Raymond Panko's review of spreadsheet audits found errors in at least 86% of the spreadsheets examined. The incentive teams I meet know this already. They keep Excel anyway, because it is flexible, it is theirs, and a wrong first payout on a new system would cost months of trust with the field.
That fear is the right one to plan around. An Excel to incentive software migration succeeds or fails in its parallel runs. A closed period is recalculated, and every payee difference from the spreadsheet is explained before anyone is paid from the new system.
The work divides into four phases: plan inventory and data mapping, configuration, a first parallel run and a second run with sign-off. Five decisions made before kickoff, from frozen plan versions to a written reconciliation rule, decide how long it takes. A reference build reaches a first live payout in 8 weeks.
A question worth asking the head of sales operations this week: how many files fed the previous payout cycle, and who could rebuild them if needed?
Your data is in safe hands. Check out our Privacy policy for more info
Raymond Panko's review of spreadsheet audits found errors in at least 86% of the spreadsheets examined. The incentive teams I meet know this already. They keep Excel anyway, because it is flexible, it is theirs, and a wrong first payout on a new system would cost months of trust with the field.
That fear is the right one to plan around. An Excel to incentive software migration succeeds or fails in its parallel runs. A closed period is recalculated, and every payee difference from the spreadsheet is explained before anyone is paid from the new system.
The work divides into four phases: plan inventory and data mapping, configuration, a first parallel run and a second run with sign-off. Five decisions made before kickoff, from frozen plan versions to a written reconciliation rule, decide how long it takes. A reference build reaches a first live payout in 8 weeks.
A question worth asking the head of sales operations this week: how many files fed the previous payout cycle, and who could rebuild them if needed?
Raymond Panko's review of spreadsheet audits, presented to the European Spreadsheet Risks Interest Group, found errors in at least 86% of the spreadsheets audited in the more rigorous recent studies. Cell error rates in those audits ran from about 0.4% to 2.5%.
At that rate, a payout workbook with thousands of formulas carries errors as a matter of course. The cost can be large. In 2013 the US Securities and Exchange Commission said JPMorgan had admitted to spreadsheet miscalculations that caused large valuation errors, inside losses of about $5.8 billion.
Payroll has its own examples. EuSpRIG records a US county that overpaid $494,157 after salary worksheets lost their formulas through cutting and pasting.
Sources: Panko, Spreadsheet Errors: What We Know. What We Think We Can Do, EuSpRIG Symposium, 2000 · SEC press release 2013-187 and JPMorgan Management Task Force report, January 2013 · EuSpRIG horror stories, Clallam County, 2017
A workbook with thousands of formulas carries errors as a matter of course, and incentive workbooks are no exception.
Excel stays because it works for the people who built it. The incentive analyst can change a slab in minutes, test a scheme overnight and answer a sales head the same day. No ticket, no vendor and no release cycle.
The hesitation about moving comes from the field. A payout that changes method in the middle of a year can confuse thousands of reps, and a wrong first payout costs months of trust. That fear is reasonable, and a migration plan should answer it directly.
Whether incentives can be run in Excel at all is covered in the guide to managing sales incentives. This post assumes the decision to move is made and sets out how to do it safely.
Excel stays because it is flexible and familiar, so a migration has to protect the field's first payout above everything else.
Four signs show a payout process has grown past what a set of workbooks can carry.
Signs that incentive calculation has outgrown Excel
| Sign | What it means | What it costs |
|---|---|---|
| Seven or more files feed one payout cycle | Links between files break and versions drift | Errors that repeat every cycle |
| Payout arrives a month or more after the close | The calculation is rebuilt by hand each time | The field acts on the previous quarter's plan |
| One person holds the logic | The process stops when that person is on leave | Key-person risk the auditor will note |
| Plan changes are made in the live workbook | No record of which rule paid which period | Disputes that cannot be settled from the file |
Each of these produces the commission calculation errors that repeat every cycle, and each one sends reps to keep their own shadow account.
