In October 2025 we launched Certyo with a hybrid subscription plus per-record pricing model. Customers paid a base fee and a variable rate for each durable record anchored. It was defensible on paper — usage-based pricing is the modern SaaS default — and it mapped to our underlying unit economics. By April 2026, after six months of pilot conversations and procurement reviews, we killed it. This post explains what the buyers told us, what we measured, and why annual volume bands are structurally better for both sides.
What buyers actually said
The per-record model ran into three consistent objections, and once we heard the same language three times we started keeping a tally. The objections were not about the absolute price. They were about the shape of the price.
The three objections, in the buyers' words: "per-record is finance-unfriendly," "we can't justify variable spend against a fixed-budget compliance control," and "volume bands map to how we already budget SaaS." The last one is the key. Finance teams spend most of their time in the same four or five pricing patterns. A novel pattern adds friction even when the absolute number is reasonable.
What we measured
We tracked three things through the pilot conversations:
- Pilot-to-paid conversion velocity — how long from first demo to signed MSA. Per-record deals took roughly twice as long as we'd originally modeled, and the delay was concentrated in the procurement and finance review, not the technical evaluation.
- Procurement review complexity — we watched buyers try to model variable spend in their SaaS budgeting tools. None of their tools natively supported per-record projections, so they built spreadsheets. Spreadsheet-grade procurement review is a signal that the pricing model isn't fitting.
- Questions asked that weren't about value — "how do we cap it?" "what if our volume doubles?" "what does the forecast look like?" None of these are bad questions, but all of them are about pricing mechanics, not about whether the product solves the problem.
The new model
The April 2026 restructure produced two SKUs with simple annual shapes. Self-hosted starts at $90,000 per year. Managed starts at $24,000 per year. Both use annual volume bands rather than per-record pricing. Overages reconcile at renewal, never mid-contract. The bands align to how finance already budgets SaaS, and the starting point is a single fixed line item.
We kept the things that mattered. No per-seat charges. No variable gas costs. No surprise mid-contract bills. The change was in the shape, not in what we charge for. Finance signs the annual line; engineering gets a predictable cap; legal reviews a single renewal term.
What we preserved from the old model
The per-record model wasn't wrong in its spirit. It reflected a real truth about our cost structure: we do pay gas per anchored record, and very-high-volume customers do cost more to serve. The restructure preserves that truth — it's just expressed as volume bands, which are finance-compatible, rather than per-record counters, which aren't.
No surprise gas costs. No per-user seat charges. Overages reconciled only at renewal, so mid-contract spikes don't generate mid-contract invoices. Those guarantees were always implicit in the per-record shape and are now explicit in the banded shape.
Who benefits from the change
The restructure changes the shape of the procurement conversation for three specific roles:
- Finance and procurement — Predictable annual line item. Fits the tools they already use. No spreadsheet-grade variance modeling required. Approval cycle typically cuts by more than half.
- Engineering leads — No chargeback dynamics internal to the business. If volume spikes, the budget doesn't spike mid-quarter. The cost of anchoring more records is a renewal conversation, not an invoice surprise.
- Compliance and audit sponsors — The integrity control is a fixed-cost line. That matches how compliance budgets are defended at the executive level — predictable spend on required controls, not variable exposure on regulated processes.
The meta-lesson
A pricing model that makes sense internally can still be the wrong model if it doesn't fit how buyers actually make decisions. We launched with the model that mapped to our costs. We switched to the model that mapped to their approval processes. That's not a concession — it's the job. A vendor that hasn't revisited its pricing model isn't listening to its buyers. For the current SKUs and FAQs, see /en/pricing.
The per-record model was defensible on paper. Annual volume bands are defensible in a procurement meeting. That's the one that mattered.