Disputes are not a sign that something went wrong. People care about money, plans are complicated, and someone will always want to check a line. The question is what happens next.
Without provenance, a dispute becomes a week of reconstruction. With it, a dispute becomes a five minute reading of one chain. The plan text is usually identical in both cases. The difference is whether anyone can show the line behind the number.
A week without a trace
Monday
Finance sends the payout run. The file looks final and the thread stays quiet for a few hours. Quiet is not agreement. Quiet is usually the interval before a private spreadsheet comes out.
Tuesday
A rep flags one line. The message is short and the attachment is a personal workbook. There are now two numbers for the same deal and no shared map between them, which means the conversation has no floor.
Wednesday
A manager, a finance partner, and the rep reconstruct history out of email, chat, and memory. Someone locates an old SPIF announcement. Someone else opens a different plan PDF. The rate becomes a question of who remembers with more confidence.
Thursday
The escalation widens, not because the plan was wrong, but because nobody can prove which clause produced the line. Time leaves the building. Some trust leaves with it, and it does not come back at the start of the next cycle.
The plan was never the problem. The missing piece was provenance, and without it both sides will invent one.
The same week with a trace
Now run the identical week with a certified run behind it. The dispute still happens on Tuesday. The archaeology does not, because every line already points at a chain with four links.
- The deal. Open the opportunity that fed the line: closed amount, close date, owner. Those commercial facts stay in the CRM, where they belong. Taurza read them on a read-only basis and wrote nothing back, so the deal record is exactly what the sales team sees.
- The attainment position. Show where the rep sat on the published ladder at the moment that deal applied. Base band or accelerator, stated as a calculated position against quota rather than a recollection of roughly where they were.
- The rate applied. Name the rate in plain units, and if the deal crossed a band edge, show that crossing explicitly. This is the step where most arguments actually end, because the rate stops being an assertion and becomes a consequence.
- The clause. Point at the clause in the plan of record that produced that rate. Finance and the rep are now reading the same sentence in the same document, and the conversation is about what the sentence says rather than who remembers it correctly.
Four links, in order, on every line. The rep can walk the chain without asking for permission, and finance does not have to reconstruct anything to answer them. Notice how much of the old week was spent proving that a number was legitimate rather than discussing whether it was correct.
What a trace is not
It is worth being clear about the limits, because provenance gets oversold and then disappoints people who expected something else from it.
- It is not a guarantee that the plan was well designed. A trace will faithfully show you a clause that pays badly.
- It is not a substitute for telling reps about a change. Provenance explains a number after the fact; publishing prevents the surprise.
- It is not an approval workflow. Deciding whether an exception should be granted is still a human judgment.
- It is not a legal opinion. Compliance with wage and employment law stays with you.
What a trace does is narrow. It removes the question of whether a number is legitimate, so the conversation can be about whether it is right. That is a smaller claim than most tools make and a more useful one in practice, because almost all of the time lost in a bad cycle is spent on the first question rather than the second.
House rules that keep the cycle clean
A trace only stays useful if the surrounding habits protect it. These are the rules worth holding:
- Publish exactly one run for the cycle, and treat it as the map.
- Require every disputed line to open the trace before anything else happens.
- Land mid-cycle exceptions in the plan of record before recalculating, never after.
- Keep payroll as the only place money actually moves.
- Answer the first dispute of the cycle with the trace, because it sets the pattern for the rest.
What to refuse
Equally important is the short list of things to say no to, each of which quietly dismantles provenance:
- A second shadow sheet, created just for this cycle, just this once.
- A private exception that never lands in the plan of record as text.
- A CRM write-back intended to paper over a calculation question.
- Resolving a line by seniority when the chain is available to read.
That refusal list is not culture theater. It is how you keep the next cycle from inheriting this cycle's exceptions, which is the actual mechanism by which comp processes decay.
You still confirm that your plans and payments comply with wage and employment law wherever you employ people. Taurza calculates and certifies; it does not give legal advice and it never holds or moves money. Payroll pays. What the platform provides is provenance, and provenance is what turns a dispute into evidence instead of an argument.
Disputes will keep happening. The difference is whether Tuesday becomes a reading or an excavation.