Why We Chose Outcome-Based Pricing Over Traditional SaaS
Most SaaS companies charge per seat, per message, or per API call. We charge based on what conversations actually accomplish. Here's why we made that choice and what we learned.
The Problem with Traditional SaaS Pricing
When we launched earlier in 2024, we started with the standard SaaS playbook:
Pricing model:
- Starter: $29/month - 1,000 messages
- Pro: $99/month - 10,000 messages
- Enterprise: Custom - Unlimited messages
- Sent 8,500 messages in month 1
- Only 12 actually resulted in sales
- Paid $99 for mostly noise
- Canceled after 2 months
- Sent 450 messages in month 1
- 89 turned into qualified leads
- Paid $29, considered it cheap
- Upgraded to annual plan
- A business doesn't want 10,000 messages
- They want leads, sales, support resolutions, bookings
- Session opened, no messages sent
- 42% of all sessions (users just browsing)
- User asks questions, gets answers
- Reading knowledge base, checking FAQs
- 38% of sessions with messages
- Something actionable happens
- Lead captured, email sent, settings changed
- 16% of sessions with messages
- Sale completed, booking confirmed, payment processed
- 4% of sessions with messages
What happened:
Customer #1: E-commerce company
Customer #2: B2B SaaS company
The misalignment was obvious: Customer #1 got terrible ROI and churned. Customer #2 got incredible ROI but we underpriced.
Per-message pricing didn't correlate with value delivered.
The Insight: Outcomes, Not Outputs
One afternoon, our CEO looked at the data and said: "Why are we charging for messages? Nobody cares about messages. They care about whether conversations work."
It seemed obvious in hindsight:
The *conversation* is the product. The *outcome* is what we should charge for.
Designing Outcome-Based Pricing
We spent 3 months analyzing conversation data from 1,200 customers to understand what outcomes actually happened:
Tier 1: No Interaction ($0.00)
Tier 2: General Enquiry ($0.50)
Tier 3: Useful Outcome ($0.99)
Tier 4: Revenue Event ($1.99)
Key insight: The distribution is heavily skewed toward low-cost interactions. Most conversations are informational. Only a small percentage create significant value.
The Technical Challenge: Outcome Detection
Pricing by outcome requires accurately detecting what happened in each conversation.
Our approach:
1. Tool call tracking
Every action the AI takes is logged:
javascript
{
conversationId: "abc123",
tools: [
{ name: "search_knowledge_base", timestamp: "..." },
{ name: "search_knowledge_base", timestamp: "..." },
{ name: "send_email", timestamp: "..." },
{ name: "create_lead", timestamp: "..." }
]
}
2. External API calls
Calls to customer systems tracked:
3. Conversation metadata
javascript
{
conversationId: "abc123",
outcome: {
tier: "useful_outcome",
confidence: 0.95,
triggers: [
"tool:send_email",
"api:salesforce.create_lead"
],
value: 0.99
}
}
4. Manual overrides
Customers can mark conversations:
Customer Response
We A/B tested outcome-based pricing with 100 customers in June 2024.
Group A (50 customers): Per-message pricing
Group B (50 customers): Outcome-based pricing
The results were clear. Customers preferred outcome-based pricing because:
1. Predictable ROI - $1.99 charge = we made a sale
2. No penalty for efficiency - better AI doesn't cost more
3. Encourages exploration - users can ask questions without cost worry
4. Aligned incentives - we succeed when customers succeed
What Changed for Our Business
Revenue
Before (per-message):
After (outcome-based):
Product Development
Before: We optimized for engagement
After: We optimize for outcomes
Customer Success
Before: "How can we help you send more messages?"
After: "How can we help you get better outcomes?"
The conversation shifted from usage to value.
Challenges and Solutions
Challenge 1: Education
"Wait, so I pay per conversation? What's a conversation?"
Solution:
Challenge 2: Outcome Disputes
"This conversation should have been free, not $0.99"
Solution:
Challenge 3: Revenue Predictability
Harder to forecast revenue when it's outcome-based
Solution:
Challenge 4: Gaming the System
Some customers tried to avoid outcome-triggering actions
Solution:
The Results: 6 Months Later
We switched all customers to outcome-based pricing in August 2024.
Business metrics:
Customer feedback:
Product improvements we made:
Would We Do It Again?
Absolutely. Outcome-based pricing is harder to implement, harder to explain, and harder to forecast. But it's the right model for our business and our customers.
Why it works:
1. Aligns incentives - we win when customers win
2. Transparent value - clear what you're paying for
3. Encourages efficiency - better product = happy customers, not less revenue
4. Scales naturally - as customers grow, so do we
For Other Founders
Should you use outcome-based pricing? Questions to ask:
Can you accurately detect outcomes?
If you can't measure it reliably, you can't price on it.
Do different actions have different value?
If all usage is equally valuable, per-seat/per-usage is simpler.
Is value creation clear?
"We captured a lead" is clear. "We improved productivity by 10%" is not.
Are you willing to invest in the infrastructure?
Outcome detection, billing systems, and customer education require investment.
Are your customers sophisticated?
SMBs love outcome-based pricing. Enterprises sometimes prefer predictable per-seat pricing.
What We'd Tell Our Past Selves
1. Test it earlier - we waited too long to experiment
2. Start with simple tiers - we originally had 7 tiers, simplified to 4
3. Over-communicate - customers need more explanation than you think
4. Build dispute resolution first - critical for trust
5. Keep per-seat option for enterprises - some prefer it
Questions?
We're happy to share more about our pricing journey. Email us at hello@ilq.ai.
Want to see how outcome-based pricing works?
---
*This post is part of our series on building a sustainable SaaS business. Next: "How We Reduced Churn from 12% to 5% in 6 Months"*