Today's businesses must derive insight from decentralized and multi-modal data near real time (NRT) in order to hyper-personalize customer interactions. There's a lot to unpack in this sentence:
- Businesses must be able to intercept the right information quickly.
- They must know how to tap decentralized information.
- They must be able to analyze information instantaneously.
- They must be able to deliver relevant and compelling recommendations back to their customers.
The days of one-off, stand alone, and asynchronous predictive models are long gone. It's not the analytics that went out of fashion (predictive modeling methodologies are still valid and much used); it's the infrastructure and scale of data that have changed. Point-in-time access to customer profile has proven to be insufficient and often non-representative of the complete picture of the customer. Because of this, businesses feel a sudden avalanche of technical debt, which has them clamoring for companywide upskilling programs to avoid falling further and further behind and losing wallet share. As for the Analytics Advisor, whether coming from the outside or an internally sourced role, the challenge is far more pronounced.
First, just who are these Analytics Advisors?
Analytics Advisors are change agents. They tend to be mature in their hands-on data science abilities. They usually have years of experience and have seen data science transform through the years, given the advancement in computing power and data. It's not enough for them to be a walking dictionary of such changes; they must also constantly be able to map traditional processes and data to emerging ones, and know the best path for an organization to take in their digital transformation journey. Their advice melds experimental design, statistical knowledge, different machine learning techniques, software, data types, and emerging data architecture, so that there is scalability and eloquence in multiple threads of work happening at once. Their data architecture savviness is what grounds their advice on practicality and business needs, which is a vital part of the "how" conversation they must have with key organizational stakeholders.
Transformation Visioning and Talent Gap Assessment
While data architecture is where digital transformation conversations get tactical, mapping business needs to analytics is still the reason and starting point. The Analytics Advisor draws a change management roadmap from organizational visioning discussions with executive sponsors and transformation champions in the organization. This visioning must include a talent gap assessment and upskilling plan to ensure sustainable paths towards the vision. Analytics Advisors meld strategic and tactical plays by not only facilitating the articulation of the company vision, but more importantly, planning the course for the company to achieve different milestones simultaneously on their way to fulfill the vision. They are aware of the importance of showing multiple and frequent successes to hedge companywide adoption. The number of successes is, of course, a function of the number of trials, and therefore Analytics Advisors have to be masters of experimental design. Multiple trials without rhyme or reason lead to unreadable, uncompelling results. It's not the number of trials; it's the eloquence of the experiment that matters. Many a time, organizations would hire out strategic consulting and are left alone to execute on the advice received. This has often proven to be ineffective, as important tactical steps are missed or underestimated. The Analytics Advisor solves for this disconnect.
Why do Analytics Advisors talk about Event-Driven Intelligent Architecture?
Analytics Advisors should be able to lead the organization down the right path of scalable experimentation, where NRT insights are derived and deployed to the right venues relevant to customers when they matter. While analytics would accomplish NRT decisioning, the biggest boost comes from the infrastructure engineering of the data. Nowadays, for example, graph databases have become popular. Yes, a graph database is a good start, but it's still not enough. The most valuable role of the Analytics Advisor is to show where artificial intelligence can be installed in the data "wiring" so that proactive learning on the graph can take place. These are all data science happening pre-analytics. For example, in a recommender engine build, the innovation is in exporting a customer's static profile to the UI where dynamic, collaborative filtering (analytics) is happening in NRT. Without the proper event-driven intelligent (self-learning) system in that "export," the pairing to the right collaborative filters is not likely to take place. In other words, Analytics Advisors must be globalists in strategy as well as laser-focused tactically, able to drill down to the data infrastructure so that digital transformations are analytics-conscious. After all, what good is a treasure trove of data or the most intelligent data architecture or the most elegant analytics if no well-thought-out use cases await them?