I started working in pricing in 2000. First from an academic standpoint, where my interest was to inject data mining concepts into the evaluation of pricing performance in large organizations. Then quickly, the “real world” started to knock on my door and I started my first consulting engagements. What happened during these last 18 years has been an amazing acceleration driven by data and analytics. Companies in the 2000s were software and consultants dependent. Implementing price corridors across Europe was the fad then, based on excel sheets and incomplete data. Analytics were starting to make a difference through large-scale software deployments in an almost “pre-BI” era. Now companies have a fantastic opportunity to embrace the data revolution. As in the French Renaissance, the world is divided between the Ancient and the Modern.
So, we wondered: what makes a modern pricing organization today?
In our opinion, these four traits differentiate “ancient” vs. “modern”:
1. Architecture: Old World, New World
Cloud is no longer an option. To state it clearly, mature companies are massively moving away from legacy towards hybrid cloud architectures. Concerns about security for example, that were holding back large insurers or banks from the transition are no longer roadblocks. We recognize now that big cloud players (AWS, Google Cloud, Microsoft Azure) are far better equipped to fight increasing attacks and breaches than the best IT department. Modern architecture platforms are a powerful accelerator of data initiatives, tacking two main challenges: storage and elaboration. Every company willing to go for context-based, granular and near real-time insights is generating a need for scalable storage. The objection that “my company is not in the big data world” is not relevant anymore. As you blend internal and external data streams together (as you go granular), you increase what is called in statistics dimensionality, or hitting the limits of legacy systems in most cases. However, the biggest win of cloud platforms is what we call elaboration, which is the set of end-to-end bricks provided by cloud platforms supporting processing and preparation. In pricing, the latest is one of the biggest pain points. Data cleaning and preparation can now be semi-automated. “Ancients” will spend days and weeks to clean poor pricing data while “Moderns” will automate the process and deal with it in a few hours.
2. Redefining Strategy, Governance and Revenue Models
Generally speaking, pricing guidance and strategy formulation is extremely poor. As data gets cleaner and flows through the organization, visibility improves, time to action gets streamlined and decision outcomes can be measured objectively. We are strong believers that data maturity (see below) is the only way to address internal bottlenecks and to free-up energy and value for the organization. By automating low value workflows and leveraging analytical thinking through the organization hierarchy, from the strategic to the tactical layers, several improvements can be isolated:
1) Strategy and guidance formulation improves through measurable objectives (plural is important), granular scenarios definitions, and actions expressed as rules that can then flow quickly into the system. Modern pricing also reflects an elevation of the pricing topic to the C-Suite: several companies have repositioned their pricing function towards a “Net Revenue Growth” function (Coca-Cola, AB-InBev, Unilever, Danone to name a few).
2) Governance takes a different shape, when human and machine roles blend: humans addressing high value-added tasks and machines supporting the tactical layer with optimization and automation. In that sense, the pricing tactical layer, in high data and analytics maturity contexts, evolves towards trading, as observed in finance for example.
3) Lastly, extra time and visibility on high value activities allows companies to efficiently blend different revenue models, subscriptions and dynamic pricing, such as UBER recently announced for example.
3. Data Science and Literacy
Pricing is highly complex. This complexity can be exciting and very frustrating at the same time. Data science plays a very important role in addressing complexity, as a model is a simplification of reality. We usually say in statistics that all models are wrong, but some are useful. In the last years, model building has considerably evolved. The first trait of this evolution is universality, as the increase in open source languages adoption (R, Python) turns models and algorithms into a universal language, generating flows of exchanges and a massive global community. It is now highly probable that the problem your company is struggling with has already been partially addressed somewhere else. The other facet of the change
is democratization of model building and consumption, reshaping data organization towards senior data scientists sitting in a centralized Center of Excellence, owners of data product design, and an army of data translators and analysts in the functional areas, rolling out models and predictions across the organization.
4. Laser-Focused Deployment
Software is dead. Long live DevOps! Don’t get me wrong. We use more software than ever, but deployment has evolved and software are now consumed as apps, APIs, web interfaces, etc. Development has evolved too. Agile and DevOps are king and microservices are the latest hot thing
putting pressure on architects and IT executives to change the way they used to work and deliver value to the organization. We use the term of Acceleration Bricks as a way to define data-driven process reengineering.
A business process can be accelerated by the use of the right combination of models, packaged in the more efficient format (web interface, Rest-API, Java script, etc). How to get there?
An Acceleration Path for Deployment
This is where change management, impact delivery, governance and strategy come to play a crucial role. A data or tool-only approach can only fail. On the shelf solutions are the same, as pricing commands high levels of customization and configuration. We think that end-to-end transformations should follow an exponential shape as described in the figure below:
The acceleration starts by establishing solid foundations (data preparation, visibility, insights) in a very quick turnaround, followed by the traditional “quick wins” capture. What comes next is central. Execution should be the first focus of deployments. Pricing is all about control and performance measurement at the beginning of the journey. Installing both bricks sustainably into the systems and processes marks all the difference. Then come the times of acceleration, focusing on democratizing access to insights and recommendations through pieces of a custom toolkit, enriching models and data step-by-step, automating processes and workflow towards a trading mindset, and aiming at large-scale optimization.
Pricing is a long, exciting and rewarding journey, where methodology and expertise make a big difference. Go modern!