Pricing is one of those things that looks simple from the outside (hopefully), yet is anything but from the inside.
At Fin, pricing and packaging (P&P) is more than a finishing touch. It’s a research problem, a forecasting challenge, a commercial decision, and ultimately, a strategic statement, requiring deep cross-functional work. We must balance the needs and wants of our customers, the value delivered by our product, and the broader vision we are building towards.
Our approach to developing P&P has evolved over the years, but it is still very much a living system.
The process
Pricing and packaging work is typically triggered when we launch something new that needs to be monetized, like Fin, our AI Agent. From there, we move through a sequence of research, analysis, and decision-making. The work is shared primarily across product, research, and data science.
Step 1: Foundational research
We start by talking to buyers to understand the world they live in. How do they define value? What does ROI look like? What pricing models do they expect in this category? What feels intuitive, and what feels off? This is deep qualitative work.
It informs two of the most important decisions we make: the pricing model and the pricing metric.
- The pricing model is the overall structure; value-based, usage-based, access-based, fixed fee, and so on. With Fin, we chose a value-based model: you only pay when Fin delivers value. Our research clearly showed that buyers don’t want to pay for usage, they want to pay for results.
- The pricing metric is the unit of value within that model, the unit we anchor pricing to. For Fin, the pricing metric is “outcomes.” An outcome is defined by Fin successfully handling a customer service query.
The details really matter. A small change in how you define the metric can completely change how customers experience value.
Our goal is to land on the model and metric that best aligns to buyers’ goals, needs, and expectations. That requires a lot of interpretive work. We can’t expect buyers to know the right price model for our products; they share how they evaluate value, and we deduce the best fit model and metric.
At every stage, we bring in input from exec, finance, GTM, and engineering. Not always formally, but enough to ensure alignment before moving on. Pricing decisions cut across the business, they can’t be made in isolation.
Step 2: Willingness to pay
Once we have a model and metric, we move to price point. This is where we run quantitative willingness-to-pay (WTP) research.
We use the model and the metric as the foundation from which we ask WTP questions.
- Example survey question:
- You would only pay when Fin delivers an outcome (→ the model).
- An outcome is counted when the AI Agent resolves a customer query with no further help needed (→ the metric).
- Would you be willing to pay $X per outcome for Fin?
The foundational qual is so important as a first step. It helps us decide what we should be asking about before we start asking how much people will pay. Without the qual ground work, you risk building a very convincing answer to the wrong question.
The goal isn’t to find a perfect price. That doesn’t exist. The goal is to ground our discussions in the reality of the market.
We use methods like Gabor-Granger and Van Westendorp to understand WTP. These studies produce outputs a bit like this (illustrative data):
This chart shows us what percentage of the market is willing to buy the product at various price points. The demand curve shows that 69% of buyers were willing to pay for the product at $0.86 per outcome, whereas only 39% were willing to pay at $1.42.
The dashed line shows the price point at which revenue for the business would be maximized (by multiplying adoption by the dollar amount).
This allows us to debate knotty questions like:
- What’s the right balance between growth and revenue?
- How sensitive is demand to price changes?
- At what price do we start losing the market?
- If we wanted to increase adoption, would lowering our prices by $X make a meaningful difference?
The data help us balance both customer and business needs.
This is where price point decisions become more tangible, but it’s not where the price point is finalized.
Step 3: Modeling
- A model.
- A metric.
- A strong signal on willingness to pay.
Now we translate that into something that works in the real world to land on a final price point. This is where data science and finance come in.
We take the WTP data and layer it with the realities of the business: margin expectations, segment differences, discounting behavior, purchase friction, sales capacity, competitor pricing, and other commercial constraints. This process requires extensive modeling.
What this looks like for Fin:
- We start with a list price aligned to strategy and commercial goals.
- Then we adjust for likely discounting to estimate actual realized price (e.g. $0.99 might look more like $0.89 in reality).
- Next, usage. We use beta data to understand how many outcomes customers generate on average and how that varies by segment, producing an average ARR per customer.
- Then combine usage data with WTP data to model attach rates (likelihood to buy), modeling a range of scenarios from conservative to optimistic.
- Finally, we feed these into the long-range plan with finance to estimate logos, ARR, and, importantly, margins. This is rarely a one-pass exercise.
We iterate until the numbers, and the story they tell, hold together.
The modeling step is important because willingness-to-pay data is somewhat theoretical. It reflects intent, not behavior. Modeling helps us bridge that gap.
The goal of this step is to land on a price point recommendation, alongside forecasts for ARR and adoption. It allows us to understand the real business impact of the decisions we’re making.
Alongside all of this, we need to ensure any decision we make falls in line with our pricing principles and broader business objectives.
Step 4: Sign-off
Finally, we bring everything together into a full pricing and packaging recommendation and take it to the execs. If it’s approved, it moves into build.
These four steps form what we think of as P&P strategy at Fin. In some ways, making the decision is the easy part.
Operationalizing it is harder: educating the sales team, rolling it out to customers, building tooling to demonstrate ROI, and ultimately, interpreting whether we got it right.
That operational work is substantial, and probably deserves a post of its own.
Do we do this every time?
No, this is the ideal process.
In reality, we probably follow this full process around half the time. Time and resource constraints are real, so we need to be smart about when we run the process thoroughly.
The rigor in our process should mirror the importance of the decisions we’re making. Some things just aren’t that important, and that’s fine. Or the answer is simply more obvious.
In practice, we often run scrappier research between stages when something feels uncertain. Pricing work is rarely linear. The process keeps us grounded, but it doesn’t make the work tidy.
The ongoing challenge
Over the last year or so, the breadth of our product has expanded significantly, which is a great thing. Fin can do a lot more than it used to. But it does mean our pricing system needs to evolve alongside it, and some of the early principles we anchored on (like simplicity) are being stretched in new ways.
For a while, we could price products in a fairly modular way. Each product had its own logic, optimized for the specific value it delivered. That worked well when the product set was simpler.
But as we’ve added more products, more Agent capabilities, and more outcomes, the question starts to change from “what is the right P&P for this one product?” to “how does everything fit together into a coherent pricing system?” That’s a very different kind of problem.
We must recognize that pricing isn’t something you set once and leave alone. As products evolve, especially in a world where AI is rapidly changing how value is created and delivered, it’s important to regularly step back and review the bigger picture, not just the component parts.
For example, outcome-based pricing has served us well, particularly when our products were tightly tied to clear, measurable outcomes. But as our products become more varied, and as we continue building toward a broader platform, it becomes less straightforward to apply a single model cleanly everywhere.
The challenge becomes less about replacing one model with another, and more about continually looking up and asking: what pricing philosophy best reflects the value we’re delivering today? And how do we deliver that philosophy in a way that still feels right for customers?
In short, there is no finish line, pricing is never “done” – and that’s exactly how it should be.
Why this work matters
Pricing and packaging is often noticeable only when it goes wrong. A confusing model, a bad metric, or a price that feels disconnected from value.
And we hear about those quickly.
But when it is done well, it’s almost invisible, and it quietly does a lot of work.
It shapes how people perceive value, helps buyers understand what they are paying for, and makes the product easier to sell, easier to buy, and easier to scale. It also forces the company to be honest about what the product is really worth.
That is why we take it seriously at Fin.