What Are In-the-Moment Business Decisions?
Leveraging real-time intelligence to improve your company’s future
by Monte Zweben and Syed Mahmood
We all make tens of thousands of snap decisions every day, such as where to go to dinner tonight, which route to take on our way to home or office, etc. Some of these decisions may be small or routine, while others may have the potential to alter the course of our lives. We use the information available to us at that given time to make a spur of the moment decision. Let’s call these decisions in-the-moment decisions.
An interesting question that we at Splice Machine have been pondering is, what are the in-the-moment decisions that businesses make in their enterprise applications? With the rise of artificial intelligence and machine learning, businesses are now able to make hundreds of thousands or even millions of intelligent, in-the-moment decisions every day. Online product recommendations, ad placements, and decisions to accept or deny a credit card application are just a fraction of the in-the-moment decisions that businesses now routinely make.
But isn’t “in-the-moment decisions” just a fancy term for real-time decisions that we have been talking about for the past two or three years? Some in-the-moment decisions are made in real-time (like the examples we gave earlier), but other in-the-moment decisions are unique because each industry and business defines its own in-the-moment decisions which may or may not be real-time. For example, when an insurance company decides whether to pay a claim or route it to an analyst for further investigation, there is a time window of several days or weeks which defines that company’s in-the-moment decision. In this case, the decision is not in real-time, but instead within the desired service window that is designed to maximize customer satisfaction and protect the company’s bottom line.
Time is just one of the many variables that define a company’s in-the-moment decision. For a mobile ad placement company, it’s not only critical that the ad is rendered as soon as the page is rendered. Factors like context, location, and weather all play a vital role in determining whether the ad is going to resonate with the audience and convert into clicks.
Now that we have an understanding of in-the-moment decisions, let us tell you how we have been thinking about this topic at Splice Machine. We believe that in-the-moment decisions should drive intelligent actions that businesses take in their mission-critical applications using data at scale. Or in other words, intelligent actions at scale are a function of in-the-moment decisions.
Unlike individuals, businesses cannot make in-the-moment decisions based on gut feeling or intuition. They must rely on all the information at their disposal to make them. That’s why businesses are so excited about the potential of artificial intelligence and machine learning to make in-the-moment decisions automatically to drive intelligent actions. Given the volume of data stored behind Enterprise firewalls and in data warehouses and lakes, and the petabytes of new data from IoT devices, server logs, and clickstreams, using artificial intelligence to make decisions is not just a luxury. It’s imperative for success.
The question on every executive’s mind must be, what are the pieces of the puzzle I need to have in place before my enterprise can use these moments to make intelligent decisions? We believe that your organization needs to have three competencies that serve as the foundational elements for making in-the-moment decisions. These are:
- Access to the most recent available data in ML models and the ability to take action at the right time
- Continuous training of ML models on the most recent data as the market conditions change constantly and your model must adjust accordingly
- Enable data scientists to freely iterate and experiment with new features representing moments, and to compare and contrast models efficiently to continuously improve and put the best models into production
Let us explore the role of each of these in making in-the-moment decisions.
Access to the Most Recent Available Data in ML Models and the Ability to Take Action at the Right Time
It is fairly intuitive to think that if you are going to build a model that is going to predict a future event, you must build it using the same data that contains all or most of the attributes that describe that specific event – the feature vector in data science terminology. The more updated data that you have at your disposal, the higher the likelihood that it contains the features that are predictive of the “moment”. In the past, due to data storage and computational constraints, businesses could only afford to use a slice of the data to build their models and in most cases, this slice of data contained old information. Even though data storage restrictions have been largely lifted due to the decline in storage prices, the current data infrastructure in most enterprises is not conducive to making the most updated data available to data scientists to build their models. This infrastructure typically consists of an operational (OLTP) database and a decision support (OLAP) data warehouse that has been duct-taped together. Others have implemented so-called Lambda Architectures on Hadoop. Both of these infrastructure architectures require large amounts of data to be moved repeatedly and as a result, the data that makes its way to the data scientists is stale.
Why does this matter? Because the models will then make bad decisions. Millions of dollars of fraudulent claims can be paid out inappropriately with little chance of clawing those payouts back. Predictable outages in oil rigs and utility grids can be missed, costing millions of dollars a day. Conversion rates on marketing treatments can suffer, affecting revenues. Patients being discharged from a hospital too soon, only to have to deal with costly readmittance, let alone the safety issues to the patient. The business ramifications of poor models are serious. And data latency is one of the worst culprits.
Having access to the most recent available data to build your model is just one part of the puzzle, the ability to deploy the model at the point of decision and take action in-the-moment is the other. We cover this topic in the modernizing applications section of this blog.
Continuous Training of Models
The second competency that you must have in place to act upon in-the-moment decisions is the ability to continuously train your machine learning models. The reason is that if the model is operating on rapidly changing data, the model must be trained much more frequently. For example, if you have built a machine learning model to detect cyber attacks on your network, the chances are that the hackers will change their strategy frequently, and if your model is not continuously trained on new attack patterns, sooner or later, it will miss an important signal that will result in the network being compromised.
The latency built into the current enterprise data infrastructure does not foster continuous training of models as the data required for retraining is simply not available at the frequency desired by data scientists. This forces the model to operate at a suboptimal level and, in certain extreme cases, make predictions that are incorrect.
Ability to Experiment Freely or Feature Factory
When data scientists embark on the task to build a machine learning model, they start with a large number of attributes or features that are predictive of the event that they are interested in. In a sense, this is the process of defining a moment. The features define the representation of a moment with both a real-time context as well as a summary of the historical context. The model then maps this representation into a prediction.
You can think of data scientists as working in a factory and feature vectors as their raw material to shape into a model. In order to build a robust model, data scientists continuously experiment with different data sets and parameters. They try out different combinations of features, often swapping out one or adding a new feature to replace a less predictive one, or using a transformed or derived feature from other features into the new run. Only through continuous experimentation on the feature production line can a data scientist define the in-the-moment decisions and build a model that is accurate enough to be put into production, stays accurate as the world changes, and in fact, continuously improves over time.
Modernize Applications to Make Decisions in the Moment
The question is then once you have identified the in-the-moment decisions for your organization, how do you make them a part of your mission-critical business process. In order to extract true business value from the ML model that is capable of identifying the in-the-moment decisions, you must inject that model into a purpose-built application. These are the applications that have been built and customized over time to represent the unique way in which your company does business and they are your company’s secret sauce. These may take the shape of a recommendation engine on your website, or a claims management or a customer onboarding application depending on your industry. The goal is for these purpose-built applications to make decisions and take action in the moment. We see a lot of companies trying to build brand new intelligent applications as part of their digital transformation initiatives. However, we feel that modernizing existing purpose-built applications represents an efficient and low-risk approach to intelligent applications that often gets overlooked as part of the digital transformation – as we like to say at Splice Machine – don’t rewrite them… modernize them.
Artificial intelligence and machine learning have been described as the oil for the next industrial age, but unless enterprises can successfully use these tools to capture the in-the-moment decisions and leverage them fearlessly to drive intelligent actions, data science initiatives will likely remain back-room activities.
In-the-Moment Business Decisions Webinar
Watch our on-demand webinar to learn about Splice Machine’s approach to making in-the-moment decisions part of your mission-critical business process by injecting predictive and ML models into your purpose-built applications.