How to price your AI product
Outcome as a service is cool, but how do you define outcome?
In the rapidly evolving world of AI, startups face a unique challenge when it comes to pricing their products.
Traditional pricing models often fall short in capturing the true value of AI solutions, especially for outcome-as-a-service (OaaS) products - and on the other hand, you don't want to invent new ways of pricing and in the process, confuse the customers.
A look at AI startups website and you won't find any pricing details (they all ask to 'get into a demo') as I believe most of them are still iterating on pricing.
What exactly is Outcome-as-a-Service (OaaS)
Before diving into pricing strategies, it's crucial to understand what OaaS means in the context of AI products.
OaaS is a business model where customers pay based on the outcomes or results delivered by the AI solution, rather than for the technology itself or the time spent using it.
Key factors to consider in AI product pricing
Value Delivered: The most critical factor in pricing an AI product is the value it delivers to customers. This is especially true for OaaS models, where the focus is on outcomes.
Cost Structure: Understanding your costs, including development, infrastructure, and ongoing maintenance, is essential for setting a price that ensures profitability.
Market Dynamics: Research your competitors and understand the market's willingness to pay for similar solutions.
Scalability: Consider how your pricing model will scale as your customer base grows and your product evolves.
Customer Segment: Different customer segments may derive different levels of value from your product, which could influence your pricing strategy.
Pricing Strategies for AI Startups offering Outcome as a service model
1. Value-Based Pricing
This strategy aligns well with the OaaS model. Price your product based on the quantifiable value it delivers to customers. For example, if your AI solution increases a customer's revenue by $100,000 per year, you might price it at a percentage of that increased revenue.
2. Tiered Pricing
Offer different levels of service at varying price points.
This allows customers to choose the tier that best fits their needs and budget, while also providing upsell opportunities.
3. Usage-Based Pricing
Charge based on the volume of usage, or the number of AI-powered decisions made.
This model can work well for products where usage correlates closely with value delivered.
4. Performance-Based Pricing
Similar to value-based pricing, this model ties the cost directly to the performance of your AI solution.
For instance, you might charge a base fee plus a bonus for exceeding certain performance thresholds.
5. Subscription + Success Fee
Combine a base subscription fee with a success fee tied to specific outcomes.
This hybrid model ensures a steady revenue stream while also aligning your interests with those of your customers.
How to come up with pricing for Outcome as a service AI product?
When implementing an OaaS pricing model, consider the following steps:
Define Clear Metrics: Establish clear, measurable outcomes that your AI solution will deliver.
Define Clear Non-metrics: That is, metrics that you won't be judged on (i.e. outcome you do not have any control on. Read notes from Gokul Rajaram*)
Set Baselines: Work with customers to establish performance baselines before implementing your solution.
Implement Tracking: Develop robust systems to track and measure the outcomes your product delivers.
Align Incentives: Structure your pricing so that your success is directly tied to your customers' success.
Communicate Value: Clearly articulate the value proposition and ROI of your product to potential customers.
Challenges and Considerations
Measurement Complexity: Accurately measuring outcomes can be challenging, especially for complex AI systems.
Long-Term Value: Some AI solutions may deliver value over extended periods, which can complicate pricing models.
Risk Sharing: OaaS models often involve sharing risk with customers, which needs to be carefully managed.
Education: Customers may need to be educated on the OaaS model and how it differs from traditional pricing approaches.
*Notes from Gokul Rajaram on Outcome as a service AI product pricing
AI-Native Software Founders: I'm sure by now, you've all read the posts around how you should price your products based on end-customer outcomes (vs seats or other old-school methods).
While this is all good, I have a caveat for you. Be careful that you only price based on metrics that you have control over. For example, if you're building an AI SDR agent that's used by other companies as a replacement for human SDRs, the temptation might be to price based on appointments scheduled by the AI agent.
The challenge with this is that you don't have control over the product quality / differentiation, the competitive dynamics, or even the customer mindset when they get a message from you. For all these reasons, the right metric to price on is an intermediate metric under your control, such as "activities" (where an activity might be generating a lead list, composing a personalized email to a prospect, etc).
One can estimate how many activities a SDR completes per month, and the AI SDR agent should be able to tirelessly accomplish multiples of this number at a fraction of the price. And so hopefully, the cost per appointment should be lower than with a human SDR. But it's a folly to price on this.
A good analogy is search advertising. Google could have priced search ads as cost per conversion instead of cost per click, since Google has tremendous amounts of conversion data (from Google Analytics, Chrome, etc). However, ultimately the conversion rate on the website is impacted by many factors not under Google's control (product, offer, marketing message, website design, etc), and it doesn't make sense to price on conversions. Google reports the number of conversions per campaign, since they can track it, but they don’t price on it. Don’t confuse being able to track something with being able to influence or control it.
Tl;dr Figure out a leading indicator that's fully under your control, and price based on this intermediate metric, vs on the eventual outcome (via).
Conclusion
Pricing an AI product, especially one based on an outcome-as-a-service model, requires careful consideration of multiple factors. By focusing on the value delivered to customers and aligning your pricing with outcomes, you can create a win-win situation that drives both customer satisfaction and your startup's growth.
Remember, pricing is not a one-time decision. Continuously monitor and adjust your pricing strategy based on market feedback, customer insights, and your product's evolving capabilities.
With the right approach, your pricing model can become a key differentiator and driver of success for your AI startup.