A short, practical guide to mapping pricing decision rights.
If you do not actively tailor your pricing process, it tailors itself.
Not well. Every deal adds a stitch, every exception adds a seam, and nobody ever steps back to look at the whole garment. If that process made suits instead of prices, you would not leave the house wearing one.
Ask five people in your company who approves a 25 percent discount. Do it separately, in writing, and do not let them talk to each other first.
In most companies below a few hundred people you will get three different answers, and at least one of them will be a question.
Your price can be perfectly researched, well packaged and correctly positioned, and you will still lose most of it here, in the gap between the price you set and the price you actually charge. The problem most companies have is with pricing execution.
Having worked on more than 100 pricing projects, I see the same four symptoms almost every time we open the hood at a company that has grown faster than its commercial machinery.
The instinct is to fix this by buying something. A CPQ tool, a RevOps hire, a discount policy copied from a blog post. All three fail for the same reason: you cannot route decisions you have not named. Automate an unmapped process and you get faster confusion. (Frankly, it does not matter what tools you use to run your business. You can run a company using pen and paper - just as the folks did back in the past. And some managed to build large enterprises without computers).
So start with the map. It is unglamorous, it costs you half a day, and it is the highest-leverage thing most companies can do to their pricing this quarter.
Good. That makes this easier, not harder.
Decisions exist whether or not a process does. Somebody approved that 22 percent last Tuesday. The absence of a process means the decisions are invisible, and invisible is a worse place to be than immature.
It is also the wrong order to work in. A process is the sequence of steps by which a decision gets made. Build one before you know which decisions exist and who owns them, and you have drawn a flowchart with question marks in the boxes. Everything in this guide works better in a company with no process at all, because you are reconstructing what actually happened rather than auditing documentation that was optimistic the day it was written.
You do not need a CRM, a deal desk, a discount policy or a RevOps hire to do this work.
What this guide covers
Step 1. Find the pricing decisions that actually exist in your company.
Step 2. Find out which of them are broken, and in which of four specific ways.
Step 3. Assign each one, and decide in advance how you would know it is being ignored.
There is a companion spreadsheet built to run exactly these three steps. The guide explains the thinking. The sheet does the work.
One thing to get straight first. A pricing decision is not a process step. It is a moment where a human being chooses, and could have chosen otherwise. "Send the quote" is a step. "Approve 22 percent because the incumbent renewal is in three weeks" is a decision.
Every company's list is different. A company selling self-serve alongside enterprise has decisions that a services firm does not. A usage-based model creates decisions that a per-seat model never faces. Copying someone else's list, including ours, gives you a document that describes a company you are not.
So derive your own. Use the three probes listed below.
Pick a closed deal, ideally a messy one that people argued about. Write the list price on the left of a page and the first invoice amount on the right. Now reconstruct everything that happened in between.
Every point where the number changed, or could have changed and did not, is a decision. Write down what was decided and stop there. Do not yet ask who decided it.
This takes about an hour and it is the single best hour you will spend on this. It gives you decisions grounded in what your company actually does, rather than what your process documentation claims.
Discount is one of roughly ten things you give away, and it is the only one most companies govern. Go through the list and ask what you have conceded in the last six months under each heading.
Anything that changes what you get, what it costs you to serve, or what risk you carry, is a pricing decision even if nobody in your company calls it one. This probe is where the surprises are. In most engagements it uncovers more leakage than the discount column does.
The ones that generated internal argument, the ones that sat in someone's inbox for a week, the ones where a Slack thread got long.
For each, ask one question: what had to be decided here? You are looking for the recurring question. Every question that shows up more than twice is a decision that already exists in your company, documented or not.
Three probes will hand you somewhere between 40 and 60 candidate decisions. That is far too many. Nobody can hold 60 rules in their head, so they hold none, and the whole exercise dies quietly in week three with a beautiful spreadsheet nobody opens.
Keep a decision only if it passes at least two of these four tests.
Everything that fails goes on a parking list. You are not deleting it, you are deciding not to govern it yet. Keep the parking list visible, because the fastest way to lose trust in this exercise is for someone to notice their pet issue silently disappeared.
You should land between 12 and 20 decisions. If you have fewer than 10, you have not been honest with Probe 2. Go back and look at what you gave away for free last quarter.
