Second
Opinion.
Your team drafted the pricing architecture, likely with AI in the loop. Before it ships, get the verdict: an expert-led, evidence-backed ship, ship with fixes, or do not ship.
Second Opinion is SPP’s expert review of a pricing architecture your team drafted itself: the three structural decisions (licensing model, packaging model, pricing model) that determine how a software company captures value. A named SPP practitioner pressure-tests the draft across five audit lanes, instrumented by LevelSetter where your deal records allow, and signs the verdict before you ship, with every fix ranked.
Smart teams draft their own pricing now, usually with AI in the loop, and the draft is a legitimate starting artifact. AI collapsed the cost of producing a pricing recommendation. It did nothing to the cost of shipping the wrong one: a mispriced metric reaching an installed base, and a repricing your market reads as an event.
When drafting is cheap, value concentrates in verification. Second Opinion treats your draft as the input, never the exhibit. Findings attach to evidence gaps and structural risk, and the review ends in a verdict you can act on. A competent draft shipping clean is an expected outcome, and the memo says so plainly.
deal patterns.
review.
fixes ranked.
A pricing recommendation used to take a quarter to produce. Now a capable team drafts one in an afternoon. The cost that did not move is the cost of shipping the wrong one.
Second Opinion is the verification step for the draft you already have: five audit lanes and a signed verdict, before a customer ever sees it.
Chris Mele
Ranked #1 on OpenView’s list of B2B SaaS pricing experts. Chris leads every Second Opinion review, surrounded by a team that has held CFO, CPO, and CIO seats inside software companies. The practitioner who signs your verdict runs your review, not an associate.
LevelSetter does the heavy lifting where deal records allow: ingestion, and the drafted architecture played against the deals you actually closed. The expert makes the calls only humans can make: what the evidence gaps mean, and whether the draft is ready to ship.
Read more about Chris →Five lanes,
one verdict.
Every lane examines a decision, because that is what a pricing architecture is. The decision layer is where a software company’s licensing, packaging, and pricing choices are made, distinct from the runtime layers (metering, billing, entitlement enforcement) that execute them. The draft in your hands is a set of decision-layer choices about to go live. The review reads each one before it does.
Licensing: the metric
and the grant
Whether the metric the draft selected is a true value metric, one that scales with the value a buyer receives and that the buyer can read and forecast, or a surrogate unit such as credits or compute units: vendor-defined and vendor-controlled. The lane also reads exposure at the far end of the usage distribution, and whether the definitions of user and usage still describe reality when agents do the consuming.
Packaging: boundaries
and Customer Groups
Whether edition boundaries map to Customer Groups and their use cases, or hide a usage threshold in disguise. Whether the packages compose from your actual capability inventory or from a good-better-best template imported from the public web, and whether a customer who outgrows an edition can see a legible next one.
Pricing: surface fidelity
and list to net
AI drafts characteristically produce three clean price points and silence between them, and the gaps are where deal desks improvise. This lane reads whether the model computes a defensible price at every configuration and volume the licensing and packaging decisions can produce, and whether list-to-net discipline exists: a pricebook worthy of the name, and a scheduled net at every volume.
Calibration: what the
draft was built on
The question a model cannot answer about itself: what evidence produced these recommendations? This lane inventories the draft against the nine lines of decision evidence below, maps which lines were missing, and traces which recommendations rest on the missing ones.
Governance and transition:
the customers you have
No public corpus contains your contract base, so drafts reliably skip this lane. It reads which existing accounts pay more under the drafted architecture, where the impact concentrates and at which renewal moments it lands, whether existing customers have a protected transition path, and who owns the pricing decision once the change is live.
Nine lines of evidence.
Most drafts saw two.
A draft can be thorough and still be calibrated on almost nothing of yours. The model never saw your losses or your contract base; it cannot report what it was missing, and it answers confidently anyway. The calibration audit reconstructs what evidence the draft actually rests on, line by line, before your customers test it for you.
Decision evidence is the full corpus a pricing decision runs on: win and loss events, the buyer’s choice set, billing, usage, and cost data, the roadmap, the value driver repository, validated customer perceptions of value, and expert judgment.
Whatever internal data your draft saw was almost certainly billing and usage exhaust. Billing data is survivorship data: every invoice is a deal you already won, so a draft calibrated on billing exhaust has seen winners only, and never the losses, collapsed quotes, and rejected configurations a pricing decision calibrates on. Two lines of nine, and the two written entirely by deals that went your way.
Win and loss events
Every deal you closed and every deal you did not. The losses carry the pricing signal, and no meter ever writes a row for a loss.
The buyer’s choice set
The configurations you offered and the alternatives each buyer weighed. It lives in deal records and debriefs, not in any runtime system.
Billing data
What invoices actually said. A legitimate input, and usually the only internal line an AI-assisted draft ever saw.
Usage data
How the product is consumed, and by whom. Written entirely by customers who already said yes.
