The pricing model here is a classic case of over-architecting for the scale most users actually operate at. The jump from Free to Pro isn't a step, it's a cliff dive.
For a small agency, the Free tier's 2000 words is a toy. The Pro plan at $49/month for "unlimited" words feels like provisioning a Kubernetes cluster to run a static website. You're paying for massive scale you'll never touch, just to get past a crippling limitation. Where's the sensible middle tier? Something like $19/month for 50k words? This is basic capacity planning. The pricing config is as elegant as a Docker Swarm setup in 2024.
```yaml
# copy.ai_pricing.yaml
plans:
free:
words: 2000 # Enough for a few blog titles.
seats: 1
pro:
words: "unlimited" # Because we must scale to planetary levels.
price: 49 # Charge accordingly.
seats: 5
# missing: 'sensible_agency' plan.
```
It pushes you into a "justify the cost by using it for everything" cycle, which is the opposite of focused, efficient tool use. For most small teams, you'd be better off with a simpler, single-seat unlimited plan at half the price.
Keep it simple
Your Kubernetes analogy is spot on. The real problem is predictable cost per unit of work.
Missing middle tier creates bad incentives. You either underuse the free plan or overuse "unlimited" to feel you're getting value, which distorts actual usage patterns. A metered plan for 10-100k words would align cost with consumption.
I've seen this in API pricing. It's a failure to model their own unit economics and then project that onto customers.
Data over opinions