When it comes to demand forecasting, most companies have way too many forecasting mechanisms in play across their organization. Each segment of the business ends up siloed from the others, relying on its own data and analysis, which impacts both efficiency and effectiveness.
For a retailer to sense, shape and synchronize demand across a broad range of business functions, a single centralized mechanism is required. The payoff is essentially one version of the truth, an impossible end when using multiple forecasting systems.
The ability to deep dive into the data to identify very specific elements of the demand signal is vital, and these elements include understanding the baseline, what is incremental, what is halo, what is cannibalization, and how trends and seasonality factor in. By using the right forecasting system, you can then reconstitute that signal to help drive a coordinated pricing strategy.
By combining the right data and the right analytics you’ll know, with confidence, the best actions to take on all aspects of pricing, from base and promotion, to markdown and lifecycle. For example, are you better off with a buy-one-get-one-free promotion or a buy-two-get-one-free? This may seem like a promotion optimization capability, but what’s actually foundational to the optimization is forecasting. By simulating the incremental lift that each of these approaches will likely yield, the right combination of forecasting and optimization can steer you to make the best decision for your business.
An effective forecasting process always starts with a series of questions, such as these:
- Do you have a proper demand signal repository, and have you gone through the proper scrutiny to make sure the data is actionable?
- Are you applying the proper analytical ETL (extract, transform, load) to ensure there's consistency in the way the data is represented?
- Are you looking at data holistically, and are you able to consider analyzing the data through different lenses, such as a consumption view and a shipment view?
- Have you captured historical causal, and explained the types of promotion, and the combination of causal support, to get a specific lift?
- Have you spent the time building out attributes and consumer decision trees for the products you carry?
Did you notice these questions all center on data quality as a starting point for effective analytics? Clean, actionable and complete data is integral for sophisticated forecasting systems to unearth significant insights and deepen a retailer’s understanding of the factors and demand levers in play to enable optimal decisions.
The analytics world is at a tipping point; data has become a core asset of virtually every organization and the ways in which they view it, act on it, and leverage it for better decision making, will ultimately determine their success and continuation.
Organizations with superior analytics capabilities have a deeper understanding of their business to more efficiently drive productivity improvements. These organizations will prosper, while those retailers that are lacking will likely end up focused on mere survival.
Whether the forecast is being leveraged for planning, pricing, operations and/or fulfillment, the strategy is the same; leverage your data as an asset, apply the best analytics possible to derive valuable insights, and be consistent with how you leverage those facts within and across the organization. If done well, you’ll enhance collaboration and productivity within and across functions, with outcomes that drive revenues, margins and profits.
An optimal forecasting capability empowers your organization to see the business through a single lens, leveraging the same foundational facts about how the business has performed and how the business is predicted to perform in the future. It also enables agile course correction at all levels of the organization, from planning and buying, through allocation and replenishment.
For additional information about quality data and demand forecasting, see our eBook on “6 Reasons Why COOs Are Embracing Demand Forecasting.”
