Why most operator software is overpriced at the wrong things

The operator software stack costs more than it should because it's priced on volume, seats, or AI tokens — when the actual leverage is decisions per dollar. Here's what gets overpriced and what doesn't.

Most operator software is priced on the wrong axis.

The pricing models the SaaS industry settled on in 2010-2020 — per-seat, per-user, per-volume — made sense in a labor-priced era when each seat was a real human doing real work. They make less sense now that a meaningful share of the work is automated.

Here's what gets overpriced and what doesn't.

What's overpriced

Per-seat tools that don't scale with seats. A team chat tool at $15/seat/mo for a 25-person company costs $4,500/year. Same tool for a 100-person company costs $18,000. The cost-of-goods-sold doesn't scale 4x for the SaaS vendor; the customer's value doesn't scale 4x either. The tool is priced on a vector that doesn't track actual leverage.

AI tools billed per token or per generation. A content tool that bills $0.10 per generation costs $30/mo at 300 generations and $300/mo at 3,000. The leverage is in the consistency of output across all 3,000 — the same brief, the same voice, the same quality. Pricing per generation incentivizes the customer to generate less, which is exactly the wrong incentive.

"Enterprise" tiers that exist to hide pricing. When the only difference between "Pro" and "Enterprise" is a contact form, the actual price discrimination is happening in the negotiation. Buyers who don't know how to negotiate pay 2-3× more than buyers who do. The tool isn't worth different prices; the buyer's procurement skill is.

Marketing automation priced on contact count. Customer of 10,000 contacts pays $400/mo. Customer of 100,000 pays $4,000. The marginal contact does not cost the vendor 10× more; the customer's leverage from the marginal contact is usually less, not more.

What's correctly priced

Per-seat tools where each seat is genuinely doing different work. Salesforce per-seat for sales reps actually scales — each rep is doing the work that justifies the seat.

Usage-based tools where usage maps to customer outcome. Stripe taking a percentage of payments. Twilio taking per-message. The vendor's revenue scales with the customer's revenue; the alignment is correct.

Capability tools at a flat rate. A linter, a CI runner, a build system — these have a fixed value to the customer regardless of how often they fire. Charge a flat rate for the capability, not per-execution.

What this means for buyers

If you're evaluating software, ask: is the pricing axis tracking my actual leverage from the tool, or is it tracking a metric the vendor invented?

If it's tracking leverage, the price is honest. If it's tracking a metric the vendor invented, the price will scale faster than your value.

The third option — flat-rate pricing for capability — is rare in the operator stack. When you find it, it's usually the most honest pricing in the category, because the vendor is pricing on the work the tool does, not on the friction they can extract.

What this means for Merkava's pricing

Merkava prices flat per executive, not per seat or per output:

• $99-599/mo per exec (or full team), regardless of seats
• No per-output billing
• No "talk to sales" tier hiding price discrimination
• All pricing visible at withmerkava.com/pricing

The reason: the leverage is in having an exec, not in producing more outputs. A GROWTH exec that ships 8 articles is more valuable than one that ships 4 — but pricing it per article would punish customers who actually need the work shipped consistently.

Pricing for capability matches the customer's actual decision: do I want this exec running this function, yes or no. The price is the same either way.