The four tests ask whether a decision is worth governing. They do not ask whether it should exist at all, and some of what the three probes surface should be removed rather than routed.
The tell is that the decision is a symptom of a problem somewhere else. "Approve free migration" exists because onboarding is broken. "Approve deferred payment" exists because invoicing is unreliable. "Approve a discount above 50 percent" exists because nobody has ever worked out what those deals actually earn.
Govern a symptom and you make it permanent. From the moment it has an owner and a row in a register, it looks handled, and nobody goes looking for the thing that caused it.
So allow yourself a third verdict alongside govern and park. Eliminate: the correct answer is always no, and the fix is to remove the ability rather than build a tidy path to it.
The obvious way to run this step is a checklist: does this decision have an owner, yes or no. Do not do that. It is too coarse, and it will make you feel considerably better about your company than the evidence supports.
There are four ways a decision right fails, and they need different fixes.
Nobody decides. The thing simply happens. Scope reduction is the classic example: a seller quietly drops a module to close a deal, and no approval was ever sought because there was nothing to approve. Missing decisions are invisible in every audit that starts from the approval process, because they never enter it.
Everyone assumes it belongs to someone else. Sales says it is a finance call, finance says it is a commercial call, and the CEO assumes both of them have it covered. Orphaned decisions do not go unmade. They default to whoever is under the most pressure in the moment, which is reliably the person with a quota and a quarter closing on Friday.
Two or more people each believe the decision is theirs. This is the worst of the four and it produces the behaviour you probably recognize: an approval is given, then overturned, then escalated again by whoever is least willing to drop it.
Contested decisions do more than slow you down. They teach your commercial team that an approval is an opening position rather than an answer, and once that is learned it is very hard to unlearn.
Someone owns it who should not. Usually because they happened to be the only person available when it first came up, and it stuck. The sharpest version: the approver carries the quota on the deal in front of them. Self-approval under quota pressure is not governance. It is paperwork with a signature block.
This is the cheapest useful diagnostic in the whole of pricing governance. It takes 20 minutes to set up and it will tell you more about your company than a week of interviews.
It is also the point where the two reactions described at the start of this guide get decided. You are asking who decides, not who is to blame, and the survey has to read that way. If it reads like an audit, the answers you get back will be the careful ones, and careful answers are worthless here.
Pick five people: two sellers, a sales leader, someone from finance, and whoever runs revenue operations. Send each of them your decision list separately, in writing. One question against every row.
Who decides this today?
No meeting. No group thread. No "let me check with someone first". And do not tell people you are testing whether their answers agree, because if you do, they will make sure the answers agree. Then compare what comes back.
The disagreement is the output. Do not resolve it inside the survey and do not tidy the answers before you show them. A page of contradictory responses, in your colleagues' own words, ends an argument that no slide about governance ever will.
Worth adding to the same survey
Ask every seller one extra question: what is the largest discount you can approve on your own, without asking anyone?
The spread in those answers is usually wider than anything else you will measure this quarter. I have seen a five-person sales team give five different numbers, ranging from 5 to 30 percent, at a company that was convinced it had a discount policy.
Six fields per decision. Not more. The moment this becomes a twelve-column matrix, it becomes a thing that gets maintained instead of a thing that gets used.
Field 3, as almost everybody fills it in, says one thing: above this number, someone more senior decides. That is an escalation boundary, and a register built only out of escalation boundaries is a ladder with no top. There is always somebody who can say yes.
Every decision needs the second kind as well.
Write both, in the same cell, and write the hard limit even when it feels theoretical. "Up to 25 percent with the CRO, and nothing above 40 percent leaves this company as a discount" does work that the first half alone cannot do.
One warning about the numbers themselves. Every threshold you publish is an advertisement. Put "above 50 percent: CRO and CFO" into a widely circulated matrix and you have just told the sales floor that 60 percent exists somewhere in the building. The top rung of a discount ladder gets read as a menu, not as a control. Keep the deep end with the people who approve it, not in the version that goes to everyone.
An 80 percent discount is the usual test case. Sometimes the company genuinely wants that deal: a lighthouse logo, entry into a new market, blocking a competitor out of an account that matters.
The answer is not a taller ladder. It is to move the decision out of pricing altogether.
