CRM hygiene campaigns fail because reps gain nothing from them. The fix is not another campaign. Make the CRM fields that matter the inputs to each rep's pay, show the result live, and reps will correct close dates, amounts, owners and products themselves, because a wrong field is now a wrong paycheque. Data quality stops being a project and becomes a side effect of how people are paid.
Below: why hygiene campaigns fade, which fields matter, how to design a plan so the right fields drive pay, the risks and how to control them, how to measure data quality, and what RevOps gets downstream. It includes new data from 2,375 public sales job ads on how companies ask reps to look after the CRM.
Key findings
- 9.0% of 2,375 open sales job ads (879 companies, read 2 November 2026) ask reps to keep the CRM accurate, updated or clean.
- 36.7% of the ads mention the CRM at all. 21.1% name a CRM product, most often HubSpot (13.6%) or Salesforce (10.4%).
- 21.8% say the role pays commission or OTE.
- 0.04%: one ad in 2,375 put CRM and commission in the same sentence, and it was about administering commission plans, not about how reps are paid.
Why do CRM hygiene campaigns fail?
CRM hygiene campaigns ask reps to do work that mainly helps other people: the forecast, management reporting, marketing attribution. Reps comply while the pressure lasts and stop when it moves on. The data decays back to where it was, and the next campaign starts.
The cycle is familiar to most RevOps teams. A board meeting or a quarterly review exposes a forecast nobody trusts. RevOps runs a cleanup sprint, adds required fields and sends a weekly report of deals with missing data. For a few weeks the numbers improve. Then the quarter ends, attention moves, and the fields go stale again.
Each common tactic has a built-in weakness:
- Required fields produce values, not truth. A required close date gets filled with the last day of the quarter. A required amount gets a round placeholder.
- Nag reports work as long as someone reads them aloud in a meeting. They stop working when the meeting changes agenda.
- Manager inspection covers the deals the manager asks about, usually the big ones. Small deals and old records are rarely checked.
- Training explains why data matters to the company. It does not change what the data is worth to the rep.
The common thread is the payoff. A rep who spends ten minutes fixing a close date gets nothing they can feel. Until that changes, hygiene is an ask, and asks fade.
Why does live commission change how reps treat the CRM?
When a rep's commission is calculated from CRM fields and shown to them while the period is open, a wrong field becomes a wrong number on their own statement. People check the numbers they are paid on. The rep becomes the first person to notice a missing deal or a wrong amount, and the person best placed to fix it.
This is a change of incentive, not of discipline. The rep is asked to do nothing new. They look at their earnings, see that a deal they signed is missing, open the CRM and find that the deal is still in "contract sent". They move it to closed won. The CRM is now correct, and nobody had to ask.
Anders Goosmann, CEO of Prowi, saw the other side of this at Webshipper, where he ran a 10-person sales team up to its exit in 2021. Commission ran in a spreadsheet, and the errors surfaced at month-end as disputes, clawbacks and long email threads. The data problems were the same ones a live statement would have shown during the month. They just arrived too late to fix cheaply.
His summary of the difference: "A dashboard shows what a deal is worth to the company. Prowi shows what it's worth to the rep." A dashboard is read by managers. A personal earnings view is read by the person who owns the deal.
Prowi's founders see this effect across customers: when reps and their managers see the same live numbers, deals get kept up to date without anyone being told to. It is an observation, not a measured result, and we do not put a percentage on it. You can test it in your own CRM with the metrics later in this article. The motivation side is covered in why reps who see commission in real time perform better.
Which CRM fields actually matter for commission?
