Lenss for product managers and product leads

Send engineering a PRD that has already survived its hardest review.

Lenss reads your PRD the way a skeptical staff engineer and a CFO would, before they do: what it takes for granted, where scope quietly grows, and which trade-off it never names.

The moment this is for

A PRD gets one first read from engineering leadership. If the user assumption is vague or the scope hides a second project, that read becomes the review, and the quarter starts with a rewrite.

What it catches

Vague user assumptions

Claims about what users want or will do, quoted from your PRD, with the cheapest way to check each one this week.

Scope that hides a second project

Requirements that sound like one feature but need their own design, data or migration to ship.

The trade-off nobody wrote down

Two goals in the document that pull against each other, each tied to the sentence that states it.

The number the decision turns on

When your PRD gives the figures, the exact value of the riskiest input at which the plan stops being worth building.

An example, both ways

The same draft read by a standard AI assistant and by Lenss.

Stakes$147,000 MRR · 3,000 accountsAdversarial Review · 3/4

Source draft

467 chars

Memo: moving from flat $49 to usage-based pricing

Head of Growth · for exec review

Today. 3,000 customers pay a flat $49 per month, which is $147,000 MRR. Customers make 3.6 million API calls per month in total, and the top 10% of customers make 70% of all calls. Proposal. Replace the flat plan with a $19 platform fee plus $0.02 per API call. The typical customer makes 1,500 calls a month (median of our top 500 accounts), so the typical bill stays at $49 and revenue holds. Rollout. Migrate every account at the next renewal. No grandfathering.

Adversarial review

pinned · quoted · computed
01

Pinned objective

Move to usage pricing without losing revenue or the accounts that carry it.

Underneath: Revenue is set by total calls across all 3,000 accounts, not by a typical one.

02

Hidden assumption

The top-500 median stands for every customer. MRR is the platform fee times all accounts plus the price per call times all calls, and the memo states both totals.

“The typical customer makes 1,500 calls a month (median of our top 500 accounts), so the typical bill stays at $49 and revenue holds.”
verbatim · highlighted in source

If it’s wrong: MRR falls 12.2% on day one, while bills for the top 10%, who make 70% of calls, rise from $49 to $187 at a forced renewal.

03

Flip threshold

Price per call at which MRR holds at $147,000

$0.025memo proposes $0.020
Figures behind the threshold
Platform fees: 3,000 × $19$57,000
Usage at $0.02 × 3.6M calls$72,000
New MRR vs $147,000 today$129,000 (−12.2%)
Top-10% bill: 8,400 calls avg$49 → $187

At $0.02 the memo cuts revenue by $18,000 a month before any churn, and concentrates the increase on the accounts that matter most.

Cheapest test: Re-price last month's actual invoices under the new model, account by account, and read the top-300 bills before the exec review.

Read the full teardown: Usage pricing priced off the wrong median →

Documents you bring

  • — PRDs and one-pagers
  • — Deprecation and sunset plans
  • — Experiment briefs
  • — Pricing and packaging changes
  • — Launch go / no-go memos

Jobs for this role

Run it on your own draft.

Free, no account. The draft is never stored. A full audit takes two to three minutes.

Audit your draft