Cost data
What each configuration costs to serve, across the volumes the pricing surface has to hold.
The product roadmap
What ships next quarter. Packages drawn without it are stale on arrival.
The value driver repository
The capabilities that move buyer economics, mapped to the Customer Groups that feel them.
Validated value perceptions
What customers say the product is worth, collected and tested rather than inferred.
Expert judgment across markets
Pattern recognition from markets your own data never saw. The line that lets a pricing decision borrow evidence at scale.
The calibration lane delivers the missing-evidence map: which of the nine lines your draft never saw, and which of its recommendations rest on the absent ones. It is the one-screen answer to “our AI draft was pretty thorough,” because thoroughness is about coverage of the questions, and calibration is about what evidence answered them. A model cannot self-report that, however good the draft.
A draft goes in.
A verdict comes out.
Hand over the draft
In whatever form it exists: a document, a spreadsheet model, a deck, or the AI transcript itself. The transcript is welcome and useful. The calibration lane reads the prompts and the evidence as much as the output.
Connect deal records, where you have them
Supply deal records and the review runs instrumented: LevelSetter ingests won and lost deals, line items intact, so the drafted architecture can be played against the customers you actually have. Without records, the review still runs; the simulation lane waits.
The five lanes run
A named SPP practitioner works the draft through licensing, packaging, and pricing, then the calibration audit and the governance read. Findings attach to evidence gaps and structural risk, never to how the draft was made.
The verdict is walked, not emailed
A working session delivers the verdict, the ranked fixes, and where the instrumented lane ran, the customer-by-customer impact. Your decision owners are in the room, so the fixes get argued and sequenced rather than filed.
Draft in hand and a ship date on the calendar? Twenty minutes with a pricing expert scopes the review against both.
Talk to an ExpertThe verdict is one page.
It reads one of three ways.
The memo is signed by the reviewing practitioner, with the reasoning traceable to lane findings. The verdict is the product; everything else in the review exists to support it.
Ship
The draft holds across all five lanes, and the memo says so plainly and without surprise. A competent draft shipping clean is an expected outcome of this review, and the signed memo becomes the evidence your board or your sponsor reads.
Ship with fixes
The architecture is sound and specific findings need correcting before launch. Each fix is tagged to its lane, ordered by the cost of being wrong, with a direction attached, so your team knows what to change first and why.
Do not ship
The review found risk that outweighs the launch date, and the memo names it. Do-not-ship is a full-value outcome: the fee is fixed at intake and buys the verdict, whichever verdict it is. A review that can only say yes would not be worth signing.
Yours to keep,
whichever verdict lands.
The signed verdict memo
One page. Ship, ship with fixes, or do not ship, signed by the reviewing practitioner, with the reasoning traceable to what each lane found in your draft.
The ranked fix list
Every finding tagged to its lane and ordered by the cost of being wrong, each with a fix direction. Directions, not designs: producing the corrected architecture is engagement work, and the memo states that boundary.
The Roll-forward simulation, where deal records allow
The drafted architecture played against your historical deals: customer-by-customer revenue impact, a forecast against the customer mix you actually have, and a customer-transition map for the accounts that need a designed path.
The verdict working session
The verdict is delivered live, with the people who own the pricing decision in the room. Questions get answered by the practitioner who signed the memo, and the fix sequence leaves the session agreed rather than pending.
Who Second Opinion
is for.
Right fit when
- You drafted it yourselves. An internal team produced the architecture: a document, a model, a deck, or an AI-assisted working session. The review needs a draft to read, and the draft you have qualifies.
- B2B software. The review and the pattern library behind it are built for software companies pricing software.
- A ship date is near. The verdict is worth most while the decision is still open and the launch can still absorb fixes.
- The decision owners will be in the room. The review requires the people who own the pricing decision, and who drafted the architecture, at the working session.
Probably not a fit if
- Looking for a rubber stamp. The fee buys the verdict, whichever verdict it is. If only one answer is acceptable, this is the wrong instrument.
- Expecting implementation. The review ends at the verdict, the ranked fix directions, and rollout guidance. Producing the corrected architecture or operating the rollout is engagement work.
- No draft yet. If the question is still what to build, score the architecture you have with the free Pricing Architecture Assessment, or run Pricing Ground Truth to see what your current pricing is actually doing.
- Pre-revenue or stealth. The review reads a draft against evidence, and most of the nine lines do not exist yet at that stage. The free Assessment scores the design itself in four minutes.
Already shipped it?
The same review runs in rescue.
Some teams arrive after the ship date, watching the drafted architecture bite: renewal pushback, and a deal desk improvising around the model. The five lanes hold; the verdict grammar shifts to hold, fix in place, or re-architect, and the working session moves to the front of the sequence because urgency is real. The job is stabilization, and the earlier the review runs, the more options stay open. If that is where you are standing, talk to an expert this week rather than next.