Approvals exist to protect two things reps cannot see alone: margin and risk. Done well, they stop a careless discount or an unenforceable clause before signature. Done poorly, they add days to every deal, train reps to route around the system, and leave finance guessing at the real average selling price. This guide walks you through the full build: setting thresholds anchored to real cost data, naming the right approver at each tier, writing routing rules that react to deal size and term length, and locking in the SLAs that keep approvals from becoming the slowest step in your sales cycle.
Time: 120 minutes
Every sales approval policy sits on three axes: how much is being discounted, how far the terms stray from standard, and how large the deal is. Build the matrix on paper before you touch the CRM. Along the first axis, list discount bands anchored to your floor price, not list, because list-based bands lie as soon as a rep bundles products. Along the second axis, list the specific non-standard terms that trigger review: net-sixty or longer payment terms, waived auto-renewal, custom SLA credits above five percent, mutual indemnity caps, uncapped liability clauses, country-specific data processing requests, and anything your legal team flags as a repeat offender. Along the third axis, segment deal size by annual contract value brackets that match your segmentation, usually SMB, mid-market, and enterprise. The intersection of these axes is your approval map. A fifteen-point discount on a two-hundred-thousand-dollar deal with net-forty-five terms is a different decision than the same discount on a twenty-thousand-dollar deal with standard terms, and the system has to treat them differently.
Most approval systems fail inside six months because they are built on list price. List is marketing fiction once bundling, multi-year commits, and competitive responses enter the room. Floor price is the number below which a deal destroys margin or signals a pricing problem, and it is the only number worth routing against. Pull your floor price from unit cost of delivery plus your minimum acceptable gross margin, calculated per product. For software with blended costs, pull it from your lowest acceptable annual contract value per seat across the last four quarters of closed-won data. Then define discount bands as the gap between what the rep is quoting and the floor. Zero to ten percent off floor can be rep-approved. Ten to twenty requires manager sign-off. Twenty to thirty requires a director or regional VP. Anything deeper requires the CRO or CFO depending on your governance. Publishing bands in floor terms also kills a destructive pattern: reps padding the quote with a fake list price so they can grant a discount on paper without touching the real number.
The single most common failure in sales approvals is routing to a role instead of a person. "Director of Sales" is not an approver. Sarah Chen is an approver. Role-based routing collapses the moment someone is on vacation, moves between teams, or leaves the company, and your approval queue backs up with no one accountable. Build the chain as a named person at each tier, with a named backup for every seat. The backup is active only when the primary approver has set an out-of-office flag or has not responded inside the SLA you will define later. For segments or regions with their own leaders, build parallel chains and let the routing rules pick the right one based on the deal territory. The result is a chain where every approval queue has a single owner at every moment, every escalation path is pre-drawn, and no deal sits waiting because the person who should answer has not been identified.
Translate the paper matrix into CRM routing logic now. Every rule reads from fields that are already required on the opportunity, which is why Step One of a healthy pipeline insists on required fields. The rule set should fire automatically when a quote is submitted for approval, read the discount percent, deal size, term length, and any non-standard terms flags, then route to the lowest tier that covers all three axes. If a deal triggers multiple tiers across axes, it routes to the highest tier that applies. For example, a twenty-percent discount routes to a director, but if the same deal has net-sixty terms, it routes to the director and the CFO in parallel because terms belong to a different approver axis. Keep the rule set declarative and visible, not buried in code. Reps and managers should be able to pull up any opportunity and see exactly why it routed where it did. Opaque routing is unlearnable, and unlearnable systems get worked around.
Discount routing is the easy half. Terms routing is where most teams drop the ball. The fix is to turn every non-standard term into a required picklist or checkbox on the opportunity, each one tied to a specific approver. Payment terms is a picklist with the standard options plus "net sixty," "net ninety," and "custom." Auto-renewal is a checkbox that defaults to on, and turning it off triggers legal review. SLA credit percentage is a numeric field with a default and a validation rule that routes anything above five percent to the head of customer success and the CFO. Indemnity cap deviation is a checkbox linked to legal. Each flag has a specific owner, a specific SLA, and a specific audit trail. The point is to make non-standard terms impossible to hide inside a redlined contract. A rep cannot route a deal around terms review by burying a clause change in a Word document attachment. The flag is the gate, and the gate is checked before the quote even leaves the CRM.
An approval policy without SLAs is a wish list. Every tier gets a response window measured in business hours, not calendar days. Rep-tier auto-approval is instant. Manager tier is four business hours. Director tier is eight business hours. VP, CRO, CFO tier is one business day. Legal review of non-standard terms is one business day for standard flags and two for anything requiring new redlines. Set these targets against your competitors, not against your approvers comfort. If your largest competitor can turn around a quote in four hours and your approval queue adds three days, you will lose deals you never knew were close. The SLA has three teeth: the clock starts the moment the request lands, the approver sees the countdown in their queue, and if the clock expires, the request auto-escalates to the named backup with a time-stamped record of the delay. Over the first quarter, pull every SLA breach report and share it with the executive team. If a particular approver chronically breaches, either the SLA is wrong or the approver is wrong, and the system makes the choice visible.
