≡ Menu

Getting to the Right Price with Oracle Data Mining


Rachel Scales presented on Getting to the Right Price: Using BI Apps with Oracle Data Mining to Improve You Company’s Margins. Pricing is increasingly complex as the world is changing and becoming more competitive. Customer loyalties are changing, resources are constrained and competition is more global. Price management is necessary to ensure your share of the “global profit pool”. In particular need to:

  • Digest cost increases in a profitable way
  • Capture value across product categories
  • Segment customers (by price and attitude to price) to address local needs
  • Efficiently address global price competition in local markets

Most companies today are doing this in spreadsheets! No single system of record for price, information hard to capture and use in the pricing process.

Oracles’ approach has a number of key concepts or activities:

  1. Price Analytics
    Current prices, historical prices, competitive prices and social norms
  2. Price Planning
    Some companies are still reactive but increasing use of price optimization
  3. Price admin
    Managing and publishing price lists
  4. Price Execution
    Guideline enforcement, price and promotions, pricing service
  5. Deal Management
    Recommendations and negotiations for a single deal

Price Segmentation – customer profiling, segmentation, price strategies and price determination – supports the analytics/planning stages. Getting the price right involves this  and price analytics.

For Oracle, Price Analytics is a standard OLAP application built on Oracle BI EE and Price Segmentation is a data mining application likewise built on Oracle BI EE.  The applications are targeted at B2B pricing. Oracle’s applications profile customers based on a wide range of values (cost to serve, breadth of product use, brand cachet, revenue….), segments them and starts with this framework. Additional information from the deal like competitors are applied within this framework to provide the right guidelines to sales teams/people/systems. Typically have 3-15 price profiles, 20-50 price strategies per year and anything from 1 to 500 price segments per price strategy. Could be simple or very complex.

Oracle Pricing Analytics monitors the prices offered and helps analyze the “price waterfall” – all the ways in which price is lost or price erosion. Shows trends and unexpected changes, rolls in how accrual-based things like incentives affect realized price and does analytics to spot pricing offenders and support pricing decision making. Has a bunch of pricing-specific KPIs and can show the effectiveness of the pricing strategies and plans developed. Although they call this an analytic application it looks like a dashboard/query environment to me. Oracle Price Segmentation uses data mining to identify segments and strategies that have the most potential for improvement. Uses attribute importance, for instance, to see what drives the variance in margin. Can determine pricing floors aligned with strategies, publish these floors and use statistics and data mining to ensure that pricing science is effective in operations. Both this and the Pricing Analytics module run on top of the analytic applications platform in Oracle BI EE and have pre-built reports, data models, ETL and more.

Rachel demonstrated the product and it was clear that the analysis tools would give some nice insight into the current discounting and pricing approaches used by a company. The data mining tools can be used to find the natural groupings of customers based on their value and behavior. Nice use of data mining to create statistically significant customer segments instead of judgmental ones. The historic price elasticity and discount patterns can be analyzed for each segment, helping to show where there is inconsistency and opportunity.

While I liked the analytic approach in this and the understanding of the flexibility that a salesforce needs to negotiate prices, it seems to me that this solution is too focused on the manual price negotiation aspects with not enough thought on how automated channels might interact with customers who also have access to manual prices. The guidelines, and models, could be used to drive rules and analytics-based pricing decisions that are unattended while also supporting manual decision-making.