What use is a value model? (REDCAPE)
Value models are a way to formally measure the value of a solution relative to the alternatives. They are foundational to value-based pricing and market segmentation and are important in product and service design. There is even an emerging discipline of value innovation that tries to organize innovation around how it will create value for stakeholders.
There are three main aspects of value to be concerned with.
Economic - the impact that a product or service has on a customer or customer segment's business. This can range from growing revenues to reducing costs and operating capital, to more efficient use of capital assets, risk reduction and the creation of future options.
Emotional - the impact on how buyers and users experience the world and their emotional response to the product or service. "How does this product make me feel?"
Community (or social) - an increase in positive externalities or a decrease in negative externalities. An externality is "a side effect or consequence of an industrial or commercial activity that affects other parties without this being reflected in the cost of the goods or services involved, such as the pollination of surrounding crops by bees kept for honey."
What is a value model?
Value models are generally used to quantify the economic value. There are several approaches to this (remember Scott Page's advice to use several models to inform any complex decision, value innovation and pricing includes many complex decisions). A conventional approach is Total Cost of Ownership (TCO) or Return on Investment (ROI) calculators. Net Present Value (NPV) calculations extend this to include risk (which is what the discount rate measures). In pricing work, we most often use the Economic Value Estimation (EVE) approach developed by Tom Nagle and his co-authors in The Strategy and Tactics of Pricing. The structure of an EVE style value model is shown below. Note that this is a model of value, it is not a pricing model.
The building material for EVE style value models is value drivers. A value driver is an equation that states how a solution impacts one of the six categories of value driver: Increase Revenue, Decrease Operating Costs, Decrease Operating Capital, Decrease or Defer Capital Investment, Decrease Risk and Increase Options. Value drivers can also be organized around metrics, such as increasing Customer Lifetime Value or improving Net Dollar Retention. I will look at alternate ways of organizing value drivers in a follow up post.
Some simplified examples of revenue value drivers and cost value drivers follow.
Revenue value driver
Imagine a new AI that allows one to create content that will attract higher quality prospects. There are two value drivers. The conversion rate per lead goes up and the size of the deal closed goes up.
One point to remember, it is not the revenue that creates the economic value but the gross profit. This is called out specifically in the value drivers as we want to have gross profit per deal as one of the variables in the model.
Gross Profit Increase = Change in Number of Leads Closed x Revenue per Lead x Gross Margin
Cost Value Driver
The same AI application can also generate the content. To build the value driver one needs a concept of 'content unit' which could be a video, a webinar, a blog post and so on. In the real world, this can be more complex as generally more than one content unit combines to generate a sales qualified lead (lead scoring systems can be repurposed into value models).
Applications like Frase do this, generating SEO optimized content - see end of this post for some notes on Frase's pricing).
Cost of Content Creation = Content Units x Hours to Create Content Unit x Change in Hours to Create Content Unit
One can then connect the two equations by looking at the number of leads generated per content unit.
The value model is a system of equations (the value drivers) that estimates economic value provided over time. Variables and constants are shared across the value driver equations used in the model. Value models can be built for a specific customer or for a segment of customers that get value in the same way (that all have the same value drivers). In this case, one generally uses the same constants across all the customers in the segment and an average value for the variables. Of course averages are false friends in pricing work, so one needs to look at the distribution of the values for the variable. If the distribution is, perhaps, bimodal, then in fact one is looking at two segments.
How are value models used?
In his book The Model Thinker, Scott Page proposes a simple mnemonic for all of the ways that models can be used. REDCAPE. Models can be used to Reason, Explain, Design, Communicate, Act, Predict and Explore. All seven of these are at play with value models.
Reason
The fundamental use of value models is to reason about the economic impact of a solution. By formalizing benefit claims as value drivers and then quantifying the value drivers by interviewing users and gathering data one can begin to ask questions about who a solution creates value for and how it creates that value. One can trace causal paths and ask what would happen if certain features are removed or enhanced.
In Ibbaka Valio, one can see the impact of turning a value driver on or off or of changing the variables in the value driver equations. This also makes it possible to test value with customers and to track value delivery over time.
Explain
By making explicit and clarifying the value drivers one can explain the intent of a solution and how it delivers value. These explanations are important to many people. Marketers use value models to inform their value propositions and to segment and target markets. Sales uses these models directly in the sales process with potential customers. Customer success uses the value models to document value being delivered. Product managers use the models to explain their intent to designers and developers, and to all those who work downstream. Finance can use value models with other models to make sure that innovation investments make sense.
When combined with customer journey maps, value models can help to explain how value is delivered over time. This provides insights critical to pricing design.
Design
When designing a product of solution one wants to make sure that it will deliver value. Value models help do this. One can begin by understanding the customer and the value you want to deliver and then using that to inform the solution and set the price. Once you understand the value, by using a value model, you can work out how much the solution should cost to develop, deliver and support. For every pattern there is an anti pattern, unfortunately in a lot of innovation work the anti pattern is more common than the successful pattern.
