Comp Plan Complexity Audit: How Complex Are Your Plans, Really?

Complexity, Not Headcount, Should Drive the Purchase

Ask a finance leader how many people are on variable pay and you get a number in four seconds. Ask how complex the comp plans are and you get an adjective. That gap is why so many incentive compensation purchases land in the wrong tier: headcount is the only complexity proxy most buyers have, so headcount is what both the market and the buyer price on.

The expensive version of the error runs in one direction. A company has a calculation and crediting problem — payouts late, disputes constant, the process held in one analyst's spreadsheet — and buys a full suite: incentive compensation plus territory and quota plus sales planning plus analytics. The engine problem is solved in month five. The territory module is never configured, the planning module runs once a year, and the invoice carries all three. You paid a suite price and a suite implementation to fix an engine problem. Enterprise implementations here run roughly 50,000 to 150,000 dollars and up, so the surcharge is not marginal.

The reverse failure is quieter and often costs more. A team scores its plans simple because most of the plan is simple, buys a self-serve tracker, then meets the one feature the tool cannot express — a split derived from a CRM field, a clawback after payment, an effective-dated change to a paid period. The tool handles 90 percent of the calculation, the spreadsheet returns for the rest, and the shadow process the purchase was meant to kill is still running.

What follows is an instrument, not an essay. It takes about an hour with the plan documents in front of you, and it belongs before the first demo.

The Six Dimensions That Drive System Requirements

Plan complexity is not one thing, and the dimensions correlate less than buyers assume. Plenty of organisations run trivial rate mechanics and brutal crediting; plenty run single-plan simplicity and still need contract-level amortisation. Each dimension below is where a particular class of tool breaks.

Crediting structure

Who gets credit for a transaction, and how many times. Buyers describe the standard case and forget the exceptions: splits between co-selling reps, overlays credited on the same booking as the account owner, manager roll-ups, matrix crediting. The hard version is not the payee count but where the split rule lives — percentages typed into a field are configuration; percentages derived from account hierarchy or partner data are engineering.

Rate mechanics

How attainment becomes money. Flat rates are trivial, tiers are easy; the feature that separates tiers of tooling is retroactive true-up — whether crossing a threshold re-rates every dollar from the first, or only dollars above it. Retroactive tiering forces a system to restate prior periods whenever attainment moves, which period-by-period arithmetic cannot do cleanly. Add decelerators, caps interacting with draws and quarterly rate cards, and the entry tier is gone.

Timing and adjustment

When credit is earned, when it is paid, and what happens when it is taken back: draws, guarantees, clawbacks, quota retirement, and the choice of bookings, invoice or cash as the crediting basis. The multiplier is different bases for different purposes — retiring quota on bookings while paying on cash collected means every deal carries two states, and reconciling them is where disputes start.

Scope of variable pay

How wide the population is: multi-currency with an FX rate-lock policy, multiple statutory entities, MBO incentives for non-sales functions, and channel, distributor or loyalty programmes. Scope is most likely to be scored low and be wrong, because non-sales bonuses and partner rebates sit with other teams in other spreadsheets and nobody counts them as comp. If the new system does not cover them, those spreadsheets survive the purchase.

Governance

How often the rules change and how well changes are controlled: effective dating, retroactive change frequency, variant count, approval and acceptance. The number that matters is variants per role, not variants in total — three plans across three roles is a different problem from three variants of one role, because the second implies exceptions negotiated one at a time. The hardest capability here is re-running a paid period without corrupting the record of what was paid.

Finance requirements

What the close needs from the system: accrual reporting, cost capitalisation and amortisation under ASC 340-40, journal export by entity and cost centre, an audit trail with an adjustment register. This is where vendor claims deserve most scrutiny — capitalisation and amortisation is often asserted at a high level and thinly documented. Treatment of your own contracts depends on your facts and belongs with your auditors, not a vendor datasheet.

Scoring Your Plans: The Instrument

Score each dimension 1 to 5 against the anchors below, interpolating within each column's range. One rule decides every close call: score the plan you actually run, including the exception you were about to call an edge case. Where the anchor is a count, count it.

Reading the total: three bands

The total runs from 6 to 30. Simple is 6 to 12 — light tooling and a small team; you are buying transparency, rep self-service and the end of manual re-keying, and should refuse to pay for configurability you will not use. Moderate is 13 to 20 — a real calculation engine with plan versioning, dispute workflow and clean CRM integration; the most contested part of the market, which is good for price. Complex is 21 to 30 — crediting flexibility at scale, effective-dated recalculation, finance-grade controls, and an implementation with a named internal owner.