Every approval decision must be audited in a way that survives a finance close, a legal discovery request, and a renewal negotiation three years later. The audit trail stores five things per approval event: who requested it, what the request contained at submission, who approved or rejected it, what the approver changed before approving if anything, and a timestamp for every state change. Store this history on the opportunity itself, not in a side system, so a renewal rep two years later can see exactly why the original deal closed at the price it did. The audit trail also feeds a quarterly discount discipline report that pulls the median and ninety-fifth percentile discount by segment, product, and rep, flags the deals that approved on exception, and surfaces the approvers who authorize the deepest cuts. That report is the single most useful artifact in a pricing review meeting, and it does not exist unless the audit trail is instrumented on day one.
Resist the urge to light up every segment and every region on day one. Pilot on a single segment with a single approval chain for two to four weeks. Pick a segment where deal volume is high enough to stress-test the rules but low enough that an SLA breach does not break a quarter. During the pilot, every exception a rep raises becomes a test case. Did the rule fire correctly, was the right approver reached, did the SLA hold, did the audit trail capture the right data. Fix every issue in the pilot segment before you expand. When you do expand, do it one segment at a time, and give each expansion its own two-week stabilization window during which rules can be adjusted without a formal change-control process. Full rollout should take six to eight weeks for a mid-sized sales organization, and the time spent in pilot pays itself back the moment you avoid the alternative, which is a company-wide approval outage during quarter close.
Approval thresholds are not static. Pricing changes, products are added, segments mature, and competitors shift. Every quarter, pull the last ninety days of approvals and ask four questions. First, which thresholds generated more than ten percent of their tier approvals as rubber-stamps, because a tier that never gets rejected is just friction. Second, which thresholds generated more than twenty percent rejections, because a tier that constantly rejects is likely set too tight for the real market. Third, which approvers chronically breached SLA, and did breaches cluster around a specific product, segment, or deal size. Fourth, did any single rep account for a disproportionate share of exception approvals, which usually points to a coaching gap, not a policy gap. Make at most two threshold changes per quarter, and never change thresholds mid-quarter without a signed change-control note from sales leadership and finance. Discipline here is what prevents the slow drift that turns a well-designed policy into shelfware within a year.
Most organizations do best with three to five tiers. Rep self-approval covers low-discount, standard-terms deals. Manager, director, and VP tiers cover progressively deeper discounts and larger deals. A CRO or CFO tier handles the exceptions. More than five tiers creates routing confusion and SLA stacking that slows every meaningful deal. Fewer than three usually means the organization is pretending exceptions do not exist.
Use percentage against floor price for routing, and surface the dollar impact alongside it on the approval request. Percentage gives you consistent policy across deal sizes. Dollar impact gives the approver the context they need to make a sound judgment on larger deals where even a small percentage is meaningful. Approving on percentage alone invites rubber-stamping. Approving on dollar alone creates inconsistent policy as deal size varies.
Build an expedited lane with a shorter SLA, usually half the standard tier window, and require a documented reason from the requesting manager. Do not create a secret back channel. The expedited lane is visible, auditable, and tracked separately so you can tell whether urgency is real or routine. If more than ten percent of deals route expedited, something is wrong with your pipeline forecasting, not your approval policy.
Yes, with the modification logged in the audit trail. An approver who can only vote yes or no either rubber-stamps or rejects deals that could have closed with a small adjustment. Give the approver the ability to amend discount, term length, or non-standard flags, require them to document the reason, and send the modified terms back to the rep for final acceptance before the quote generates.
Every tier has a named backup with equal approval permission. The primary approver is required to set an out-of-office flag before an absence of more than two business days, which activates the backup automatically. If no flag is set and the SLA expires, the system auto-escalates to the backup with a time-stamped audit entry. Shared inboxes or role-based routing are not acceptable substitutes.
Review quarterly. Change at most twice per quarter and never mid-quarter without a signed change-control note. The data you need is rubber-stamp rate per tier, rejection rate per tier, SLA breach clustering by approver and segment, and exception approval concentration by rep. Thresholds that never get rejected are friction. Thresholds that constantly reject are miscalibrated. Both deserve a change, but neither deserves it in the middle of close.
Strkr gives you declarative approval routing, named-approver chains with backups, SLA countdowns, and audit trails that renewal reps can read three years later. Configure once, review quarterly, and keep your discount discipline report honest.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
