Machine Learning vs Rule-Based Content Management Systems
Karthik Guruswamy, Data Scientist at Teradata, got us thinking about how machine learning is replacing rule-based systems in how we manage and use enterprise content, knowledge, and ideas. Karthik explains that for the last 25 years rule-based systems have typically been built along the forms of:
If X then do Y or else if P then do Q etc.
These types of rules are extremely easy to understand and easy to code. But, Karthik points out, things quickly get out of hand: “When a system gets operationalized, one starts with 100 scenarios with 100 rules to handle it. As time goes by we encounter more and more exceptions and start making more rules to keep exceptions under control. What’s wrong with this approach? Think US Tax Code – things get unwieldy and cumbersome over time.”
Rule-Based Content Management Systems
Content Management Systems have traditionally been dumb boxes that were designed to make it easy to store information and didn’t give much through to information retrieval. In the past few years rule-based content management systems have emerged – think a system that leverages location software to prevent a rep from accessing content that is not intended for a specific geography, or a system that includes a quasi programming language that admins use to build Access Authentication Authorization (AAA) rules to manage content. While well intentioned, both the traditional storage approach and now the rules based approach to sales content management have fallen flat and fail to solve the problem of delivering “the right information at the right time in the right format and in the right place to assist in moving a specific sales opportunity forward.”
Content Management and Sales Enablement
Many companies built and implemented a well thought out place for sales reps to find content. These solutions provided some small scale value but the problem sales reps face of finding the right information at the right time still exists because, as is always the case, different repositories, portals, and tools across the enterprise house the diverse and relevant pieces of content designed and used for very specific purposes, and it’s not possible to keep these single-repository solutions updated with all the most relevant content, and it’s also not possible to keep the system up to date with all the rules required to manage content access.
Guruswamy says “rules based systems are like dinosaurs in the big data world – the volume, velocity, complexity and variety of data makes it near impossible to do well. Increasing number of false positives and negatives can wreak havoc in your operational systems with no useful actionable results …”
If Keeping Up To Date With All The Rules Doesn’t Kill You, The False Positives Certainly Will
Now think about this in relation to sales content management. Rule-based system will work only if you know all the situations under which decisions can be made. But how is this possible in today’s complex world where buyers expect messaging, collateral, and presentations, really everything they hear or see, to be tailored to their specific industry, geography, company size, solution needs, etc? It’s not possible.
Rule-based systems try to predict all the content needed within a sales cycle, so reps aren’t forced to go hunting for the right stuff (or forced to create their own).
In a rule-based system a set of content might be loaded into a CRM, with relevant content attached at each stage of the selling cycle. This simplifies things for the rep: as she promotes an opportunity to the next stage, the relevant content for that stage is attached right there at the opportunity level. She can grab it for her next meeting without having to go search or email people asking for the right content.
But what happens when the rep faces a scenario they haven’t encountered before? Unique situations arise all the time in B2B sales, it’s not possible to predict everything, and today’s world can’t be simplified down to personas. Rule-based systems can’t handle this and force reps into using material that was designed for a completely different use case. If a rep is forced to do or use something, and if the results aren’t perfect, a rep will never use the solution again. If programming an infinite number of rules doesn’t get you, it’s the false positives that will kill your prescriptive approach to sales content management.
Machine learning – A new way
Modern technologies allow a new approach:
- Synch with real world content and data, and then prompt users with the most relevant content based on explicit signals such as search and implicit signals such as CRM activity.
Rather than re-creating another repository, a modern machine-learning approach would connect to all existing repositories and auto-discover content and metadata. This approach abstracts the complexity away from implementation, keeping every underlying content repository in sync in real time, drastically reducing the amount of manual effort required to maintain the solution compared to a rule-based repository approach.
Most importantly, this abstraction of the complexity away from the user, while providing more relevant content recommendations – drawing on a larger corpus of content with a smarter relevance engine – means users actually use the platform and can count on the results.
Very simply put – the benefits of a machine learning approach to content management trump rule-based because of it’s convenience, scalability and low operational cost – especially as content these days is unstructured, structured and semi-structured. Machine Learning has the ability to measure effectiveness and improve itself by changing algorithms or tweaking the weighting of different inputs that feed into the algorithms.
To put it in Guruswamy’s words “Machine Learning system removes the manual task of classifying and tweaking rules each time. Fixing rules manually over time is like fixing bugs as the code gets bigger – problem becomes harder like adding to a house of cards.”