Every deal that affects pay must answer five questions: what was sold, whose deal it is, when it counts, what state it is in, and how much it is worth. The fields behind those five questions are the ones a commission plan can keep accurate. Most other fields need a different mechanism.
| Question | Typical CRM fields | What usually goes wrong | Effect on pay if wrong |
|---|---|---|---|
| What was sold? | Products or line items, deal type (new, expansion, renewal) | No line items, deal type left at the default | Wrong rate applied, product kickers missed |
| Whose deal is it? | Deal owner, second owner or split field, SDR field | Owner left as the person who created the record, SDR not recorded | Commission paid to the wrong person, or split disputes |
| When does it count? | Close date or won date | Expected close date never updated at signing | Deal lands in the wrong period or the wrong tier |
| What state is it in? | Deal stage or status, cancellation flag | Signed deals left open, cancelled deals still marked won | Earned commission not paid, or a clawback missed |
| How much is it worth? | Amount, currency, contract term, discount | Amount is the early estimate, not the signed value | Overpayment or underpayment |
The names differ between systems. HubSpot has deal owner, close date, amount and line items. Pipedrive has owner, value, won time and products. Salesforce has opportunity owner, close date, amount and opportunity products. The five questions are the same. For system-specific detail, see HubSpot commission tracking and Pipedrive commission tracking.
Which fields drive pay, and which do you still clean by campaign?
Pay keeps a field accurate only when the field changes the rep's number. Fields that never touch pay, such as lead source, industry or lost reason, still need automation, enrichment or manager review. Be explicit about which is which, so nobody expects the plan to fix everything.
| Field | Drives pay in most plans? | Who notices an error first | Best way to keep it accurate |
|---|---|---|---|
| Amount | Yes | The rep | Pay on it, and check it at approval |
| Deal owner | Yes | The rep | Pay on it, and lock it after close |
| Close or won date | Yes | The rep | Pay on it, with one written date rule |
| Stage or status | Yes | The rep | Pay only on closed won |
| Products or line items | When rates differ by product | The rep | Pay on it where the plan differs by product |
| Contract term | When there is a multi-year multiplier | The rep | Pay on it, and check it against the contract |
| Split or SDR field | When credit is shared | Both reps | Pay on it, with a written crediting rule |
| Lead source | No | Marketing | Set by automation at creation |
| Industry and company size | No | RevOps | Enrichment |
| Next step and activities | No | The manager | Pipeline reviews |
| Lost reason | No | Nobody, usually | Required when a deal moves to lost |
| Forecast category or probability | No, and it should not | The manager | Manager review, never pay |
The last row matters. Paying on a field the rep sets freely, with no outside check, invites inflation. Probability and forecast category are judgement fields. Keep them out of the plan.
How do you design a commission plan so the right fields drive pay?
Start from the data you need to be true, make those fields the plan's inputs, and write one rule per field that a rep can understand. Pay only on values that can be checked: signed amounts, real dates, closed stages. Then make the effect visible, because a rule nobody sees changes nothing.
Six design principles do most of the work:
- Pay on the fields you want accurate. If forecasting needs correct close dates, the plan should credit deals on the won date. Reps will then move deals to won on the day they sign.
- Use one date rule. Decide whether the won date, the signature date or the invoice date counts, and write it down. Two date rules guarantee disputes at every period boundary.
- Pay on signed values, not estimates. Use the amount or line items at closed won, not the early estimate. Where you invoice from an ERP, consider paying on the invoiced amount.
- Make shared credit a field. If SDRs or second reps get credit, store it in a field, not in a note. A split that lives in a note cannot be calculated or checked.
- Tie clawbacks to a state. If a deal cancels within the clawback window, the stage or a cancellation flag should say so. Then the clawback follows the data instead of someone's memory.
- Keep judgement fields out. Never pay on probability, forecast category or any field the rep can set without a check.
These principles sit inside the broader rules for a plan people trust, set out in how to design a fair and scalable commission model. For role-specific plans with the fields each one reads, see commission plans by role. To check your current plan, run the commission plan health check.
What does it look like on one rep's statement?
