Northbeam's MMM+ onboarding is designed specifically to your company's needs, goals, and ability.
Stage 1: Outline Primary KPI & Objectives
The first step is to define business objectives and outline questions that you’re looking to answer. It’s important to understand the goals and outline the scope to help the data science team determine how your MMM+ model will be designed. Learn how to choose a Primary KPI for MMM+.
Stage 2: Gather and Organize Data
Data collection is an essential part of the MMM+ process to yield data accuracy. Northbeam MMM+ takes daily sales, daily spend per media channel, promotion calendar, and any other categorical information to address business objectives with data-driven insights and scenarios. We use at least 5-months of both online and offline data. If you’re an MTA customer, most of this data collection is already automatically done. Read through Northbeam's data collection tips and maintenance best practices.
Stage 3: Build & Calibrate Model
Once the data has been cleaned, organized, integrated, and validated in your data input spreadsheet, we start the iterative process of analyzing historical data to make predictions of what your spend will be tomorrow or what your outcome would be if you had x budget. We follow best-in-class machine learning methodology to get a highly-accurate model, such as preventing an overfit using cross validation, automatic hyper-parameter tuning, feature selection, adstock transformations, and seasonality adjustment.
Stage 4: Analyze & Assess Initial Recommendations
Part of the calibration stage, a dedicated customer success manager and a data scientist will use the first data review to further hone your goals, assess initial outputs, and equip you with the skills to use the diagnostic tools to build a strong foundation for decision-making and get the most out of Northbeam MMM+.