Retail and online sellers are dealing with a complex environment of competition while at the same time managing their own multi channel environment. These companies are already lean so there is not much room to cut costs. Therefore, the ability to increase revenue and improve margins through an advanced price optimization process can make a huge difference in their business.
In Three Steps toward Price Optimization Success we described the steps that can be taken to achieve price optimization success in an organization. These steps were to predict the price elasticity, or Price to Volume relationship for products for sale, define a price optimization model given the organization’s goals, and install a change management program to bridge any gaps between the technical and commercial teams. When organizations follow these steps, they position themselves to not only to improve commercial performance, but to better understand how their actions influence both revenue and profit in their organizations.
This first phase of work could be considered a Base Implementation Phase of the price optimization project. At this stage, the team has a method to define the price elasticity, taking into account all necessary business factors from data feeds. There is also a production optimization model to achieve a business goal such as maximize revenue given cost margin constraints. And also in this stage, the commercial team has entrusted the technical team with pricing products under this new type of pricing model while checking for consistency.
Once this base implementation phase is completed, there are three areas to explore in order to enhance the results. They are Performance Measurement, Algorithm Enhancement, and Software Modification.
Because the base implementation was at a smaller scale – subset of products, percentage of time in practice, etc. – we run the risk of experiencing performance that is inconsistent with our expectation. When scaling, it is both natural and recommended to assume that the historical subset performance will expand to a larger set. Therefore, collecting statistics on algorithm performance is required to determine whether any changes should be made to the process.
A/B testing is typically the technique data science teams use to measure performance. The definition of A/B testing is reserved for running two strategies on the same product at the same time. But there are also other ways of testing such as same product, different time or different/related product, same time.
Statistics that reveal that the strategy is performing well indicate that no immediate modifications should be investigated for process. Poor statistics indicate that the algorithm is failing, that another strategy performs better, or that the performance is good for the wrong reason (implied luck). In these cases, the statistics reveal that modifications to the approach should be made – and asap.
In this context, algorithm enhancement specifically refers to changing the way elasticity curves are created. We assume that the optimization model around them is true – and that the source of potential error is in predicting the Price to Volume relationship.
These are some techniques that we have applied to enhance model performance:
Smoothening elasticity curves – any randomness in the curves could enforce suboptimal answers with high variance
Adding/Removing Input factors – changing input data into the model may dramatically influence the shape of the elasticity curves, and hence the optimal prices
Change the actual decision model – the model itself is not always ideal. For instance, a random forest may work one month, but a neural network may perform better in subsequent months.
Once these potential enhancements are investigated and researched, it is important to backtest the strategy on historical data to gain confidence to play it forward in a live commercial environment.
So far, we have discussed algorithm performance from a commercial perspective. There is also a technical perspective on how well the algorithm performs – examples being computational complexity, run time, and code organization.
Here are some specific techniques that we have applied:
Runtime improvements – streamline code to be as efficient as possible
Database Management – set up real-time connections to the databases and design efficient queries
Memory Management – store data in a way that enables solving complex optimization problems faster
Code Organization – modular design to simplify training and debugging activities
In conclusion, it is important to note that price optimization is most likely a continuous process for an organization unless the purchase behavior is stable and there is no competition. Teams should continue to look for better (more accurate/efficient) ways to use historical data to predict future patterns – as the specific methods typically change over time. Following the steps we have outlined will position your team to achieve pricing optimization success.
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.