The bands are not evenly spaced in cost: Simple to Moderate is roughly a doubling of per-payee price, while Moderate to Complex adds an implementation line that often exceeds year-one subscription. Within two points of a boundary, treat the band as unresolved and let the rule below decide it.

Mapping Score to Vendor Tier

Analyst standing appears only where confirmed. Forrester's Wave on SPM Solutions for Incentive Compensation, Q1 2025 named Varicent and CaptivateIQ Leaders, and Everstage, Performio and Forma.ai Strong Performers; other placements in that Wave sit behind the paywall and are not asserted here. Gartner retired the SPM Magic Quadrant after roughly 2021 in favour of a Market Guide, so there is no current quadrant to cite. Pricing figures are signals, not quotes.

When the complexity is breadth, not depth

The fourth row deserves expansion. If your score is driven mainly by dimension four — variable pay for people outside sales, plus channel, distributor or loyalty incentives alongside the sales plan — the mid-market ICM tools handle your rate mechanics competently and leave you administering the MBO scorecards and partner programmes somewhere else. That is a coverage problem, not a tier problem.

Vulki by Akeron is the name worth adding to that shortlist. Akeron S.r.l. sits in Lucca, Italy, founded by former Tagetik people (Tagetik went to Wolters Kluwer in 2017); it is PE-backed by White Bridge Investments with approximately 30 million euros raised, including 12 million in July 2024, and has appeared four times as a representative vendor in Gartner's Market Guide for SPM. Its positioning is breadth: one system of record for variable pay spanning sales commissions, non-sales MBO, and channel and loyalty incentives, on a calculation-and-simulation engine with no-code plan configuration aimed at comp administrators rather than engineers. There is also an agent layer, Akyba — four agents, a shared knowledge base and bring-your-own-LLM support across GPT, Claude and Gemini — in which the Compensation Admin Assistant drafts plan proposals for an administrator to review, approve and launch. On our own hands-on evaluation that layer is improving quickly: an emerging capability, not a settled one.

Be equally plain about the limits, which map onto three of the six dimensions. Third-party review coverage is thin — a handful of Capterra reviews and no meaningful G2 presence — so you cannot triangulate with crowd data as you can with CaptivateIQ or Spiff, and reference calls carry more weight. Enterprise scale is unproven in our research: if your total sits at the top of the Complex band on payee volume and multi-entity reach, do not set this vendor against Varicent, Xactly or SAP on that axis. If dimension six is your high score, treat ASC 606 and IFRS 15 handling as unverified — ask for the amortisation engine in product documentation, with a worked example on your own contract pattern, before it counts. Pricing is quote-only and there is no free trial. A strong candidate where the complexity is breadth and you will accept being an earlyish North American reference; not a default.

Using the Output: Requirements, Shortlist, Proof of Concept

A score you do not convert into artefacts is forgotten by the second demo. Three conversions.

Requirements. Turn each dimension into statements containing a test, not an adjective. Not "flexible crediting" but "credits one booking to the account owner, one overlay and one manager, with split percentages read from two CRM fields." Not "606 support" but "produces a contract-level amortisation schedule where the period exceeds the initial contract term because renewal commissions are not commensurate." A requirement a vendor can agree to without demonstrating anything is a talking point.

Shortlist. Three vendors from one band, chosen on your highest-scoring dimension rather than category reputation. Longer lists dilute attention across demos designed to look alike. Add one vendor from an adjacent band only if you are within two points of a boundary, then make the boundary question the explicit subject of both demos.

Proof of concept. Run your single most complex plan, on your own data, for a period you have already closed and whose correct answer you know. Comparing against a known result turns a demo into a test. Force these features into it:

• A retroactive tier true-up: a rep crosses 100 percent in month eight and every prior dollar re-rates. Show the catch-up and its audit trail.
• A three-way split with percentages read from CRM fields, including one deal where the splits do not total 100 percent.
• A mid-period plan change with an effective date, then a re-run of the prior period showing it unchanged.
• A clawback on a deal cancelled after payment cleared, and the mechanism that recovers it.
• A non-recoverable draw against a monthly minimum, for a rep below quota two months running.
• A payout in a second currency, on your FX rate-lock policy and rate source, not the vendor's default.
• One manual adjustment — then the register of every manual adjustment made during the test.
• The GL export by entity and cost centre, plus the accrual for a deal booked but not yet invoiced.

Score the POC on three things and write them down: variance against the answer you already knew, elapsed hours to configure, and who did the configuring. The last decides your run rate. If the vendor's services team built everything while your admin watched, you tested the vendor and not yourself, and your first mid-year change becomes a support ticket rather than an afternoon. Insist your administrator configures one variant and one mid-period change unaided. Then ask for a payout-accuracy service level in writing; reluctance is itself a finding.

Frequently Asked Questions