to store everything you can find at once.
    • Most financial institutions should organize data around the end-customer, which is the profitability center

Fourth – Only once they have completed the above do they select their tools

Now we can turn our attention to tools. At a minimum, organizations who are serious about managing data will need to find a set of tools that allows them to:

    • Aggregate the data and store it in a flexible and efficient way
    • Report on the data across products, channels and customer segments
    • Analyze the data – often including model building
    • Execute on obtained insights across e-channels, branches and the contact center • Measure the results

These tools need to be knitted together into an integrated platform that can be easily accessed by decision makers.

    • Despite what some vendors tell you, effective data management will require a suite of tools integrated into a coherent data management platform
    • Try to buy the least number of tools that leave you with the flexibility you need to execute the business cases
    • Increasingly, pre-configured tools that integrate several functions are resolving many if not most of the data challenges financial institutions have – often at much lower costs than traditional providers
    • Be wary when your own team tells you they can “build it themselves”. Unless they have the skillsets and multiple platform builds behind them there is a high chance of failure.

Example: “Analytics Core”

Fifth – They implement methodically, use case by use case

With all of the pieces in place, it is time to implement. Rather than a big-bang project that tries to accomplish all of an organization’s data goals in one step, it has been our experience that steady, step-by-step implementations are the most effective and lowest risk options. Starting with the ‘least risk/highest ROI’ use case, organizations should gather the data required (generally a contained set of core system and customer data), build the analysis and execution process and ‘get it working’ before moving on to the next use case/data load etc.

    • Start by focusing on one, achievable use case and using it to test the process and tools and to achieve some ROI for the project prior to moving forward
    • Systematically build out additional use cases as they are articulated and prioritized 7
    • Focus on building insights and capabilities in sequence – starting with straightforward use cases like predicting campaign profitability – and moving systematically toward advanced use cases like mining big data or predicting consumer demand elasticity

And whatever you do – avoid these ROI-destroying mistakes

Finally, we focus our clients on avoiding making un-forced errors that kill ROI and delegitimize the project – often before any tangible results are achieved. While these are numerous, there are several that are common and can be easily avoided.

    • Letting an inexperienced IT team run the project – these are complex projects that take experienced teams of specialists
    • Buying too many expensive tools that the organization doesn’t have the competency to utilize effectively – advanced statistical tools are a prime suspect here – as are complex campaign management systems, CRM systems, etc.
    • Expecting a data “guru” to develop profitable insights alone – Top analysts and data scientists are important, but alone they will accomplish very little. What’s more – they often do not have the backgrounds necessary to build data infrastructure and execute on the insights.
    • Building functions into tools that are not their core strength – Such as building data warehouses “under” a CRM or presentation tool like Tableau
    • Hiring consultants with no industry background – There are consultants everywhere that claim to know how to manage data. What they don’t tell you is that to really understand your problems they need deep industry specialists to write requirements and develop the metrics you will need to begin modeling. Focused financial institution consultants are out there – find them.
    • Starting with Big Data – when there is a lot of small data ROI to be had at much lower cost. Value can be created through efficient management of data before a data scientist ever needs to be hired.
    • Building too fast – be disciplined about the use case methodology and avoid big bang projects. If you try to do everything at once you will likely accomplish very little.
    • Bet on finding some exclusive insight – even the deepest AI teams can only expect to get a few months’ head start on their rivals – better to focus on being better at executing on and learning from insights – even if they are mundane

Summary

Making money with your data can be a difficult and challenging endeavor and the risks of failure or expensive misstep are high. However, organizations that are disciplined and follow processes similar to what we have expressed above will over a few years find that they have built the competencies necessary to experience significant returns on their investment. Even more importantly, these will have built a strong base to allow them to accelerate their data management and insight development agenda in the future.