When a rep sees a statement built from CRM fields, the corrections they make show up as money. The example below is illustrative: one rep, one month, a flat 10% commission rate (an assumption), amounts in EUR. It shows four common kinds of fix.
| Fix made in the CRM during October | Deal value after the fix (EUR) | Change in October commission (EUR) |
|---|---|---|
| Amount updated to the signed value after an upsell (was €18,000) | €24,000 | +€600 |
| Owner corrected from the team lead to the rep | €12,000 | +€1,200 |
| Close date corrected to the October signing date | €9,000 | +€900 |
| Duplicate deal record merged | €4,000 | −€400 |
| Total change | +€2,300 |
Three things are worth noticing. The first three fixes were made by the rep, because each one raised their number. The close-date fix moves money between months rather than creating it: without it, the €900 would have been paid in November. And the last row is a fix reps rarely volunteer. It was caught at approval, because the manager sees the same statement. Pay-driven accuracy works best when the rep and the approver look at the same numbers.
How does this play out by team size and CRM?
The mechanism is the same everywhere, but the fields that matter most depend on how you sell. A small SaaS team mostly fixes dates and amounts. A product-heavy business fixes line items. A large team with SDRs fixes owners and splits. The scenarios are illustrative, not drawn from specific companies.
Scenario A: 8 reps on HubSpot, subscription software (illustrative)
Each deal has an amount in EUR and a close date. Before live commission, close dates were the expected date set at qualification and rarely updated, so the forecast drifted at every month-end. The plan pays on closed won, credited on the won date. Once reps see a signed deal missing from this month's statement, they move it to closed won at signing. The close date becomes the real signature date, and the month-end forecast stops drifting for the reason it used to.
Scenario B: 30 reps on Pipedrive, wholesale with product-based rates (illustrative)
The plan pays 4% on standard products and 8% on a strategic product line, in USD. Before, most deals had only a total value and no products, so the higher rate was applied by hand from notes. Once the plan reads products from the deal and the statement shows the rate per line, reps add products to their deals, because missing products cost them the higher rate. RevOps gets product mix by rep and by month for the first time.
Scenario C: 120 reps on Salesforce, SDR and AE split (illustrative)
SDRs earn a share of the deals they source. Before, the SDR was recorded in a free-text field or not at all, and sourcing credit was settled in a monthly spreadsheet review. The plan now reads a dedicated SDR lookup field. SDRs check their statements and chase AEs to fill the field at handover. Attribution of pipeline to outbound becomes reliable as a side effect. The rules for shared credit are in our glossary entry on split commission.
| Dimension | 8 reps, HubSpot | 30 reps, Pipedrive | 120 reps, Salesforce |
|---|---|---|---|
| Fields that drive pay | Amount, close date, stage | Products, value, owner | Owner, SDR field, amount, close date |
| Most common error before | Close date never updated | Deals without products | SDR not recorded |
| What reps start fixing | Stage and close date at signing | Line items on every deal | SDR field at handover |
| What RevOps gets | A forecast that holds at month-end | Product mix by rep | Reliable outbound attribution |
| Main risk to control | Deals pushed across month-end | Products added at a token price | Owner disputes at handover |
Before and after: what changes in the pipeline?
With a hygiene campaign, errors are found late by people who cannot fix them. With pay-driven accuracy, errors are found early by the person who owns the deal. The table compares the two on the points RevOps cares about.
| Dimension | Hygiene campaign | Pay-driven accuracy |
|---|---|---|
| Why a rep fixes a field | Because someone asked | Because their statement is wrong |
| When errors are found | At the review or at month-end | During the period, often the same day |
| Who finds them | RevOps, finance or the manager | The rep who owns the deal |
| How long the effect lasts | Until attention moves on | As long as the plan pays on the field |
| Fields covered | Whatever the campaign targets | The fields that drive pay |
| RevOps effort | Repeated sprints and reports | Rule design once, then exceptions |
| Side effect | Reps see data entry as admin | Reps see the CRM as their pay record |
Can paying on CRM data make the data worse?
