In This Post
Fragmented analytics tools
Fun fact: There are roughly the same number of different currencies circulating in the world today as there are Analytics/BI and Data Science products listed on Gartner’s Peer Insights platform. Talk about fragmented!
Benn Stancil broke it down one time in a tongue-in-cheek way:
— Benn Stancil, “Business in the back, party in the front: Sorting through the chaos in the consumption layer”
The problem with fragmented tools
Having choices is a good thing (unless you’re indecisive). But even after you’ve decided what tools to use, most organizations end up with five to 20 tools in their toolset. This has downsides:
- You have to ensure that tools can work together in a cohesive workflow, from data to analysis to consuming insights
- There are multiple steps in the end-to-end workflow, slowing down the overall process and making iteration difficult
- Tools tend to be narrow in scope and cater to specific personas, keeping analytics creators (business intelligence analysts and data experts who generate insights) and analytics consumers (business users who make decisions on insights) siloed in their own tools
Analytics and data science silos
For many companies, analytics and data science exist in separate silos.
Analytics, business intelligence, and SQL query-based tools are used primarily by data analysts and business teams who use dashboards and reports to track key performance indicators.
Data science and machine learning are the realm of data scientists and experts who look to fine-tune models for predictive analysis and advanced insights. Findings are shared to business teams by transferring data to the BI system. Users of both sides of the aisle often speak different languages, and they sometimes have different names for the same thing (e.g., what one calls a “feature” could be a “calculated column” to another).
These silos are gradually coming together, especially for organizations building new data teams who have purview over both analytics and data science. But the organizational shift is just the starting point—why should businesses want the technology silos to be broken as well?
Why silos should dissolve
The truth is that businesses don’t care about what is a simple query problem versus what requires advanced segmentation of hundreds of variables. They don’t care about which team of analysts is answering the question they have.
Businesses want to monitor high-level metrics, perform a root cause analysis, and then make data-informed decisions—and they want the fact-finding exercise to be as seamless as possible. In the ideal world, they can access descriptive analytics to understand what’s happening in the business, diagnostic analytics to learn the reasons why things change, predictive analytics to grasp what will happen in the future, and prescriptive analytics to decide what next best action to take.
Because analytics tools focus on a specific type of analysis of data, organizations end up using multiple tools and approaches:
- Business users and analytics consumers use business intelligence tools and dashboards for descriptive analysis to understand what’s happening in the business and visualize performance metrics
- Analysts write SQL queries or use spreadsheets to slice-and-dice data for diagnostic analytics to uncover why things change
- Data experts employ data science and machine learning automation for predictive and prescriptive analytics to formulate recommendations and next best actions
But hold up! Analytics is a team sport, requiring analytics consumers and creators to collaborate. With analytics/BI and data science/ML tools separate from each other, traditional approaches using disparate tools slows down data-backed decision-making, exacerbates analytics backlogs, and forces decision-makers to rely on gut and intuition.
But what if analytics and data science were not separated?
Two sides of the same coin
At Tellius, we have always developed our product thinking that Analytics/BI and Data Science/ML are two sides of the same coin. Both approaches should be used hand-in-hand to provide the decision intelligence necessary for making better data-driven business decisions.
Combining BI and artificial intelligence (AI) creates a much higher-value analytics currency:
Analytics/BI: Beyond dashboards and reports, anyone can ask questions of data and rapidly explore terabytes of data in an ad hoc way with a Google-like natural language search interface and automated visualizations.
Data Science/ML: In addition to flexible machine learning modeling (e.g., AutoML, point-and-click, or Python/SQL), the auto segmentation, signal detection, proactive intelligence, and no-/low-/full-code capabilities make it easier to get recommendations for decision-makers to take action.
Unified Analytics Experience: A unified analytics UI allows analytics creators and consumers to seamlessly flow between various analytics modes to answer what, why, and how types of business questions. People across various roles can apply their technical expertise and business acumen to collaborate on generating and validating breakthrough insights.
Intelligent Automation: AI-powered insights automatically evaluate every combination of data points to surface the most important findings by providing automated key driver analysis, cohort comparison, and root cause analysis insights, accompanied by NLG summaries, to dramatically cut down on time spent analyzing data.
Architecture: Tellius has designed a unique dual analytics engine that allows for lightning-fast ad hoc exploration, coupled with a robust distributed compute engine for ML processing, to support both BI and AI workloads in a single product.
GenAI: We can’t forget about generative AI, of course. The application of large language model (LLM) technology on a platform built for multi-persona analytics provides even more opportunities for collaboration.
Heads, you win; tails, you win
Organizations use Tellius’ AI analytics platform to leapfrog their analytics and BI capabilities and make AI-powered insights accessible to more people. Our customers use a single platform to get the best of both worlds, whether they approach problem-solving from the analytics or data science perspective.
Two markets. One purpose-built product.
Want to see more?
