Ibbaka

View Original

What is digital pricing?

Steven Forth is a Managing Partner at Ibbaka. See his Skill Profile on Ibbaka Talio.

Note: This post is about the foundations of digital pricing. It will be of most interest to people who design pricing and the software and data management systems that support pricing design and execution. Looking for something of more immediate business relevance? Try one of these posts.

Digital transformation and digital pricing

Many companies are committed to digital transformation, driven by the way software and data are redefining their operating environment. As Marc Andreessen said back in 2011, ‘software is eating the world.’ And now, in 2023, artificial intelligence is eating software. Companies are under huge pressure to adapt and then to evolve.

Brian Solis defines digital transformation as

the evolving pursuit of innovative and agile business and operational models — fueled by evolving technologies, processes, analytics, and talent — to create new value and experiences for customers, employees, and stakeholders.

Note that the reason for digital transformation is ‘to create value.’ More on that later.

Digital transformation requires digital pricing.

Why is this?

Pricing is where all the different aspects of your business come together. Product and solution development, marketing, sales, implementation, customer success and finance are all critically impacted by pricing. In most companies, they also have a say in how prices are designed, pricing levels are set and how price is communicated and negotiated. It is the ultimate cross functional discipline.

To make this work, pricing has to be digital. Our definition of digital pricing is as follows.

Digital pricing is model based, data driven and adaptive, connecting price to value across the customer journey.

Let’s look deeper into each of these attributes.

Pricing is model based

Data only has meaning in the context of models, otherwise it is noise. Good models let us connect different types of data. At the same time, data can connect different models, giving us a holistic understanding of how things work together. This two step of data and models is critical to pricing work.

  • Data from different sources are used in one model

  • The same data is used in different models

This the basic structure of model based pricing.

What models are we concerned with? Three models interact.

  1. The value model

  2. The cost model

  3. The pricing model

The value and price models act as constraints on the pricing model. The value model sets a range within a price is seen as fair to the buyer and seller. The cost model determines the floor beneath which the seller cannot profitably operate.

Artificial intelligence is built on models and it is the model that is the intelligence. An application like ChatGPT is built on top of the Large Language Model GPT (Generative Pre-Trained Transformer). Pricing is built on systems of equations for value and cost that define the space within which the pricing model can be defined.

Pricing is data driven

Models are a fine and necessary thing, but they can only be built from data. One way to do this is through machine learning, in which the patterns in a set of data are discovered. In pricing we take a related approach. We gather many kinds of data and derive equations that connect the data. We then see if those equations an be used to make predictions about key SaaS metrics like Lifetime Customer Value or Net Dollar Retention.

What data are we interested in? A partial list follows (the list will always be partial as data uncovers new questions and generates new data).

  • General customer Data (and Models)

    • Operating model

    • Growth motion

    • Innovation platforms

    • Product platforms

    • Products

    • Solutions

    • Services

  • CRM Data

    • Market Qualified Leads (MQLs)

    • Sales Qualified Leads (SQLs)

    • Product Qualified Leads (PQLs)

    • Pipeline structure

    • Conversion rates

    • Pipeline velocity (how fast an opportunity moves through the pipeline)

    • Win/Loss data

    • Lead firmographics

      • Buyer role

      • Decision maker roles

      • Influencer roles

      • Industry

      • Location

      • Size

      • Profitability

      • Customer base (who the prospect sells to)

  • Financial Data

    • Contract Value

    • Contract Terms

    • Revenue components

    • Discounts (on and off invoice)

    • Renewals

    • Upsell

    • Cross sell

  • Usage Data

    • Number and types of users

    • User actions

    • Transactions

    • Value paths

    • Value path completions

    • Temporal patterns of use

    • Spatial patterns of use

  • Cost Data

    • Unit Costs

    • Unit Cost Components

    • Fixed Costs

    • Fixed Cost Components

    • Cost Scaling Parameters

    • Cost allocation rules

  • Market Data

    • About competitors

      • Pricing metrics

      • Pricing curves and pricing levels

      • Growth motions

      • Market share

    • About the customer’s customers

      • Business models

      • Operating models

      • Growth motion

      • Growth rate

      • Innovation vectors

    • About the Operating Environment

      • Growth rates

      • Input costs

That is a lot of data to gather together and organize. It can only be done using software to organize the data and to keep it current.

In general, one wants to abstract data into data types, turn these into parameters that can be used as variables in the models and then test performance across different values for the variables. One can then test different values for the parameters in order to better understand …

  • How each model behaves at different scales

  • Interactions between the three models

Make sure that the following relationship holds across all of the scales at which you are likely to operate.

Value > Price > Cost

Pricing is adaptive

The main reason to use models is that they can be tested and changed. One can test the model to see how well it describes the data on which it was based, one can make predictions, and see how accurate the descriptions are. The modeling approach fits well with machine learning. Machine learning approaches like deep learning can be used to build models and to make predictions and recommendations.

How do the models adapt?

This happens at three levels.

  1. Better estimates for the variables are found
    One often needs to start with guesstimates for each variable, but by calling out assumptions one can test and improve them.
    The goal is to have the SaaS application collect the data needed for the variables wherever possible.

  2. New equations are developed
    New value drivers can be found, new cost drivers uncovered.
    Situations or customers for which the diver (equation) is relevant are better understood and equations can turned on or off as relevant.

  3. Pricing and packaging are optimized
    Each package is designed to play a role in the overall offer architecture. Assumptions are made about the volume and revenue for each package. The assumptions are tested against actual market structure (which can change over time) and optimized according to the goal (volume, revenue or cost optimization).
    Price levels can also be adjusted to help packages meet their goals. The price of each package has to be set in the context of the other packages. The goal is to optimize for the system of packages and bundles rather than for any one package.

Applications of artificial intelligence is evolving rapidly. Beginning with traditional revenue and yield management systems we are now seeing a flowering of different approaches that are more focussed on optimizing configurations for the value they can provide to the customer. These approaches assume digital pricing.

Connecting price to value across the customer journey

Value and pricing are not static. Value is generally delivered over time, which is why subscription models are an effective pricing design. There are several common patterns for value realization. The pattern will determine the best way to price and how to organize sales, implementation and customer success.

Six Common patterns for Value over Time

It is easier to design for these different patterns if one takes a model based approach.

Two other things to consider when you think about value over time.

  • What data should be collected to document value as each touch point in the customer value journey

  • What value story should be told at each touchpoint

Of course the value story will evolve differently depending on the way value delivery changes over time.

Conclusion

With software eating the world and AI eating software most companies will be under great pressure to move to digital pricing. This means a move to model based pricing, where price is defined as part of a formal model. That model needs to cover more than just price. Value is also important; as is cost. The models are connected when they share variables.

Models need to be tested to make sure that the critical relationship holds across different scales: Value > Price > Cost.

Models without data are an empty exercise. Data is needed to develop, then test, then evolve the value, price and cost models. That data comes from many different sources and it needs to be gathered and updated at a regular cadence.

The guiding principle of digital pricing is to connect price to value over time. Doing this has many implications:

  • Pricing is seen to be fair

  • The risk discount that plagues SaaS pricing is reduced

  • Price increases can be targeted

  • Price and value can be managed over time

Digital pricing will require a new generation of pricing platforms, such as Ibbaka Valio.

Read other posts on pricing design