Yes, if the plan pays on fields nobody checks. Any number that affects pay will be pushed in the rep's favour where the rules allow it. The answer is not to avoid paying on data. It is to pay on fields that can be verified and to put controls where the incentive is strongest.
| Risk | What it looks like | Control |
|---|---|---|
| Timing games | Deals held back or pulled forward to hit a tier, also called sandbagging | Credit on the signature or invoice date, and review deals that move across period boundaries |
| Inflated amounts | Amount raised after close | Lock amount after closed won, or pay on the invoiced amount |
| Owner grabbing | Owner changed to the rep just before close | Log owner changes and require approval for changes in the last days before close |
| Deal splitting | One order entered as several deals to cross a threshold | Set thresholds per period, not per deal |
| Premature closing | Deals marked won before signature | Clawback rule tied to cancellation, and a signed contract check at approval |
| Neglect of non-pay fields | Pay fields perfect, everything else stale | Keep automation and reviews for non-pay fields, as in the table above |
None of these risks is new. They exist under a spreadsheet plan too, where they are harder to see. A live statement with an audit trail makes them visible while they can still be corrected. Our guide to avoiding commission disputes covers the approval side.
How do you measure CRM data quality once pay depends on it?
Measure the fields that drive pay at the moment they matter: when a deal is won and when a period closes. Count missing values, late changes and corrections after close. Record a baseline before you change anything, so you can see whether the plan made a difference.
- Pay-field completeness at won: the share of closed won deals with amount, owner, close date and, where relevant, products and split fields filled.
- Signature-to-won lag: days between the contract signature and the deal moving to closed won in the CRM. Shorter lag means the CRM reflects reality sooner.
- Close-date changes per won deal: how often the close date moved before the deal was won. Many moves point at optimistic dates or timing games.
- Edits to pay fields after period close: each one is either a correction you need or a control you are missing.
- Statement questions caused by data: the share of rep questions where the answer was "the CRM was wrong". It should fall as reps fix records during the period.
- Duplicates found at approval: a direct count of the errors the incentive does not fix by itself.
Report these monthly next to the commission close. They tell you whether the data is improving, and they show finance that the numbers it approves rest on cleaner records.
What does RevOps get downstream?
Clean pay fields are the same fields forecasting, attribution and pipeline reviews depend on. When reps keep them accurate for their own pay, RevOps gets a more reliable forecast, credible attribution, faster pipeline reviews and a commission cost finance can accrue without guessing.
Forecasting
Forecasts break on two fields: close date and amount. Both drive pay in a well-designed plan. When won deals carry the real signature date and the signed amount, the history your forecast learns from is real. Our article on using commission data to forecast sales goes further.
Attribution
Attribution needs to know who sourced and who closed a deal. When the SDR and split fields drive pay, they get filled, and outbound, partner and inbound sourcing can be compared on the same basis.
Pipeline reviews
When the pay fields are reliable, reviews can skip "is this right?" and go straight to "what will it take to close?". Managers spend the hour on deals, not on data.
Commission cost and accruals
Finance needs to accrue commission before it is paid. With accurate amounts, dates and stages, the commission on closed won deals is known during the period, and the cost of open pipeline can be estimated with the plan rules. That is a better accrual than last month's number plus a guess.
What are the common mistakes?
Most mistakes come from expecting the plan to fix every field, or from paying on fields that cannot be checked. A few come from rolling out the data rules without showing reps the effect.
- Expecting pay to fix fields it does not touch. Lead source and lost reason will not improve because of commission. Keep other mechanisms for them.
- Paying on estimates. If the plan pays on the early amount, reps have a reason to inflate it. Pay on the signed value.
- Hiding the calculation. If reps only see a total at month-end, the link between field and pay is invisible, and nothing changes in the CRM.
- Changing the date rule mid-year. Every change to which date counts moves deals between periods and creates disputes. Settle it before the year starts.
- Storing splits in notes. A split nobody can calculate will be settled by argument.
- Letting anyone edit pay fields after close. Lock or log them, or the audit trail means nothing.
- Skipping the baseline. Without a before measurement, nobody can say whether the data improved.
How do you roll it out in six weeks?
