Intelligent Business Planning

Many traditional Integrated Business Planning IBP platforms being hyped these days as the new deal are essentially glorified spreadsheets in the cloud with traditional planning operators like ABC-XYZ segmentation or traditional statistical forecasting. They bring the same failed on-premise manual intensive spreadsheet based S&OP paradigm in a new shiny repackaging: now called IBP in the cloud with an Excel UI interface, with an ability to crunch large volumes of data, and disaggregate / aggregate / visualize results on the fly. Although these ideas do improve the offline spreadsheets driven S&OP environments to some extent and they can be deployed relatively quickly (4-6 months), they fall short on the business value front. Why? The same basic forecast models like triple exponential smoothing that failed to deliver sustainable improvements over the last two decades are being used to drive demand review portion of the S&OP process.

This is a world in which consensus demand forecast provided by simplistic forecast models and/or sales teams is assumed to be correct at product/customer level in all future periods. At best, the demand planner can perform a few local what-if simulations to understand alternative demand scenarios (just a couple upside / downside scenarios or versions). This theoretically produces an accurate consensus demand plan. All that is left to do is to take this demand plan and explode it through Bills of Material in Supply Planning process and develop a capacity constrained Supply Plan and proceed to buy Raw Materials from Vendors.

We all know how this turns out in reality. Sales orders come in different than demand plan: what is ordered is different, when and how much are also different. Results are a lot of turbulence in operational planning / execution horizon as needed Raw Materials are short, leaving Supply Plan sub-optimal on capacity usage front, or worse non-executable. High plan to actual gaps are reported. Supply chain is blamed as On Time Delivery performance suffers. Then the company gets back to the next S&OP cycle to repeat the madness all over again.

Why are we still using demand forecast models developed half a century ago or earlier, when computing power was scarce? Cloud computing and in-memory databases are pointless, if the planning algorithms are still so last century. At GitaCloud, we believe that a fundamentally different approach is required to tame the turbulence. Instead of fighting demand uncertainty and asking for frozen demand fences, we should embrace this uncertainty. Why limit the alternative future consideration to a few ad-hoc what-if simulations? The compute power available in cloud today combined with deep learning algorithms can autonomously model the full range of risk & uncertainty in future demands by deeply analyzing customer buying behaviors for product / quantity / timing risk. 

The result is not a single deterministic demand number for every product/customer/week combination, but a full range of probabilities for all possibilities in a stochastic probabilistic demand prediction. This is essentially what-if simulation on steroids, unconstrained by number of planning combinations or planner bandwidth. 

This approach to demand modeling results in a much better appreciation of risk in the demand signal. Supply chain can now work with the commercial teams to align on what part of the risk spectrum needs to be covered for in terms of buy/make/move plans. This leads to much reduced turbulence in operational horizon as we place optimal buy/make/move calls to optimize revenue/margin with full consideration of risk. 

GitaCloud can provide buy/make/move signals as a service to the planners, who then are freed up from the forecasting/planning chores every cycle and can focus on understanding the business environment to tweak the plans on an exception basis. Think of this as your autonomous car, where you can take control at will, and at the same time, read a book on your morning commute to work.

At GitaCloud, we do not engage through just a technology implementation lens. The technical go-live of an IBP software means little, if the business users see no hard value in terms of sustained improvement in on-time delivery, reduction in expedited shipping costs, reduction in inventory levels, etc., and is a recipe for quick reversion to the same old spreadsheets. We work with you to provide accurate demand & supply signals as a service, all available on your IBP platform to further refine the signal and execute as you please. We track value-added vs. the baseline signal and audit value from our service on an ongoing basis (continuous business case validation as customers can stop the service anytime).

Curious? Send us an email at, if you would like to learn more.