Skip to main content Skip to Footer

BLOG


August 10, 2015
Better forecasts start with understanding Traffic
By: Neil Fernandes & Sumeet Mahajan

If you belong to industries such as retail, car and equipment rentals, hotels, airlines, etc., that deal with very limited point-of-sale inventory, then read on to see how something that you do almost on a daily basis can be done a lot more efficiently and effectively. If you are not from any of the industries above but rather belong to more traditional industries such as manufacturing, then you too should read on to consider a whole different way of looking at demand forecasts.

Our goal has always been to help clients across various industries reach the efficient frontier of their operations performance. In almost all cases it has been done by leveraging advanced analytics along with some progressive concepts such as end-to-end optimization, risk exposure index and segmentation, to name a few. In 2014, OPS Rules chairman, David Simchi-Levi identified, experimented and succeeded with yet another innovative concept to help move organizations towards a more efficient level of performance. This time, advanced analytics were used to improve product pricing, resulting in immediate and substantial impact to the company’s bottom-line, unlike traditional supply chain initiatives where the Return on Investment (ROI) period is at least a few quarters. This effort, first implemented at an online retailer in the US, received the 2014 INFORMS Revenue Management award.

This new concept differs in one major assumption from the traditional approach—that the point-of-sale demand is a lot more malleable, based on controllable as well as uncontrollable factors. And unlike the traditional supply chain organizations where most of the operational plans are etched out based on forecasts, this new concept works in industries where you have more power to influence the daily or weekly demand by varying factors such as price and product exposure.

Examples of such industries would be traditional and online retailers, car and equipment rentals, hotels, airlines, companies operating car and bike sharing programs, etc. The challenge across all of these industries is the same: There is limited point-of-sale stock, and there are daily (or weekly) revenue and profit targets to meet. How should you price your mix of products (or services) to maximize your performance?

This is where Dr. Simchi-Levi challenged the existing methods of "cost-plus" pricing or "following the competition". He proposed a mix of predictive analytics using machine learning techniques, and prescriptive analytics using some interesting efficient algorithms. From internal and external historical data, you learn the conditions at which potential demand became actual demand (how many customers looked at the product versus how many bought the product), and then set optimal price and exposure points to maximize performance, while working within your operational constraints like limited stock and minimum performance targets.

Because of the constraint of limited stock, a key to optimally price your products is predicting the range of potential demand, also known as traffic. Traffic is the number of entities that could be interested in buying the product or service, and show up at a physical or virtual location. Eventually, a fraction of the total traffic ends up buying the product thus becoming actual demand.

Here's an illustration of an application of traffic forecasting based on our work with an equipment rental company. Figure 1 is a sample price-demand demand elasticity curve for a product, say a bulldozer. P1 is the price at which we predict to rent out D1 units of the equipment to maximize revenue, without considering any inventory constraints.

Figure 2 illustrates how today’s inventory position (red dotted line, indicating number of available units) is less than the predicted D1, and hence price P1 might be optimal but not practical. Instead a better price point is P2 where we predict to sell D2 which maximizes revenue while satisfying any stock constraints? But to be able to decide the right pricing strategy, i.e. choosing between maximizing revenue and maximizing margin, we need to know what the expected traffic is. If there are few units available but we are expecting higher traffic, then we pick margin maximization, while if we have more inventory and lower expected traffic then it is better to maximize revenue (think of this as a 2x2 matrix with traffic forecast and inventory level as the axes, and the pricing strategy as the decision variable).

Industries such as hotels and e-retailers have been using some sort of traffic information as part of their pricing strategy for some time now. But they have not yet started leveraging machine learning techniques to understand how the price-demand elasticity curve changes with different levels of traffic. In fact, in our recent experience, a good quality of traffic forecast is a strong predictor of sales right behind your price and your competitors’ price. A good quality traffic forecast will not only help improve the predictability of sales but can also be a very powerful decision support tool for deciding product placements and investments in promotions and exposure.



The methods for predicting traffic are as varied as the number of applications itself. They vary from simple structural methods like exponential smoothing to machine learning methods like neural networks and regression trees. The model and the predictors of traffic chosen vary depending on the specific problem at hand. The selection process is as much an art as science.

We have been able to predict product category level traffic and even individual product level traffic with up to 90 percent and 75 percent accuracy, respectively. However, let’s not forget that predicting traffic is only the first of the many important steps in price optimization. Impact of other important predictors of sales such as competitors’ price and events, also need to be understood better. Using a mix of predictive and prescriptive analytics for pricing, can raise an organization’s operating model to a whole new level of performance.


About Accenture

Accenture is a leading global professional services company, providing a broad range of services and solutions in strategy, consulting, digital, technology and operations. Combining unmatched experience and specialized skills across more than 40 industries and all business functions—underpinned by the world's largest delivery network—Accenture works at the intersection of business and technology to help clients improve their performance and create sustainable value for their stakeholders. With approximately 373,000 people serving clients in more than 120 countries, Accenture drives innovation to improve the way the world works and lives. Visit us at www.accenture.com.

Accenture Digital, comprised of Accenture Analytics, Accenture Interactive and Accenture Mobility, offers a comprehensive portfolio of business and technology services across digital marketing, mobility and analytics. From developing digital strategies to implementing digital technologies and running digital processes on their behalf, Accenture Digital helps clients leverage connected and mobile devices; extract insights from data using analytics; and enrich end-customer experiences and interactions, delivering tangible results from the virtual world and driving growth. Learn more about Accenture Digital at www.accenture.com/digital.

 

About Accenture Analytics

Accenture Analytics, part of Accenture Digital, delivers insight-driven outcomes at scale to help organizations improve their performance. With deep industry, functional, business process and technical experience, Accenture Analytics develops innovative consulting and outsourcing services for clients to help ensure they receive returns on their analytics investments. For more information follow us @ISpeakAnalytics and visit www.accenture.com/analytics.

This document makes descriptive reference to trademarks that may be owned by others. The use of such trademarks herein is not an assertion of ownership of such trademarks by Accenture and is not intended to represent or imply the existence of an association between Accenture and the lawful owners of such trademarks.

This blogpost is produced by consultants at Accenture as general guidance. It is not intended to provide specific advice on your circumstances. If you require advice or further details on any matters referred to, please contact your Accenture representative.

Popular Tags

    More blogs on this topic

      Archive