  1. Commit to managing with data
  2. Articulate clearly what data will do for your organization and how it will achieve ROI
  3. Determine what data you will assemble and how you will organize it
  4. Pick your tool set and build on an integrated platform
  5. Implement methodically, use case by use case
Categories
Banking Big Data and Analytics Credit Union Customer Experience 2020 Digital Strategy Insurance Microfinance Uncategorized

Customer Experience 2020: The Emergence of the Autonomous Financial Institution

In the age of self-driving cars, additive “3D” manufacturing processes, and self-flying delivery drones – how can Financial Institutions capture similar opportunities enabled by continual advances in sensors, artificial intelligence and autonomous process automation?

Financial institutions need deep analytics solutions that:

1.  Start by allowing the definition of the desired business goal like growing market share, decreasing customer churn, increasing profitability – or even simultaneous combinations of these strategies.

2.  Use Artificial Intelligence in a self-learning mode, to continually optimize to the desired goal.

3.  Combine Machine Learning and AI algorithms with effective data management – all linked with opti-channel delivery channels. This gives FI’s the ability to execute sophisticated, automated marketing and customer management programs, with few manual processes and minimal human intervention.

4.  Take advantage of more and more value-creating opportunities (sales, marketing, financially accretive service interactions), to be more responsive to changes in the market place, and to significantly reduce marketing expenditures.

To achieve Customer Experience 2020, FIs need to:

  1. Develop insights faster and be smarter about prioritizing their customer interactions to achieve their financial and other goals…
  2. Act on those insights right away and be more responsive to changes in the marketplace – beating competitors to market opportunities…
  3. Automatically learn from the insights and be more efficient and operate more effectively at the lowest possible cost levels…

To learn more about Customer Experience 2020 and automated FI’s download our full white paper on our main website.

To learn about RedPort’s deep learning analytics technologies click the links below:

SmartBanker: Self-Learning analytics and marketing platform for banks, credit unions and consumer lenders.

SmartInsurer: Self-Learning analytics and marketing platform for insurance providers.

Categories
Big Data and Analytics

Analytics: Strategy or Tactic?

You can hardly go anywhere or read anything anymore without hearing or seeing something about Big Data. People who aren’t in banking, insurance retail, or consumer marketing could be forgiven for thinking Big Data is like the second coming of Godzilla or something: “Big Data. It’s Everywhere!” But while high-consumer-based merchandisers have been building their businesses around data for quite some time, financial-services operations are just starting to dip their toes in the proverbial water, relatively speaking.

We could engage in the philosophical debate about whether data analytics constitutes a strategy or a tactic. However, there may not be a right or wrong side in that debate. In fact, some organizations might choose to enjoy some of the advantages of data analytics as a tactical approach, even as they re-tool their operating models to make analytics more of a strategic underpinning. Here’s how:

  • Tactical approach: Let’s say you decide to track the performance of one particular product. Maybe it’s an insurance line of business. Maybe it’s consumer-loan offering. And let’s say you have limited tracking capabilities; but you can track the customer segment to which the product is sold, the channel through which it’s sold, the geography in which it’s sold, and the person who sold it. Just having that limited amount of information would enable you to know if the product was a potential winner or loser.
  • Strategic approach: Given what you learned from your tactical experiment, you may decide to extend data analytics farther across the enterprise — or at least beyond one department or a single line of business. At the far end of the spectrum, you may opt to aggregate and analyze data from all your core and ancillary systems and data sources to get a closer look at overall operational performance, to better understand customers, to recognize successful products, to see trends that lead to opportunities, to identify and recognize high-performing employees, to make better marketing decisions, to refine pricing, and to decrease losses and expenses.

Clichés are True

The bottom line is this: You can employ analytics productively as a tactic. Every time you do, you’ll learn something. But you’ll employ analytics more productively if you see your way to employing it as the strategic underpinning for operations and decision-making.

According to the cliché, you can’t manage what you don’t measure. And we’ll agree that there are metrics at least as important as numbers. But if you employ analytics strategically to monitor, measure, and manage all your numbers, many other metrics will take care of themselves.