Pick the fields that drive pay, write the rules, record a baseline, make earnings visible, and run one period with a formal approval before you change anything else. Six weeks is enough for a team that already has its plan written down.
| Week | Focus | What to do |
|---|---|---|
| 1 | Fields | List the fields the plan reads, using the five questions. Mark every field that drives pay. |
| 2 | Rules | Write one rule per pay field, including the date rule and the split rule. Agree field permissions after close. |
| 3 | Baseline | Measure pay-field completeness, signature-to-won lag and close-date changes on last quarter's won deals. |
| 4 | Visibility | Give each rep a live view of their own earnings, calculated from the CRM. Show the deals that count and the ones that do not. |
| 5 | Approval | Close the period with statements and an approval step where the manager sees the same numbers as the rep. |
| 6 | Review | Compare the data metrics with the baseline. Fix the rules that caused questions. Decide which non-pay fields need their own mechanism. |
If you are moving from spreadsheets to a system as part of this, our guide to commission software implementation covers timeline and effort, and commission calculator vs commission management system helps you decide whether you need one. For the wider discipline, see our guide to commission management. The operating model around all of this is in our companion piece, commission is a RevOps system, not a payroll task.
Where do the job-ad numbers come from?
The figures come from Prowi's own dataset of public job ads, read on 2 November 2026. They show what companies write in sales ads, not how they run their CRM.
- Sample: 2,375 open sales job ads at 879 companies, from the public careers pages of companies Prowi tracks. Recruitment agencies are excluded. About half of the companies use a .com domain, a fifth a Nordic one.
- Sales ads: titles for selling roles, such as SDR or BDR, account executive, account manager, business developer, sales manager and presales.
- CRM upkeep: the ad asks the rep to keep, maintain, update, log or record data in the CRM, or describes CRM data as accurate, clean or up to date. The keyword pass covers English and Danish wording, so ads in other languages are undercounted.
- CRM and commission in one sentence: a search for both terms within the same sentence. We read the single match.
- Limits: an ad is not a CRM policy, and many companies ask for CRM discipline in the interview rather than the ad. No company is named.
| What the sales ad mentions | Ads (count) | Share of 2,375 ads (%) |
|---|---|---|
| The word CRM | 871 | 36.7% |
| A named CRM product | 502 | 21.1% |
| HubSpot | 323 | 13.6% |
| Salesforce | 248 | 10.4% |
| Pipedrive | 19 | 0.8% |
| Commission or OTE for the role | 517 | 21.8% |
| Forecasting | 419 | 17.6% |
| Keeping the CRM accurate or up to date | 214 | 9.0% |
| CRM upkeep and commission in the same ad | 66 | 2.8% |
| CRM and commission in the same sentence | 1 | 0.04% |
One detail stands out. Ads that mention commission were somewhat more likely to ask for CRM upkeep (12.8%) than ads that did not (8.0%). That may only mean those ads are more detailed. Either way, almost none of them tell candidates that pay and CRM data are connected.
FAQ
How do you get sales reps to update the CRM?
Make the fields you need part of how they are paid, and show them the effect while the period is open. When a missing close date means a missing deal on their statement, reps update it without being asked. Keep automation and manager reviews for fields that do not affect pay.
Which CRM fields should drive commission?
The fields that answer five questions: what was sold (products, deal type), whose deal it is (owner, split, SDR), when it counts (won or close date), what state it is in (stage), and how much it is worth (signed amount, term). Never pay on judgement fields such as probability.
Won't reps manipulate CRM data if their pay depends on it?
Some will try, under any plan. Pay on values that can be checked, lock pay fields after close, log changes, and let managers approve statements that show the same numbers as the rep sees. Live data makes manipulation visible earlier than a spreadsheet does.
Does real-time commission improve forecast accuracy?
It can improve the inputs. Forecasts depend on close dates and amounts, and those are the fields a well-designed plan pays on. Measure it on your own data: compare close-date changes and signature-to-won lag before and after.
Do we still need CRM hygiene campaigns?
For fields that do not affect pay, yes, though usually as automation and reviews rather than campaigns. Lead source, industry and lost reason need their own mechanisms. Pay-driven accuracy covers the fields that matter most for revenue, not all of them.
Want to see your own deals turned into live commission statements? Book a demo and we will build one from your CRM data.