Planning as a Service: short-term gimmick or emerging strategy? 

At a supply chain software vendor event this summer in the United States, the CSCO of a major fast-moving consumer goods company declared, almost as an aside, that his company was in the midst of a progressive transfer of baseline demand planning to third-party service providers. Since this does not seem to have yet been formally announced, the parties won’t be identified here. But the development is consistent with anecdotal information that is circulating concerning other companies. 

With the apparent emerging trend of planning as a service now being questioned as a core competency, it’s worthwhile to consider the potential benefits and pitfalls of outsourcing demand planning and the contexts where the trade-offs could make sense. 

The potential pitfalls 

There is a temptation to consider the activity of demand planning as a purely analytical and statistical undertaking, wherein if the company has chosen the best tool and the planner chosen the best algorithm, then the resulting demand plan will be as accurate as possible and there is nothing more to be done. 

Experienced supply chain managers know this not to be the case. The most challenging part of demand planning is not finding the right algorithm or machine learning technique, it is largely about getting the market intelligence. Market intelligence may be information on customer or consumer behavior, growth, and expansion plans, media spending, or new accounts – and so applies not just to launches and promotions but also to baseline forecasts. The holders of this information, usually sales and marketing, often look upon the supply chain as a secondary function, not as a valued partner. Using an outside firm to prepare demand plans will only add friction to the information flow to planners, making it less likely that the demand plans fully reflect the best forward knowledge of future activity. 

One recent trend in demand planning is to move towards probabilistic planning, where different demand scenarios are evaluated, assigned probabilities and the resulting aggregate vision is used as the demand plan. This would be much harder to implement with outsourced planners as it requires even deeper collaboration between business functions to identify and weigh the possible scenarios. 

Similarly, it will be difficult for an outside firm to have all necessary information to properly execute the initial ‘cleansing’ stage of demand plan preparation by taking into account known past events like a stock-out or other extraordinary event that should be corrected before initiating statistical data treatment. These frictions to data flow are in addition to the more first-order technical ones such as product master files, actual sales, cost of goods that also come with set-up costs. 

There are broader managerial considerations at play when evaluating the potential pitfalls. Research has shown that 80% of large, global, mature companies suffer from demand plan manipulation, where the general manager will take the demand plan that resulted from careful analysis and construction, and force arbitrary changes to the number in order to preserve stock buffers and enable a form of shortage gaming. This unfortunate behavior is likely to increase if the planners creating the demand plan are outsourced. One of the principal reasons the sales and operations planning ‘one set of numbers’ pillar was developed was to discourage this sort of manipulative behavior.