When the files multiply, the close stretches and one person holds the logic, the workbook has become the risk.
A migration runs in four phases. The timing below follows a reference build, and scope changes it.
An eight-week Excel to incentive software migration, by phase
| Weeks | Phase | What is produced |
|---|---|---|
| 1 to 2 | Plan inventory and data mapping | Every plan, KPI, slab and gate listed, and each input mapped to its source system |
| 2 to 4 | Configuration and connection | Plans built in the engine, connectors to HRMS, SFA, CRM and ERP, uploads where no system exists |
| 4 to 6 | First parallel run | A closed period recalculated and reconciled against the spreadsheet payout, payee by payee |
| 6 to 8 | Second parallel run and sign-off | Finance signs off, reps get the app and the first live payout runs |

The parallel run is where trust is built. Every difference between the spreadsheet and the engine is explained, payee by payee. Some will be engine errors to fix. Many will be spreadsheet errors that had gone unnoticed.
A migration earns the field's trust in the parallel run, where every difference from the old payout is explained before go-live.
Five decisions shorten a migration more than any feature does.
Frozen plan versions, named data sources, agreed history and a written reconciliation rule decide the length of a migration.
Our principle is to fill the white spaces and leave the foundation alone. HR keeps working in the HRMS, sales in the CRM and SFA, and finance in the ERP. We connect to each of them and retire the spreadsheets that sit between them.
Configuration with AI. Admins define KPIs by describing them in a prompt, using our AI KPI builder, live today. The same prompt-based approach extends to whole plans shortly, which shortens the configuration phase further.
Parallel runs and audit trail. Every payout drills down to the data behind it, and every change sits in a change log. Plans are held as dated versions, so the parallel run compares the same rules.
Field launch with AI. Reps get a plain-language AI summary of what they have earned, and AI-assisted query resolution handles many first-cycle questions. Every payout traces back to the plan and KPI behind it, and tracing to the underlying data follows by the end of December 2026.
From kickoff to first live payout, our reference build takes 8 weeks. One mid-sized Indian pharma deployment has more than 1,000 medical representatives. There, the seven to ten spreadsheets used each cycle fell to none, and the incentive team saves more than 200 hours a month.
Source: Kennect claim register · time to first live payout, kickoff to first live payout, reference build · spreadsheets per cycle and admin hours, anonymized mid-sized Indian pharma deployment, 1,000+ medical representatives
A migration that connects to existing systems, configures plans with AI and proves itself in parallel runs reaches a live payout in weeks.
A financial-year plan launches on 1 April. A migration that starts in January can run two parallel cycles on the current plan and go live with the new one. Starting later means another year of the same workbooks.
A migration proved in parallel runs, payee by payee, gives the field a first payout it can trust.
Field audits found errors in 86% of spreadsheets. An eight-week reference build moves incentive calculation off workbooks without changing the systems teams already use.
See how Kennect calculates sales commission and incentive payouts.
Try the sales commission calculator →Comply · Compound · Coach




A reference build runs in about eight weeks from kickoff to first live payout. That covers two weeks of data mapping, two of configuration and four of parallel runs and sign-off. More plans, more data sources or missing history lengthen it. Starting in January leaves room to go live for an April plan year.
Formula errors, broken links between files, version drift when plans change mid-year and key-person risk when one analyst holds the logic. Audits of spreadsheets have found errors in at least 86% of files examined. Errors in a payout workbook reach every payee it calculates.
Recalculate a closed period in the new system using the same plan version and data as the spreadsheet. Compare payouts payee by payee and explain every difference in writing. Fix engine errors, record spreadsheet errors and repeat on a second period before finance signs off.
A list of every plan, KPI, slab and gate in use, plus hierarchy history from the HRMS. Add activity data from the SFA or CRM and sales data from the ERP or DMS. Name one source of record for each input, and bring around twelve months of history so recomputes can reach back.
Yes, for analysis and ad hoc reports on data exported from the engine. Keep calculation, approvals and the record of each payout in the system, so there is one version of what was paid and why.