How to Master Business Planning with Time Series Forecasting

Posted by Jess Lin

17.01.2017 10:15 AM

Every analyst report, news article, and business conference has drummed into our collective minds that predictive analytics is the way of the future. If you can predict future sales, you can funnel that knowledge into driving optimal sales and business operations. Smart businesses are leveraging big data insights to leave their competitors in the dust, or at least so they say.

As this new time series forecasting tech makes its way to the enterprise, trumpeting billions of dollars behind it, you feel ready for the challenge. You’ve learned a bewildering array of new vocabulary, sat through endless meetings on empowering a data-driven mindset, invested in top of the line data warehouses, and cracked your allotted share of jokes on the size of your Big Data. Best of all, you’ve somehow even managed to hire a crack data science team. Profit?

Not so fast.

The Three Stages of Data Science....png

Before those stacks of paper come rolling in and you deliver on all your ROI promises, let’s walk through this step by step.

  • Phase 1: Connect a beautiful data pipeline of well-cleaned and transformed data (yeah, we all wish) over to your data scientists.
  • Phase 2: Convert data into actionable insights through predictive analytics.
  • Phase 3: Feed those insights into well-executed, shareable dashboards that enable data-driven business planning decisions.

Even skipping over the problematic phases 1 and 3, let’s level set on what your data scientists can do. Even a team of rock star data scientists has limits, and working at top efficiency can still only produce a limited number of predictive models. Do you want dependable results at frequent, regular intervals? Best you can get is a single prediction for overall sales. Do you want per-unit forecasts across thousands of SKUs? Be prepared for a wait and infrequent forecast updates. Bringing in consulting firms doesn’t solve the problem either – fees can run into the multi-million dollar range for a years-long engagement that still only addresses part of your needs.

This is the situation that a major car manufacturer found themselves in when they came across Eureqa. With ever-increasing and increasingly-complex data streaming in from all sources, they had outgrown their existing, manual methods of data analysis and needed a new solution that could handle their data and provide innovative predictive analytics. Nutonian’s AI-powered modeling engine, Eureqa, seemed like the perfect fit.

Can Eureqa automatically ingest and analyze hundreds of data sources at once? Check ✓

Can Eureqa provide hundreds of highly accurate sales forecasts on a monthly basis across all levels from operations, from high-level national forecasts all the way down to individual dealers? Check ✓

Can Eureqa allow end-users to manipulate model results and simulate the impact on the sales forecast given different potential scenarios? Check ✓

Can Eureqa isolate and quantify the impact of individual attributes to the sales forecast? Check ✓

Can Eureqa identify and automatically recommend highest-impact actions for both a corporate strategist as well as an individual dealership? Check ✓

So we did all of that, and wrapped it up into a user-friendly application that works across the entire business team. With a marketing budget that runs into the millions, corporate can optimize their ad spend mix by using Eureqa to pinpoint where to increase spend and where spend can be safely decreased. At the same time, managers can use Eureqa to evaluate upcoming sales forecasts for their dealerships and design the best action plan for success, both short-term and long-term. By providing actionable insights and leveraging AI to automate interpretable predictive modeling, Eureqa is the “wheel deal” for this client.

Eureqa doesn’t stop at forecasting sales, either. Sales forecasts can be leveraged into powerful use cases like optimizing staffing, pricing, inventory, supply chain, marketing, and store expansion. Eureqa’s fundamental time series forecasting approach can also be applied to detecting financial fraud, diagnosing equipment predictive maintenance, building a portfolio trading strategy, and more.

What could you do with thousands of accurate, automated, and actionable time series forecasts? Reach out to us at to see how we can work together!

Topics: Business Planning, Eureqa, Time Series Forecasting

Follow Me

Posts by Topic

see all