Analytics

Three Design Principles for Operational Analytics

By Michie Cao - June 27, 2019

Our mission at Sisu is to help people make more informed decisions with data. In today’s business environment, that’s no small task. Operational leads, marketing managers, financial analysts, customer support experts, and store managers each make dozens of decisions every day. Add that up across even a single organization, and you quickly get to thousands of decisions every month.

Most organizations gather the data needed to inform these decisions, but even with today’s tools for analysis and dashboarding, it takes a team of technical experts (and lots of time and patience) to find the insights that really matter.

We believe that needs to change. As teams collect more data and make more immediate operational decisions, they need new tools that will enable each team member to analyze data quickly and easily. And the results from that analysis need to be understandable and easily shareable to their teammates.

To get there with Sisu, we’ve identified three key design principles for making data analytics and collaboration accessible and understandable to everyone in an organization.

  1. Low Threshold, High Ceiling
    Our first design principle ensures that anyone using Sisu – from an operational lead to an expert data scientist – can get useful results quickly. Then, for the truly expert customers, allow them to dig in deeper, perform sophisticated analyses and comparisons of the results, and verify those results, as needed, with Sisu’s rich statistical context.This is our most powerful differentiator. Traditionally, analytics tools are technical and complicated; and when they’re not, their output is a simple, descriptive dashboard that doesn’t tell you why things are changing. This makes it hard for everyone in a team to make business decisions together. At Sisu, we believe everyone should play a role in the analytics and decision-making process. Building a shared language and knowledge about the data is critical.
  2. Let the Data Speak
    Our second design principle is all about helping our customers tell a better story with their data. It’s not enough to provide a “black box” prediction and expect customers to run with results we give to them; that gets in the way of empowering people to truly understand what’s happening and encouraging them to collaborate with others. We have to provide transparency in how we got those results in order to build trust with our customers. By showing the work we’re doing and highlighting the numbers, we can bring clarity to complexity.An important benefit of this is that we’re also seizing the opportunity to teach more sophisticated data interpretation and analysis skills to people, further accelerating their confidence and ability to turn these insights into action in their organization.
  3. Deliver the “Wow”
    This last design principle serves double duty, both as a guidepost and a way to evaluate the impact of our decisions. As an operational analyst, there’s nothing more satisfying than finding the key insight needed to inform a tough decision. We build for these moments of “wow” and look to remove the friction between our customers and these magical moments, whether that means providing speedier data processing, better result discovery, or more easily digestible results.And then when it all comes together, we can ask ourselves “Was that a truly ‘Wow’ moment, or can we do better for our customers? What would make this experience even more amazing?”

Understanding what our customers need and solving that struggle in their day-to-day work is critical to our customers’ success at Sisu. It’s why I’m here, and why I’m excited to help people tackle their toughest operational analytics tasks.

For more on how we’re tackling this challenge, or how we could help your team find the facts hidden in your data, read more at www.sisu.ai.


Read more

August 8, 2019

Why Everyone Needs a Dedicated Analyst Team

With the investment we’re making in collecting structured data, everyone could benefit from a dedicated analyst team. But almost nobody does.

Read more

October 3, 2018

Systems for Software 2.0: A Lesson from Web 1.0

Our modern stack is built around a deterministic, precise model of computation. But statistical forms of correctness are the new normal for software 2.0.

Read more