May 9, 2019
To grow business, large volume commodity suppliers — including steel manufacturers, agricultural trading companies, and downstream oil distributors — traditionally focus on improving asset utilisation and volume placement. Sales efforts rarely include price optimisation or promoting margin diversification when trading with customers and these gaps in process leave money on the table.
In order to differentiate their offer from the competition, marketing departments at these companies end up trying to bundle value-added services with their products. Too often, these efforts do not pay off. By creating extra costs in processing, development and marketing, these “extras” ironically reduce already narrow margins. Offering additional services may help to incentivise customers to maintain their product volumes, but often at a cost that offsets any incremental profit. It is even worse when customers do not value the service and they are then given away at a substantial discount or even for free. Reasons could be various, ranging from pricing compliance to misjudging customer willingness-to-pay.
Instead, to improve the bottom line, B2B commodity producers and distributors need to explore new capabilities to their marketing models. Identifying dynamic segmentation and price optimisation opportunities across product-customer portfolios can yield benefits in the range of 2-5% of top-line revenue straight to the bottom line. Dynamic pricing enables organisations to differentiate pricing across their customer base and throughout their markets, react faster to changing market conditions, and improve quoting effectiveness.
To build efficient dynamic pricing capabilities, companies must introduce three key components:
1. Deploy a dynamic modelling capability to capture higher customer surplus via differentiation.
Price differentiation relies on defining – and often re-defining – combinations of products + customers that matter and explain customer willingness-to-pay in each market situation. Leveraging algorithms to build a regression-based segmentation model (a form of supervised artificial intelligence) will allow you to explore which dimensions and attributes in your data provide meaningful drivers for differentiation. Combined with cloud processing power and scalability, these algorithms enable pricing managers to build and test different models within short timeframes – hours instead of months or weeks. In simple terms, you can let algorithms figure out how to make sense of your available data. Relying on a machine learning based approach will get rid of many biases towards business segmentation.[
2. Test all possible data types for more accurate segmentation models.
All customers are not the same; their way of transacting is different; they are willing-to-pay different prices during various market events. Purely descriptive or static data, e.g. sold-to industry or end-use, can only partially capture customers’ price acceptance behaviour. Enhancing transaction data with market information provides a better way to understand customer behaviour and willingness-to-pay.
The good news is that external data is readily available in commoditised industries. It is relatively easy to access information that can provide a forward-looking indication of price development – think of market indices, freight trends, or even sophisticated data used by quant desks (e.g. volatility measures). The result could be differentiating or even dynamic adjustment of margin capture opportunities.
Companies that leverage several data sources will be able to augment pricing segments with contextual information. Knowing up-front whether your data is critical for segmentation is less significant – your modelling algorithm should figure it out. After all, segmentation is a discovery process that should provide a statistical perspective on your business.
3. Optimise prices on ‘true’ margin to keep profitability safe.
The nature of commodity marketing often carries complex contract schemes or additional costs related to customers executing options. Besides, not all customers call fully committed volumes. Thus, the premium charged per unit might not represent the actual margin realised on transactions.
Any analytical and price optimisation capabilities must provide a ‘true view’ for business profitability (including all costs attributed to transactions). It makes little sense to offer target prices with low margins which your cost to serve can quickly push into the red. A fully defined pricing waterfall supports accurate, agile, and action-biased pocket margin calculations for granular measurement and precise control.
Best practice dictates applying guardrails of profitability on top of optimised guidance to keep the margins safe – this is where knowing how the AI makes decisions is critical, as well as the ability to easily review and revise before changes are put into action. Minimum margin requirements, customer risk adjustments, discount corridors are typical examples of policies leveraged in dynamic pricing. What-if and impact analysis capabilities will help you set and verify those values.
Commodity commerce has excellent potential to increase profits, achieve better customer experience, and raise sales confidence via faster and more accurate pricing. Natural resource and basic material businesses should aim to create a closer relationship with their customers, improve on market intelligence quality, and deploy algorithmic optimisation and decision support to their value capture efforts.
Dynamic pricing capabilities supported by modular architecture (for inbound data feeds and outbound quoting solutions) and cloud scalability define the path to commercial excellence for commodity trading business.
To learn more about how an energy company transformed their pricing into a high-performing, profitable business function, read this success story: Global Pricing at Energy Giant Transformed Into Well-Oiled Machine.