Communicate
Communicating value is a critical part of marketing, sales and customer success and is how value models are most often experienced. At Ibbaka, we use value models in the following ways:
- QUANTIFY the value of the offering (Value Model)
- PRESENT the value of the offering (Value Proposal)
- MONITOR the economic value realized by the customer (Value Pricing Dashboard, Value to Customer)
- COMMUNICATE the economic performance of the offering (Value Story)
Act
Models enable action. The actions a value model enables are what gives it meaning. Value models enable a number of critical actions.
- Investment decisions - Can we create enough value to justify the innovation investment?
- Design decisions - Can we design the solution to deliver the value?
- Pricing decisions - How can we price to fairly share value?
- Buying decisions - Should we buy the solution, will it deliver the value we need?
- Adoption decisions - How can we adopt the solution to realize the value over time?
- Renewal decisions - Should we continue to invest/ Are we getting the value?
Predict
At many times during the development, commercialization and sales process we make predictions using value models. We predict what value we can create. We predict how much of that value we can capture. These are used to calculate a value ratio, which is the Value to Customer over time or V2C (calculated using a value model) divided by the Lifetime Value of the Customer or LTV.
Value Ratio = V2C / LTV and to be viable V2C > LTV
During the sales process, one often makes predictions about the value that a customer will receive from a solution. A value model can do more than communicate that promise, it can be used to document the value received. Making predictions then testing them is the only way to get better at making predictions.
Explore
Models allow us to ask 'what if' questions. We don't do enough of this in value and pricing work. It is important to explore how value changes as the economy (for example inflation and interest rates), customer business models and competitors change.
At a basic level we can turn value drivers on and off, change the constants and variables in the value driver equations, and see how value changes. It will change in different ways for different customers and may even cause the market segmentation to change.
One can go further and develop alternative value models to explore and contrast different approaches to creating value for customers. Models can be combined to explore how solutions can come together and the value chain be reconfigured.
Value models are a powerful way to ground pricing in the customer's reality and make sure that product design, marketing, sales, customer success and finance are aligned around the customer.
As Peter Drucker said, "the purpose of a business is to create a customer."
Frase Pricing (Pricing in an Emergent Category)
Frase is part of an emerging category of applications that use artificial intelligence to generate content. Karen A Chiang and I will be using this emerging category to explore themes of value innovation and pricing at the Professional Pricing Society conference this October in San Francisco.
Frase will generate content that is optimized to be found by search engines. Its value promise is
"Frase helps you research, write, and optimize content that ranks 1st on Google – in minutes instead of hours."
So the value driver here seems to be 'save time.' Content creators are expensive, so if they can be made more efficient there are cost savings.
Is this the best value driver? I don't think so. The real value here is that the Frase AI (and other similar AIs) generates content that scores better with the search engines. It is more likely to be found, recommended and viewed. From there, depending on the purpose of the content, one can translate this into an economic value driver. Depending on the customer's business model one could have two revenue value drivers.
- Leads generated (and the value of a lead)
- Ads viewed (and revenue per ad)
This suggests two market segments, based on how value is created per customer, probably two different offers and two different pricing models.
But that is for the future, and will come into focus as the category emerges and defines itself.
What does Frase's pricing look like today (and a tip of the hat to Frase for publishing their pricing)?
This is the beginning of a tiered pricing architecture, but it seems provisional and incomplete.
When I went through the buying process and selected for an internal development team, a company with 6-50 people, with my primary interest AI content generation I got the following.
There does not seem to be a free trial, which surprised me for an innovative solution that needs proof points.
There is a fee for the basic platform. and the packages are fenced by
- Number of users
- Number of documents
- AI characters per month (which is capped at 10,000 per month across all tiers)
As fences these seem like a reasonable starting point.
One only wants to use per user subscription fees when each individual user needs a unique view of data or their own workflow. I think this is true of writing (others may disagree and see management of AI generated content as a purely organization thing).
One needs to account for both articles (and most of the economic value drivers are likely to be per article or content unit) and the length of the articles, so having the article and AI characters fences makes sense. Note that the per article fence only applies to the Basic Plan. I would prefer some sort of formula that combines the articles and characters (or words) rather than having them separate, but it will require some exploration to get this formula right, and it may well vary by use case (which would be good as it would reinforce the fencing between offers).
I believe that AI generated content is in the exploratory phase of innovation. Value propositions for the new category are still being established and proof is weak. During this phase one want to price to allow for new and unexpected uses. One wants to keep barriers to entry for new users low. At the same time, one wants to begin charging for the service. Charging for the service forces a buying decision and understanding what drives buying decisions is a critical part of defining a new category. It can also encourage more systematic use. Frase seems to be in the ballpark, though (i) a free trial of some form would help gather more information, (ii) the pricing may be on the high side and prevent users from trying, especially as there seems to be no free trial, (iii) more work could be done on integrating the fences and keeping things simple.
As the AI generated content category matures, and I expect this to happen quickly, packaging and pricing is likely to change, value based pricing get adopted, and prices increase sharply.