Treat it as an investment rather than a discount. A written business case, a named sponsor, an expiry date, and a cap on how many you will do in a year. Two things change, and both matter. It stops setting a pricing precedent that the next buyer can quote back at you, because it is not a price, and it lands in your accounts as what it actually is, the cost of acquiring that particular logo.
A blanket ban you cannot hold is worse than an expensive, visible route. Rules that get broken once teach everyone that the rules are theatre.
Go through your list and mark honestly which kind of control you have on each row.
Preventive — strongest. The system will not let it happen. Quote validation blocks an out-of-policy price before it can be sent. Expensive to build, and you should have it on maybe three decisions, not twenty. Reserve it for the ones where a single mistake is very costly or genuinely irreversible.
Detective — good enough for most rows. You will see it afterwards. A monthly discount distribution, an exception log somebody actually reads, a reconciliation of approved price against invoiced price. Cheap, fast to stand up, and sufficient for the majority of your list.
Declarative — not actually a control. It is written down and you hope people follow it. Most companies have twenty declarative rules and call the collection governance. A declarative rule on a high-value decision is an unmanaged risk with a document attached to it.
What you need is an honest column.
One. Not five. A signal you check monthly beats four signals you meant to build.
Before the room. Somebody pulls the last 20 non-standard deals and walks one of them from list price to invoice. That is two hours of preparation and it is not optional, because without it the workshop becomes a discussion of opinions. Send the naming test three days ahead and collect every answer before anyone talks to anyone.
Do not write this alone and then present it. A decision rights map written by one person is a proposal that will be negotiated for the next six weeks. The same map written by five people in a room is a decision.
Completeness feels like rigour and it is the opposite. Sixty rules is zero rules, because nobody can carry them. Cut to 15 and park the rest in writing, where people can see they were not ignored.
UNRESOLVED is an honest answer in the room and a lie three months later. Put a name and a date against every unresolved row before the meeting ends, or accept that it will stay unresolved permanently.
The register needs a named owner and a quarterly review. One question on the agenda: what did we approve this quarter that the map did not anticipate? That question is how the map stays true.
This single behaviour invalidates more pricing governance than the other three combined. It does not read as an override to the person doing it. It reads as being helpful, unblocking a deal, backing the team.
But if the founder is reachable, the ladder is decorative. Everyone learns within a month that there is a faster route, and after that you do not have a decision rights map. You have a document about one.
The companion spreadsheet
Four working tabs, one per step, plus a reference library of common pricing decisions to check your list against once you have built your own. Import it into Google Sheets and it works as it is.
Step 1. Decision inventory with the four-test filter scored automatically.
Step 2. Naming test collector that diagnoses missing, orphaned, contested or misplaced on its own.
Step 3. The register: six fields, control strength, signal, owner, review date.
Reference. Around 50 common decisions, as a prompt list. Use it last, never first.
Get the decision rights workbook →
Pricing chaos does not feel like chaos from the inside. It feels like being responsive.
Every individual exception looks reasonable. Every escalation looks like good service. Every override looks like a leader unblocking their team. From the inside it looks like a company that moves fast, and that is exactly why it survives so long without being named.
You only see it in aggregate, and by the time you do, you have already taught your buyers that your price is an opening position.
The companies that get this right are not the ones with the strictest rules. Some of them have very few. They are the ones where any seller can answer, in five seconds and without asking anyone, the question of who decides. That answer is worth more than any policy document you will ever write, and it costs you an afternoon.
%20(1).png)
Maciej is a General Manager at Valueships and data-driven problem solver with strong SaaS experience who helps companies unlock higher profits and revenue growth. He leads a team of highly talented analysts who deliver valuable insights through comprehensive data analysis. He has contributed to over 100 pricing projects that have collectively unlocked tens of millions in additional monthly profit for clients. His core strengths include analytical thinking, statistics, and transforming complex data into actionable business strategies.
Schedlue a free consultation
Maciej is a General Manager at Valueships and data-driven problem solver with strong SaaS experience who helps companies unlock higher profits and revenue growth. He leads a team of highly talented analysts who deliver valuable insights through comprehensive data analysis. He has contributed to over 100 pricing projects that have collectively unlocked tens of millions in additional monthly profit for clients. His core strengths include analytical thinking, statistics, and transforming complex data into actionable business strategies.
