Pricing & Promotion Analytics
Huge Payoff for Getting Pricing Right
- What drives the decision to buy in a category at all?
- What can the consumer afford? What is he/she willing to pay?
- Where does the consumer make the pricing decision?
- How does the consumer react to weight-outs or other changes in size and count that have value implications?
- What is the lift from different promotional tactics?
- How do consumers react to thresholds and deals? What is the mental change that occurs when we offer products at whole dollar boundaries (e.g. 99¢) or "3 for $5.00" or "Buy one get 40¢ off the second?"
- What should be the price of a new product?
- What do I do in a developing economy, when the product may not even have the price marked on it, or the product may not even be displayed?
Historically, pricing has not received the attention it deserves, but it is now in the spotlight as one of the most important areas to achieve greater profitability in today's competitive markets. Business leaders who grasp the value of predictive analytics for pricing can stop looking over their shoulders to see what their competitors are up to and start taking charge of their own pricing destinies. The payoff for doing pricing right is huge.
in4mation insights offers a framework for pricing solutions across a variety of issues, both strategic and tactical. Almost all analysis must involve an understanding of consumer sensitivity but we also recognize that price plays a role inside the company and throughout the supply chain. Pricing issues impact many divisions in a company and so the first issue is to understand multiple perspectives: is getting pricing right about positioning a product, holding share or maximizing profit? The first step is therefore one of disambiguation: what are the objectives? What information is explaining or obscuring the issue? The following matrix is a good starting point for identifying issues in pricing
The Intersection of Price Analytics and the Consumer
Most of the time, we monitor price by conducting analytics on aggregated sales or panel data. This leads us to produce models that predict changes in sales based on changes in price and promotion tactics. This field, called "Econometrics," is one in which in4mation insights is well-versed and approaches with a range of advanced Bayesian models including:
- Non-linear sales response models (the typical marketing mix model).
- Choice models involving discrete sales or quantities.
- Integrated choice/response models ("MCI models" are one of several in this family).
These models work well if we know what new price we want to measure, the price has already occurred in the range of the data, and we are interested in the impact on sales. They work better in mature categories where prices have stabilized and innovation is not disruptive. These models are also useful for conducting optimizations that enable us to understand price movements that would be ideal from a technical point of view.
However, these models are not good for cases where we have new products, new attributes, or where price falls outside the historical range. Also, these results are rarely insights-rich. They often leave us with more questions than answers. For instance, most of us know about the link between high loyalty to a product and the ability to charge a price premium. But what is driving the degree of price insensitivity in this case? Are we charging enough? What kind of guidance can we get if we are the market leader? What guidance for a new product? A new package size? A truly disruptive innovation?
To address these issues, in4mation insights has pioneered a range of models that help us to understand consumer dynamics around price. We are interested both in finding the price that consumers are willing to pay and in identifying the drivers of the amount they will pay.
One of the key concepts that we emphasize is to distinguish between issues of affordability and those of substitutability. Affordability (also called an "income effect") governs whether someone can buy in a category at all, or will buy the higher quality brands in a category. Substitutability is the switching between products in cases where people have moderate preferences, are sensitive to different attributes in the product, and can be swayed by price.
Another key concept is the need-state driving the selection of the product. What underlying need-states are so important to consumers that they will pay more to satisfy them? The goal of our analysis is to both frame the price and to understand the consumer drivers of the price continuum.
in4mation insights uses the full array of Bayesian econometric and survey-based price models as a platform for optimization. Since our models capture all the SKUs and channels in a category in one modeling framework, the optimization takes account of all known sources of variation in demand.
However, our point of view is that price optimization will only work if we have a full understanding of the capabilities of our client and the competitive context. Competitive price handling and monitoring systems that claim to do automated optimization are misleading as often as they are useful. Misleading information does not improve results and leads to another decision tool that no-one uses.
Here is a short checklist of issues that must be addressed for price optimization to be actionable. Do your price actions, business rules, and/or optimization tools take account of these?
Competitive price gaps
Are you the price leader or follower?
How often do you follow price changes?
Is it the upscale or the lower quality tier that is important to gap to?
Can you produce the volume required for an optimum price?
Do you see regular consumer adaptation to higher prices (such as an annual price increase)?
How long does it take for consumers to get used to lower prices and start to expect them?
Are the goods/services in your category of very different quality?
Can only some people afford them?
Are a large group of people waiting to buy you on sale or later?
Are consumers using price as a signal of quality?
Is your category categorized by high/low loyalty?
Do promotions keep people in your franchise or only steal them?
We Offer a Broad Range of Predictive Solutions
There is growing momentum behind the shift from descriptive to predictive analytics in pricing. We offer a broad range of tools and techniques that predict future outcomes.
Full Range of Syndicated Data Modeling
- Sales response models.
- Hierarchical Bayesian models.
- Mixed choice/quantity models.
Deep Expertise in Choice Modeling
- Original innovators in the use choice experiments (CBCA) and advanced experimental designs for choice modeling.
- Choice models that are calibrated to match syndicated data on share and elasticity.
- Modeling the motive and attribute drivers of elasticity.
- New product models.
- Experts in capturing price thresholds/consumer budgets.
- Modeling choice of multiple units, build-your-own or menu tasks, and multiple unit and multiple item choices.
Predictive Analytic Tools
- Full range of fast simulators with a long shelf-life.
- Software-as-a-Service (SaaS) tools managed by us, available on the web, using high performance computation and parallel processing for quick results.
- Experts in integrating models with corporate planning tools/systems.
Tools for Developing Countries
- Price testing methodologies that work in developing countries.
- Data triangulation strategies to make up for poor quality, sparse, or intermittent data.
- Focus on affordability in traditional channels, not just applying advanced dynamics of modern trade models to other countries.
Tools for Online Pricing
- Price modeling for online retail sales, using daily or even hourly sales data.
- Web-scraping algorithms for monitoring own and competitive price.
- Web-scraping algorithms and tools to monitor the effects of online social and clickstream activity.
Case Study: Our pricing recommendation, if executed, would have saved the company $30MM.